3 \cfg{text-title-align}{left}
4 \cfg{text-chapter-align}{left}
5 \cfg{text-chapter-numeric}{true}
6 \cfg{text-chapter-suffix}{. }
7 \cfg{text-chapter-underline}{-}
8 \cfg{text-section-align}{0}{left}
9 \cfg{text-section-numeric}{0}{true}
10 \cfg{text-section-suffix}{0}{. }
11 \cfg{text-section-underline}{0}{-}
12 \cfg{text-section-align}{1}{left}
13 \cfg{text-section-numeric}{1}{true}
14 \cfg{text-section-suffix}{1}{. }
15 \cfg{text-section-underline}{1}{-}
16 \cfg{text-versionid}{0}
18 \cfg{html-contents-filename}{index.html}
19 \cfg{html-template-filename}{%k.html}
20 \cfg{html-index-filename}{docindex.html}
21 \cfg{html-leaf-level}{1}
22 \cfg{html-contents-depth-0}{1}
23 \cfg{html-contents-depth-1}{3}
24 \cfg{html-leaf-contains-contents}{true}
26 \define{dash} \u2013{-}
28 \title Developer documentation for Simon Tatham's puzzle collection
30 This is a guide to the internal structure of Simon Tatham's Portable
31 Puzzle Collection (henceforth referred to simply as \q{Puzzles}),
32 for use by anyone attempting to implement a new puzzle or port to a
35 This guide is believed correct as of r6190. Hopefully it will be
36 updated along with the code in future, but if not, I've at least
37 left this version number in here so you can figure out what's
38 changed by tracking commit comments from there onwards.
40 \C{intro} Introduction
42 The Puzzles code base is divided into four parts: a set of
43 interchangeable front ends, a set of interchangeable back ends, a
44 universal \q{middle end} which acts as a buffer between the two, and
45 a bunch of miscellaneous utility functions. In the following
46 sections I give some general discussion of each of these parts.
48 \H{intro-frontend} Front end
50 The front end is the non-portable part of the code: it's the bit
51 that you replace completely when you port to a different platform.
52 So it's responsible for all system calls, all GUI interaction, and
53 anything else platform-specific.
55 The current front ends in the main code base are for Windows, GTK
56 and MacOS X; I also know of a third-party front end for PalmOS.
58 The front end contains \cw{main()} or the local platform's
59 equivalent. Top-level control over the application's execution flow
60 belongs to the front end (it isn't, for example, a set of functions
61 called by a universal \cw{main()} somewhere else).
63 The front end has complete freedom to design the GUI for any given
64 port of Puzzles. There is no centralised mechanism for maintaining
65 the menu layout, for example. This has a cost in consistency (when I
66 \e{do} want the same menu layout on more than one platform, I have
67 to edit two pieces of code in parallel every time I make a change),
68 but the advantage is that local GUI conventions can be conformed to
69 and local constraints adapted to. For example, MacOS X has strict
70 human interface guidelines which specify a different menu layout
71 from the one I've used on Windows and GTK; there's nothing stopping
72 the OS X front end from providing a menu layout consistent with
75 Although the front end is mostly caller rather than the callee in
76 its interactions with other parts of the code, it is required to
77 implement a small API for other modules to call, mostly of drawing
78 functions for games to use when drawing their graphics. The drawing
79 API is documented in \k{drawing}; the other miscellaneous front end
80 API functions are documented in \k{frontend-api}.
82 \H{intro-backend} Back end
84 A \q{back end}, in this collection, is synonymous with a \q{puzzle}.
85 Each back end implements a different game.
87 At the top level, a back end is simply a data structure, containing
88 a few constants (flag words, preferred pixel size) and a large
89 number of function pointers. Back ends are almost invariably callee
90 rather than caller, which means there's a limitation on what a back
91 end can do on its own initiative.
93 The persistent state in a back end is divided into a number of data
94 structures, which are used for different purposes and therefore
95 likely to be switched around, changed without notice, and otherwise
96 updated by the rest of the code. It is important when designing a
97 back end to put the right pieces of data into the right structures,
98 or standard midend-provided features (such as Undo) may fail to
101 The functions and variables provided in the back end data structure
102 are documented in \k{backend}.
104 \H{intro-midend} Middle end
106 Puzzles has a single and universal \q{middle end}. This code is
107 common to all platforms and all games; it sits in between the front
108 end and the back end and provides standard functionality everywhere.
110 People adding new back ends or new front ends should generally not
111 need to edit the middle end. On rare occasions there might be a
112 change that can be made to the middle end to permit a new game to do
113 something not currently anticipated by the middle end's present
114 design; however, this is terribly easy to get wrong and should
115 probably not be undertaken without consulting the primary maintainer
116 (me). Patch submissions containing unannounced mid-end changes will
117 be treated on their merits like any other patch; this is just a
118 friendly warning that mid-end changes will need quite a lot of
119 merits to make them acceptable.
121 Functionality provided by the mid-end includes:
123 \b Maintaining a list of game state structures and moving back and
124 forth along that list to provide Undo and Redo.
126 \b Handling timers (for move animations, flashes on completion, and
127 in some cases actually timing the game).
129 \b Handling the container format of game IDs: receiving them,
130 picking them apart into parameters, description and/or random seed,
131 and so on. The game back end need only handle the individual parts
132 of a game ID (encoded parameters and encoded game description);
133 everything else is handled centrally by the mid-end.
135 \b Handling standard keystrokes and menu commands, such as \q{New
136 Game}, \q{Restart Game} and \q{Quit}.
138 \b Pre-processing mouse events so that the game back ends can rely
139 on them arriving in a sensible order (no missing button-release
140 events, no sudden changes of which button is currently pressed,
143 \b Handling the dialog boxes which ask the user for a game ID.
145 \b Handling serialisation of entire games (for loading and saving a
146 half-finished game to a disk file, or for handling application
147 shutdown and restart on platforms such as PalmOS where state is
148 expected to be saved).
150 Thus, there's a lot of work done once by the mid-end so that
151 individual back ends don't have to worry about it. All the back end
152 has to do is cooperate in ensuring the mid-end can do its work
155 The API of functions provided by the mid-end to be called by the
156 front end is documented in \k{midend}.
158 \H{intro-utils} Miscellaneous utilities
160 In addition to these three major structural components, the Puzzles
161 code also contains a variety of utility modules usable by all of the
162 above components. There is a set of functions to provide
163 platform-independent random number generation; functions to make
164 memory allocation easier; functions which implement a balanced tree
165 structure to be used as necessary in complex algorithms; and a few
166 other miscellaneous functions. All of these are documented in
169 \H{intro-structure} Structure of this guide
171 There are a number of function call interfaces within Puzzles, and
172 this guide will discuss each one in a chapter of its own. After
173 that, (\k{writing}) discusses how to design new games, with some
174 general design thoughts and tips.
176 \C{backend} Interface to the back end
178 This chapter gives a detailed discussion of the interface that each
179 back end must implement.
181 At the top level, each back end source file exports a single global
182 symbol, which is a \c{const struct game} containing a large number
183 of function pointers and a small amount of constant data. This
184 structure is called by different names depending on what kind of
185 platform the puzzle set is being compiled on:
187 \b On platforms such as Windows and GTK, which build a separate
188 binary for each puzzle, the game structure in every back end has the
189 same name, \cq{thegame}; the front end refers directly to this name,
190 so that compiling the same front end module against a different back
191 end module builds a different puzzle.
193 \b On platforms such as MacOS X and PalmOS, which build all the
194 puzzles into a single monolithic binary, the game structure in each
195 back end must have a different name, and there's a helper module
196 \c{list.c} (constructed automatically by the same Perl script that
197 builds the \cw{Makefile}s) which contains a complete list of those
200 On the latter type of platform, source files may assume that the
201 preprocessor symbol \c{COMBINED} has been defined. Thus, the usual
202 code to declare the game structure looks something like this:
205 \c #define thegame net /* or whatever this game is called */
206 \e iii iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
209 \c const struct game thegame = {
210 \c /* lots of structure initialisation in here */
211 \e iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
214 Game back ends must also internally define a number of data
215 structures, for storing their various persistent state. This chapter
216 will first discuss the nature and use of those structures, and then
217 go on to give details of every element of the game structure.
219 \H{backend-structs} Data structures
221 Each game is required to define four separate data structures. This
222 section discusses each one and suggests what sorts of things need to
225 \S{backend-game-params} \c{game_params}
227 The \c{game_params} structure contains anything which affects the
228 automatic generation of new puzzles. So if puzzle generation is
229 parametrised in any way, those parameters need to be stored in
232 Most puzzles currently in this collection are played on a grid of
233 squares, meaning that the most obvious parameter is the grid size.
234 Many puzzles have additional parameters; for example, Mines allows
235 you to control the number of mines in the grid independently of its
236 size, Net can be wrapping or non-wrapping, Solo has difficulty
237 levels and symmetry settings, and so on.
239 A simple rule for deciding whether a data item needs to go in
240 \c{game_params} is: would the user expect to be able to control this
241 data item from either the preset-game-types menu or the \q{Custom}
242 game type configuration? If so, it's part of \c{game_params}.
244 \c{game_params} structures are permitted to contain pointers to
245 subsidiary data if they need to. The back end is required to provide
246 functions to create and destroy \c{game_params}, and those functions
247 can allocate and free additional memory if necessary. (It has not
248 yet been necessary to do this in any puzzle so far, but the
249 capability is there just in case.)
251 \c{game_params} is also the only structure which the game's
252 \cw{compute_size()} function may refer to; this means that any
253 aspect of the game which affects the size of the window it needs to
254 be drawn in must be stored in \c{game_params}. In particular, this
255 imposes the fundamental limitation that random game generation may
256 not have a random effect on the window size: game generation
257 algorithms are constrained to work by starting from the grid size
258 rather than generating it as an emergent phenomenon. (Although this
259 is a restriction in theory, it has not yet seemed to be a problem.)
261 \S{backend-game-state} \c{game_state}
263 While the user is actually playing a puzzle, the \c{game_state}
264 structure stores all the data corresponding to the current state of
267 The mid-end keeps \c{game_state}s in a list, and adds to the list
268 every time the player makes a move; the Undo and Redo functions step
269 back and forth through that list.
271 Therefore, a good means of deciding whether a data item needs to go
272 in \c{game_state} is: would a player expect that data item to be
273 restored on undo? If so, put it in \c{game_state}, and this will
274 automatically happen without you having to lift a finger. If not
275 \dash for example, the deaths counter in Mines is precisely
276 something that does \e{not} want to be reset to its previous state
277 on an undo \dash then you might have found a data item that needs to
278 go in \c{game_ui} instead.
280 During play, \c{game_state}s are often passed around without an
281 accompanying \c{game_params} structure. Therefore, any information
282 in \c{game_params} which is important during play (such as the grid
283 size) must be duplicated within the \c{game_state}. One simple
284 method of doing this is to have the \c{game_state} structure
285 \e{contain} a \c{game_params} structure as one of its members,
286 although this isn't obligatory if you prefer to do it another way.
288 \S{backend-game-drawstate} \c{game_drawstate}
290 \c{game_drawstate} carries persistent state relating to the current
291 graphical contents of the puzzle window. The same \c{game_drawstate}
292 is passed to every call to the game redraw function, so that it can
293 remember what it has already drawn and what needs redrawing.
295 A typical use for a \c{game_drawstate} is to have an array mirroring
296 the array of grid squares in the \c{game_state}; then every time the
297 redraw function was passed a \c{game_state}, it would loop over all
298 the squares, and physically redraw any whose description in the
299 \c{game_state} (i.e. what the square needs to look like when the
300 redraw is completed) did not match its description in the
301 \c{game_drawstate} (i.e. what the square currently looks like).
303 \c{game_drawstate} is occasionally completely torn down and
304 reconstructed by the mid-end, if the user somehow forces a full
305 redraw. Therefore, no data should be stored in \c{game_drawstate}
306 which is \e{not} related to the state of the puzzle window, because
307 it might be unexpectedly destroyed.
309 The back end provides functions to create and destroy
310 \c{game_drawstate}, which means it can contain pointers to
311 subsidiary allocated data if it needs to. A common thing to want to
312 allocate in a \c{game_drawstate} is a \c{blitter}; see
313 \k{drawing-blitter} for more on this subject.
315 \S{backend-game-ui} \c{game_ui}
317 \c{game_ui} contains whatever doesn't fit into the above three
320 A new \c{game_ui} is created when the user begins playing a new
321 instance of a puzzle (i.e. during \q{New Game} or after entering a
322 game ID etc). It persists until the user finishes playing that game
323 and begins another one (or closes the window); in particular,
324 \q{Restart Game} does \e{not} destroy the \c{game_ui}.
326 \c{game_ui} is useful for implementing user-interface state which is
327 not part of \c{game_state}. Common examples are keyboard control
328 (you wouldn't want to have to separately Undo through every cursor
329 motion) and mouse dragging. See \k{writing-keyboard-cursor} and
330 \k{writing-howto-dragging}, respectively, for more details.
332 Another use for \c{game_ui} is to store highly persistent data such
333 as the Mines death counter. This is conceptually rather different:
334 where the Net cursor position was \e{not important enough} to
335 preserve for the player to restore by Undo, the Mines death counter
336 is \e{too important} to permit the player to revert by Undo!
338 A final use for \c{game_ui} is to pass information to the redraw
339 function about recent changes to the game state. This is used in
340 Mines, for example, to indicate whether a requested \q{flash} should
341 be a white flash for victory or a red flash for defeat; see
342 \k{writing-flash-types}.
344 \H{backend-simple} Simple data in the back end
346 In this section I begin to discuss each individual element in the
347 back end structure. To begin with, here are some simple
348 self-contained data elements.
350 \S{backend-name} \c{name}
354 This is a simple ASCII string giving the name of the puzzle. This
355 name will be used in window titles, in game selection menus on
356 monolithic platforms, and anywhere else that the front end needs to
357 know the name of a game.
359 \S{backend-winhelp} \c{winhelp_topic}
361 \c const char *winhelp_topic;
363 This member is used on Windows only, to provide online help.
364 Although the Windows front end provides a separate binary for each
365 puzzle, it has a single monolithic help file; so when a user selects
366 \q{Help} from the menu, the program needs to open the help file and
367 jump to the chapter describing that particular puzzle.
369 Therefore, each chapter in \c{puzzles.but} is labelled with a
370 \e{help topic} name, similar to this:
372 \c \cfg{winhelp-topic}{games.net}
374 And then the corresponding game back end encodes the topic string
375 (here \cq{games.net}) in the \c{winhelp_topic} element of the game
378 \H{backend-params} Handling game parameter sets
380 In this section I present the various functions which handle the
381 \c{game_params} structure.
383 \S{backend-default-params} \cw{default_params()}
385 \c game_params *(*default_params)(void);
387 This function allocates a new \c{game_params} structure, fills it
388 with the default values, and returns a pointer to it.
390 \S{backend-fetch-preset} \cw{fetch_preset()}
392 \c int (*fetch_preset)(int i, char **name, game_params **params);
394 This function is used to populate the \q{Type} menu, which provides
395 a list of conveniently accessible preset parameters for most games.
397 The function is called with \c{i} equal to the index of the preset
398 required (numbering from zero). It returns \cw{FALSE} if that preset
399 does not exist (if \c{i} is less than zero or greater than the
400 largest preset index). Otherwise, it sets \c{*params} to point at a
401 newly allocated \c{game_params} structure containing the preset
402 information, sets \c{*name} to point at a newly allocated C string
403 containing the preset title (to go on the \q{Type} menu), and
406 If the game does not wish to support any presets at all, this
407 function is permitted to return \cw{FALSE} always.
409 \S{backend-encode-params} \cw{encode_params()}
411 \c char *(*encode_params)(game_params *params, int full);
413 The job of this function is to take a \c{game_params}, and encode it
414 in a string form for use in game IDs. The return value must be a
415 newly allocated C string, and \e{must} not contain a colon or a hash
416 (since those characters are used to mark the end of the parameter
417 section in a game ID).
419 Ideally, it should also not contain any other potentially
420 controversial punctuation; bear in mind when designing a string
421 parameter format that it will probably be used on both Windows and
422 Unix command lines under a variety of exciting shell quoting and
423 metacharacter rules. Sticking entirely to alphanumerics is the
424 safest thing; if you really need punctuation, you can probably get
425 away with commas, periods or underscores without causing anybody any
426 major inconvenience. If you venture far beyond that, you're likely
427 to irritate \e{somebody}.
429 (At the time of writing this, all existing games have purely
430 alphanumeric string parameter formats. Usually these involve a
431 letter denoting a parameter, followed optionally by a number giving
432 the value of that parameter, with a few mandatory parts at the
433 beginning such as numeric width and height separated by \cq{x}.)
435 If the \c{full} parameter is \cw{TRUE}, this function should encode
436 absolutely everything in the \c{game_params}, such that a subsequent
437 call to \cw{decode_params()} (\k{backend-decode-params}) will yield
438 an identical structure. If \c{full} is \cw{FALSE}, however, you
439 should leave out anything which is not necessary to describe a
440 \e{specific puzzle instance}, i.e. anything which only takes effect
441 when a new puzzle is \e{generated}. For example, the Solo
442 \c{game_params} includes a difficulty rating used when constructing
443 new puzzles; but a Solo game ID need not explicitly include the
444 difficulty, since to describe a puzzle once generated it's
445 sufficient to give the grid dimensions and the location and contents
446 of the clue squares. (Indeed, one might very easily type in a puzzle
447 out of a newspaper without \e{knowing} what its difficulty level is
448 in Solo's terminology.) Therefore, Solo's \cw{encode_params()} only
449 encodes the difficulty level if \c{full} is set.
451 \S{backend-decode-params} \cw{decode_params()}
453 \c void (*decode_params)(game_params *params, char const *string);
455 This function is the inverse of \cw{encode_params()}
456 (\k{backend-encode-params}). It parses the supplied string and fills
457 in the supplied \c{game_params} structure. Note that the structure
458 will \e{already} have been allocated: this function is not expected
459 to create a \e{new} \c{game_params}, but to modify an existing one.
461 This function can receive a string which only encodes a subset of
462 the parameters. The most obvious way in which this can happen is if
463 the string was constructed by \cw{encode_params()} with its \c{full}
464 parameter set to \cw{FALSE}; however, it could also happen if the
465 user typed in a parameter set manually and missed something out. Be
466 prepared to deal with a wide range of possibilities.
468 When dealing with a parameter which is not specified in the input
469 string, what to do requires a judgment call on the part of the
470 programmer. Sometimes it makes sense to adjust other parameters to
471 bring them into line with the new ones. In Mines, for example, you
472 would probably not want to keep the same mine count if the user
473 dropped the grid size and didn't specify one, since you might easily
474 end up with more mines than would actually fit in the grid! On the
475 other hand, sometimes it makes sense to leave the parameter alone: a
476 Solo player might reasonably expect to be able to configure size and
477 difficulty independently of one another.
479 This function currently has no direct means of returning an error if
480 the string cannot be parsed at all. However, the returned
481 \c{game_params} is almost always subsequently passed to
482 \cw{validate_params()} (\k{backend-validate-params}), so if you
483 really want to signal parse errors, you could always have a \c{char
484 *} in your parameters structure which stored an error message, and
485 have \cw{validate_params()} return it if it is non-\cw{NULL}.
487 \S{backend-free-params} \cw{free_params()}
489 \c void (*free_params)(game_params *params);
491 This function frees a \c{game_params} structure, and any subsidiary
492 allocations contained within it.
494 \S{backend-dup-params} \cw{dup_params()}
496 \c game_params *(*dup_params)(game_params *params);
498 This function allocates a new \c{game_params} structure and
499 initialises it with an exact copy of the information in the one
500 provided as input. It returns a pointer to the new duplicate.
502 \S{backend-can-configure} \c{can_configure}
504 \c int can_configure;
506 This boolean data element is set to \cw{TRUE} if the back end
507 supports custom parameter configuration via a dialog box. If it is
508 \cw{TRUE}, then the functions \cw{configure()} and
509 \cw{custom_params()} are expected to work. See \k{backend-configure}
510 and \k{backend-custom-params} for more details.
512 \S{backend-configure} \cw{configure()}
514 \c config_item *(*configure)(game_params *params);
516 This function is called when the user requests a dialog box for
517 custom parameter configuration. It returns a newly allocated array
518 of \cw{config_item} structures, describing the GUI elements required
519 in the dialog box. The array should have one more element than the
520 number of controls, since it is terminated with a \cw{C_END} marker
521 (see below). Each array element describes the control together with
522 its initial value; the front end will modify the value fields and
523 return the updated array to \cw{custom_params()} (see
524 \k{backend-custom-params}).
526 The \cw{config_item} structure contains the following elements:
533 \c{name} is an ASCII string giving the textual label for a GUI
534 control. It is \e{not} expected to be dynamically allocated.
536 \c{type} contains one of a small number of \c{enum} values defining
537 what type of control is being described. The meaning of the \c{sval}
538 and \c{ival} fields depends on the value in \c{type}. The valid
543 \dd Describes a text input box. (This is also used for numeric
544 input. The back end does not bother informing the front end that the
545 box is numeric rather than textual; some front ends do have the
546 capacity to take this into account, but I decided it wasn't worth
547 the extra complexity in the interface.) For this type, \c{ival} is
548 unused, and \c{sval} contains a dynamically allocated string
549 representing the contents of the input box.
553 \dd Describes a simple checkbox. For this type, \c{sval} is unused,
554 and \c{ival} is \cw{TRUE} or \cw{FALSE}.
558 \dd Describes a drop-down list presenting one of a small number of
559 fixed choices. For this type, \c{sval} contains a list of strings
560 describing the choices; the very first character of \c{sval} is used
561 as a delimiter when processing the rest (so that the strings
562 \cq{:zero:one:two}, \cq{!zero!one!two} and \cq{xzeroxonextwo} all
563 define a three-element list containing \cq{zero}, \cq{one} and
564 \cq{two}). \c{ival} contains the index of the currently selected
565 element, numbering from zero (so that in the above example, 0 would
566 mean \cq{zero} and 2 would mean \cq{two}).
570 Note that for this control type, \c{sval} is \e{not} dynamically
571 allocated, whereas it was for \c{C_STRING}.
577 \dd Marks the end of the array of \c{config_item}s. All other fields
580 The array returned from this function is expected to have filled in
581 the initial values of all the controls according to the input
582 \c{game_params} structure.
584 If the game's \c{can_configure} flag is set to \cw{FALSE}, this
585 function is never called and need not do anything at all.
587 \S{backend-custom-params} \cw{custom_params()}
589 \c game_params *(*custom_params)(config_item *cfg);
591 This function is the counterpart to \cw{configure()}
592 (\k{backend-configure}). It receives as input an array of
593 \c{config_item}s which was originally created by \cw{configure()},
594 but in which the control values have since been changed in
595 accordance with user input. Its function is to read the new values
596 out of the controls and return a newly allocated \c{game_params}
597 structure representing the user's chosen parameter set.
599 (The front end will have modified the controls' \e{values}, but
600 there will still always be the same set of controls, in the same
601 order, as provided by \cw{configure()}. It is not necessary to check
602 the \c{name} and \c{type} fields, although you could use
603 \cw{assert()} if you were feeling energetic.)
605 This function is not expected to (and indeed \e{must not}) free the
606 input \c{config_item} array. (If the parameters fail to validate,
607 the dialog box will stay open.)
609 If the game's \c{can_configure} flag is set to \cw{FALSE}, this
610 function is never called and need not do anything at all.
612 \S{backend-validate-params} \cw{validate_params()}
614 \c char *(*validate_params)(game_params *params, int full);
616 This function takes a \c{game_params} structure as input, and checks
617 that the parameters described in it fall within sensible limits. (At
618 the very least, grid dimensions should almost certainly be strictly
619 positive, for example.)
621 Return value is \cw{NULL} if no problems were found, or
622 alternatively a (non-dynamically-allocated) ASCII string describing
623 the error in human-readable form.
625 If the \c{full} parameter is set, full validation should be
626 performed: any set of parameters which would not permit generation
627 of a sensible puzzle should be faulted. If \c{full} is \e{not} set,
628 the implication is that these parameters are not going to be used
629 for \e{generating} a puzzle; so parameters which can't even sensibly
630 \e{describe} a valid puzzle should still be faulted, but parameters
631 which only affect puzzle generation should not be.
633 (The \c{full} option makes a difference when parameter combinations
634 are non-orthogonal. For example, Net has a boolean option
635 controlling whether it enforces a unique solution; it turns out that
636 it's impossible to generate a uniquely soluble puzzle with wrapping
637 walls and width 2, so \cw{validate_params()} will complain if you
638 ask for one. However, if the user had just been playing a unique
639 wrapping puzzle of a more sensible width, and then pastes in a game
640 ID acquired from somebody else which happens to describe a
641 \e{non}-unique wrapping width-2 puzzle, then \cw{validate_params()}
642 will be passed a \c{game_params} containing the width and wrapping
643 settings from the new game ID and the uniqueness setting from the
644 old one. This would be faulted, if it weren't for the fact that
645 \c{full} is not set during this call, so Net ignores the
646 inconsistency. The resulting \c{game_params} is never subsequently
647 used to generate a puzzle; this is a promise made by the mid-end
648 when it asks for a non-full validation.)
650 \H{backend-descs} Handling game descriptions
652 In this section I present the functions that deal with a textual
653 description of a puzzle, i.e. the part that comes after the colon in
654 a descriptive-format game ID.
656 \S{backend-new-desc} \cw{new_desc()}
658 \c char *(*new_desc)(game_params *params, random_state *rs,
659 \c char **aux, int interactive);
661 This function is where all the really hard work gets done. This is
662 the function whose job is to randomly generate a new puzzle,
663 ensuring solubility and uniqueness as appropriate.
665 As input it is given a \c{game_params} structure and a random state
666 (see \k{utils-random} for the random number API). It must invent a
667 puzzle instance, encode it in string form, and return a dynamically
668 allocated C string containing that encoding.
670 Additionally, it may return a second dynamically allocated string in
671 \c{*aux}. (If it doesn't want to, then it can leave that parameter
672 completely alone; it isn't required to set it to \cw{NULL}, although
673 doing so is harmless.) That string, if present, will be passed to
674 \cw{solve()} (\k{backend-solve}) later on; so if the puzzle is
675 generated in such a way that a solution is known, then information
676 about that solution can be saved in \c{*aux} for \cw{solve()} to
679 The \c{interactive} parameter should be ignored by almost all
680 puzzles. Its purpose is to distinguish between generating a puzzle
681 within a GUI context for immediate play, and generating a puzzle in
682 a command-line context for saving to be played later. The only
683 puzzle that currently uses this distinction (and, I fervently hope,
684 the only one which will \e{ever} need to use it) is Mines, which
685 chooses a random first-click location when generating puzzles
686 non-interactively, but which waits for the user to place the first
687 click when interactive. If you think you have come up with another
688 puzzle which needs to make use of this parameter, please think for
689 at least ten minutes about whether there is \e{any} alternative!
691 Note that game description strings are not required to contain an
692 encoding of parameters such as grid size; a game description is
693 never separated from the \c{game_params} it was generated with, so
694 any information contained in that structure need not be encoded
695 again in the game description.
697 \S{backend-validate-desc} \cw{validate_desc()}
699 \c char *(*validate_desc)(game_params *params, char *desc);
701 This function is given a game description, and its job is to
702 validate that it describes a puzzle which makes sense.
704 To some extent it's up to the user exactly how far they take the
705 phrase \q{makes sense}; there are no particularly strict rules about
706 how hard the user is permitted to shoot themself in the foot when
707 typing in a bogus game description by hand. (For example, Rectangles
708 will not verify that the sum of all the numbers in the grid equals
709 the grid's area. So a user could enter a puzzle which was provably
710 not soluble, and the program wouldn't complain; there just wouldn't
711 happen to be any sequence of moves which solved it.)
713 The one non-negotiable criterion is that any game description which
714 makes it through \cw{validate_desc()} \e{must not} subsequently
715 cause a crash or an assertion failure when fed to \cw{new_game()}
716 and thence to the rest of the back end.
718 The return value is \cw{NULL} on success, or a
719 non-dynamically-allocated C string containing an error message.
721 \S{backend-new-game} \cw{new_game()}
723 \c game_state *(*new_game)(midend *me, game_params *params,
726 This function takes a game description as input, together with its
727 accompanying \c{game_params}, and constructs a \c{game_state}
728 describing the initial state of the puzzle. It returns a newly
729 allocated \c{game_state} structure.
731 Almost all puzzles should ignore the \c{me} parameter. It is
732 required by Mines, which needs it for later passing to
733 \cw{midend_supersede_game_desc()} (see \k{backend-supersede}) once
734 the user has placed the first click. I fervently hope that no other
735 puzzle will be awkward enough to require it, so everybody else
736 should ignore it. As with the \c{interactive} parameter in
737 \cw{new_desc()} (\k{backend-new-desc}), if you think you have a
738 reason to need this parameter, please try very hard to think of an
739 alternative approach!
741 \H{backend-states} Handling game states
743 This section describes the functions which create and destroy
744 \c{game_state} structures.
746 (Well, except \cw{new_game()}, which is in \k{backend-new-game}
747 instead of under here; but it deals with game descriptions \e{and}
748 game states and it had to go in one section or the other.)
750 \S{backend-dup-game} \cw{dup_game()}
752 \c game_state *(*dup_game)(game_state *state);
754 This function allocates a new \c{game_state} structure and
755 initialises it with an exact copy of the information in the one
756 provided as input. It returns a pointer to the new duplicate.
758 \S{backend-free-game} \cw{free_game()}
760 \c void (*free_game)(game_state *state);
762 This function frees a \c{game_state} structure, and any subsidiary
763 allocations contained within it.
765 \H{backend-ui} Handling \c{game_ui}
767 \S{backend-new-ui} \cw{new_ui()}
769 \c game_ui *(*new_ui)(game_state *state);
771 This function allocates and returns a new \c{game_ui} structure for
772 playing a particular puzzle. It is passed a pointer to the initial
773 \c{game_state}, in case it needs to refer to that when setting up
774 the initial values for the new game.
776 \S{backend-free-ui} \cw{free_ui()}
778 \c void (*free_ui)(game_ui *ui);
780 This function frees a \c{game_ui} structure, and any subsidiary
781 allocations contained within it.
783 \S{backend-encode-ui} \cw{encode_ui()}
785 \c char *(*encode_ui)(game_ui *ui);
787 This function encodes any \e{important} data in a \c{game_ui}
788 structure in string form. It is only called when saving a
789 half-finished game to a file.
791 It should be used sparingly. Almost all data in a \c{game_ui} is not
792 important enough to save. The location of the keyboard-controlled
793 cursor, for example, can be reset to a default position on reloading
794 the game without impacting the user experience. If the user should
795 somehow manage to save a game while a mouse drag was in progress,
796 then discarding that mouse drag would be an outright \e{feature}.
798 A typical thing that \e{would} be worth encoding in this function is
799 the Mines death counter: it's in the \c{game_ui} rather than the
800 \c{game_state} because it's too important to allow the user to
801 revert it by using Undo, and therefore it's also too important to
802 allow the user to revert it by saving and reloading. (Of course, the
803 user could edit the save file by hand... But if the user is \e{that}
804 determined to cheat, they could just as easily modify the game's
807 \S{backend-decode-ui} \cw{decode_ui()}
809 \c void (*decode_ui)(game_ui *ui, char *encoding);
811 This function parses a string previously output by \cw{encode_ui()},
812 and writes the decoded data back into the provided \c{game_ui}
815 \S{backend-changed-state} \cw{changed_state()}
817 \c void (*changed_state)(game_ui *ui, game_state *oldstate,
818 \c game_state *newstate);
820 This function is called by the mid-end whenever the current game
821 state changes, for any reason. Those reasons include:
823 \b a fresh move being made by \cw{interpret_move()} and
826 \b a solve operation being performed by \cw{solve()} and
829 \b the user moving back and forth along the undo list by means of
830 the Undo and Redo operations
832 \b the user selecting Restart to go back to the initial game state.
834 The job of \cw{changed_state()} is to update the \c{game_ui} for
835 consistency with the new game state, if any update is necessary. For
836 example, Same Game stores data about the currently selected tile
837 group in its \c{game_ui}, and this data is intrinsically related to
838 the game state it was derived from. So it's very likely to become
839 invalid when the game state changes; thus, Same Game's
840 \cw{changed_state()} function clears the current selection whenever
843 When \cw{anim_length()} or \cw{flash_length()} are called, you can
844 be sure that there has been a previous call to \cw{changed_state()}.
845 So \cw{changed_state()} can set up data in the \c{game_ui} which will
846 be read by \cw{anim_length()} and \cw{flash_length()}, and those
847 functions will not have to worry about being called without the data
848 having been initialised.
850 \H{backend-moves} Making moves
852 This section describes the functions which actually make moves in
853 the game: that is, the functions which process user input and end up
854 producing new \c{game_state}s.
856 \S{backend-interpret-move} \cw{interpret_move()}
858 \c char *(*interpret_move)(game_state *state, game_ui *ui,
859 \c game_drawstate *ds,
860 \c int x, int y, int button);
862 This function receives user input and processes it. Its input
863 parameters are the current \c{game_state}, the current \c{game_ui}
864 and the current \c{game_drawstate}, plus details of the input event.
865 \c{button} is either an ASCII value or a special code (listed below)
866 indicating an arrow or function key or a mouse event; when
867 \c{button} is a mouse event, \c{x} and \c{y} contain the pixel
868 coordinates of the mouse pointer relative to the top left of the
869 puzzle's drawing area.
871 \cw{interpret_move()} may return in three different ways:
873 \b Returning \cw{NULL} indicates that no action whatsoever occurred
874 in response to the input event; the puzzle was not interested in it
877 \b Returning the empty string (\cw{""}) indicates that the input
878 event has resulted in a change being made to the \c{game_ui} which
879 will require a redraw of the game window, but that no actual
880 \e{move} was made (i.e. no new \c{game_state} needs to be created).
882 \b Returning anything else indicates that a move was made and that a
883 new \c{game_state} must be created. However, instead of actually
884 constructing a new \c{game_state} itself, this function is required
885 to return a string description of the details of the move. This
886 string will be passed to \cw{execute_move()}
887 (\k{backend-execute-move}) to actually create the new
888 \c{game_state}. (Encoding moves as strings in this way means that
889 the mid-end can keep the strings as well as the game states, and the
890 strings can be written to disk when saving the game and fed to
891 \cw{execute_move()} again on reloading.)
893 The return value from \cw{interpret_move()} is expected to be
894 dynamically allocated if and only if it is not either \cw{NULL}
895 \e{or} the empty string.
897 After this function is called, the back end is permitted to rely on
898 some subsequent operations happening in sequence:
900 \b \cw{execute_move()} will be called to convert this move
901 description into a new \c{game_state}
903 \b \cw{changed_state()} will be called with the new \c{game_state}.
905 This means that if \cw{interpret_move()} needs to do updates to the
906 \c{game_ui} which are easier to perform by referring to the new
907 \c{game_state}, it can safely leave them to be done in
908 \cw{changed_state()} and not worry about them failing to happen.
910 (Note, however, that \cw{execute_move()} may \e{also} be called in
911 other circumstances. It is only \cw{interpret_move()} which can rely
912 on a subsequent call to \cw{changed_state()}.)
914 The special key codes supported by this function are:
916 \dt \cw{LEFT_BUTTON}, \cw{MIDDLE_BUTTON}, \cw{RIGHT_BUTTON}
918 \dd Indicate that one of the mouse buttons was pressed down.
920 \dt \cw{LEFT_DRAG}, \cw{MIDDLE_DRAG}, \cw{RIGHT_DRAG}
922 \dd Indicate that the mouse was moved while one of the mouse buttons
923 was still down. The mid-end guarantees that when one of these events
924 is received, it will always have been preceded by a button-down
925 event (and possibly other drag events) for the same mouse button,
926 and no event involving another mouse button will have appeared in
929 \dt \cw{LEFT_RELEASE}, \cw{MIDDLE_RELEASE}, \cw{RIGHT_RELEASE}
931 \dd Indicate that a mouse button was released. The mid-end
932 guarantees that when one of these events is received, it will always
933 have been preceded by a button-down event (and possibly some drag
934 events) for the same mouse button, and no event involving another
935 mouse button will have appeared in between.
937 \dt \cw{CURSOR_UP}, \cw{CURSOR_DOWN}, \cw{CURSOR_LEFT},
940 \dd Indicate that an arrow key was pressed.
942 \dt \cw{CURSOR_SELECT}
944 \dd On platforms which have a prominent \q{select} button alongside
945 their cursor keys, indicates that that button was pressed.
947 In addition, there are some modifiers which can be bitwise-ORed into
948 the \c{button} parameter:
950 \dt \cw{MOD_CTRL}, \cw{MOD_SHFT}
952 \dd These indicate that the Control or Shift key was pressed
953 alongside the key. They only apply to the cursor keys, not to mouse
954 buttons or anything else.
956 \dt \cw{MOD_NUM_KEYPAD}
958 \dd This applies to some ASCII values, and indicates that the key
959 code was input via the numeric keypad rather than the main keyboard.
960 Some puzzles may wish to treat this differently (for example, a
961 puzzle might want to use the numeric keypad as an eight-way
962 directional pad), whereas others might not (a game involving numeric
963 input probably just wants to treat the numeric keypad as numbers).
967 \dd This mask is the bitwise OR of all the available modifiers; you
968 can bitwise-AND with \cw{~MOD_MASK} to strip all the modifiers off
971 \S{backend-execute-move} \cw{execute_move()}
973 \c game_state *(*execute_move)(game_state *state, char *move);
975 This function takes an input \c{game_state} and a move string as
976 output from \cw{interpret_move()}. It returns a newly allocated
977 \c{game_state} which contains the result of applying the specified
978 move to the input game state.
980 This function may return \cw{NULL} if it cannot parse the move
981 string (and this is definitely preferable to crashing or failing an
982 assertion, since one way this can happen is if loading a corrupt
983 save file). However, it must not return \cw{NULL} for any move
984 string that really was output from \cw{interpret_move()}: this is
985 punishable by assertion failure in the mid-end.
987 \S{backend-can-solve} \c{can_solve}
991 This boolean field is set to \cw{TRUE} if the game's \cw{solve()}
992 function does something. If it's set to \cw{FALSE}, the game will
993 not even offer the \q{Solve} menu option.
995 \S{backend-solve} \cw{solve()}
997 \c char *(*solve)(game_state *orig, game_state *curr,
998 \c char *aux, char **error);
1000 This function is called when the user selects the \q{Solve} option
1003 It is passed two input game states: \c{orig} is the game state from
1004 the very start of the puzzle, and \c{curr} is the current one.
1005 (Different games find one or other or both of these convenient.) It
1006 is also passed the \c{aux} string saved by \cw{new_desc()}
1007 (\k{backend-new-desc}), in case that encodes important information
1008 needed to provide the solution.
1010 If this function is unable to produce a solution (perhaps, for
1011 example, the game has no in-built solver so it can only solve
1012 puzzles it invented internally and has an \c{aux} string for) then
1013 it may return \cw{NULL}. If it does this, it must also set
1014 \c{*error} to an error message to be presented to the user (such as
1015 \q{Solution not known for this puzzle}); that error message is not
1016 expected to be dynamically allocated.
1018 If this function \e{does} produce a solution, it returns a move
1019 string suitable for feeding to \cw{execute_move()}
1020 (\k{backend-execute-move}).
1022 \H{backend-drawing} Drawing the game graphics
1024 This section discusses the back end functions that deal with
1027 \S{backend-new-drawstate} \cw{new_drawstate()}
1029 \c game_drawstate *(*new_drawstate)(drawing *dr, game_state *state);
1031 This function allocates and returns a new \c{game_drawstate}
1032 structure for drawing a particular puzzle. It is passed a pointer to
1033 a \c{game_state}, in case it needs to refer to that when setting up
1036 This function may not rely on the puzzle having been newly started;
1037 a new draw state can be constructed at any time if the front end
1038 requests a forced redraw. For games like Pattern, in which initial
1039 game states are much simpler than general ones, this might be
1040 important to keep in mind.
1042 The parameter \c{dr} is a drawing object (see \k{drawing}) which the
1043 function might need to use to allocate blitters. (However, this
1044 isn't recommended; it's usually more sensible to wait to allocate a
1045 blitter until \cw{set_size()} is called, because that way you can
1046 tailor it to the scale at which the puzzle is being drawn.)
1048 \S{backend-free-drawstate} \cw{free_drawstate()}
1050 \c void (*free_drawstate)(drawing *dr, game_drawstate *ds);
1052 This function frees a \c{game_drawstate} structure, and any
1053 subsidiary allocations contained within it.
1055 The parameter \c{dr} is a drawing object (see \k{drawing}), which
1056 might be required if you are freeing a blitter.
1058 \S{backend-preferred-tilesize} \c{preferred_tilesize}
1060 \c int preferred_tilesize;
1062 Each game is required to define a single integer parameter which
1063 expresses, in some sense, the scale at which it is drawn. This is
1064 described in the APIs as \cq{tilesize}, since most puzzles are on a
1065 square (or possibly triangular or hexagonal) grid and hence a
1066 sensible interpretation of this parameter is to define it as the
1067 size of one grid tile in pixels; however, there's no actual
1068 requirement that the \q{tile size} be proportional to the game
1069 window size. Window size is required to increase monotonically with
1070 \q{tile size}, however.
1072 The data element \c{preferred_tilesize} indicates the tile size
1073 which should be used in the absence of a good reason to do otherwise
1074 (such as the screen being too small, or the user explicitly
1075 requesting a resize if that ever gets implemented).
1077 \S{backend-compute-size} \cw{compute_size()}
1079 \c void (*compute_size)(game_params *params, int tilesize,
1082 This function is passed a \c{game_params} structure and a tile size.
1083 It returns, in \c{*x} and \c{*y}, the size in pixels of the drawing
1084 area that would be required to render a puzzle with those parameters
1087 \S{backend-set-size} \cw{set_size()}
1089 \c void (*set_size)(drawing *dr, game_drawstate *ds,
1090 \c game_params *params, int tilesize);
1092 This function is responsible for setting up a \c{game_drawstate} to
1093 draw at a given tile size. Typically this will simply involve
1094 copying the supplied \c{tilesize} parameter into a \c{tilesize}
1095 field inside the draw state; for some more complex games it might
1096 also involve setting up other dimension fields, or possibly
1097 allocating a blitter (see \k{drawing-blitter}).
1099 The parameter \c{dr} is a drawing object (see \k{drawing}), which is
1100 required if a blitter needs to be allocated.
1102 Back ends may assume (and may enforce by assertion) that this
1103 function will be called at most once for any \c{game_drawstate}. If
1104 a puzzle needs to be redrawn at a different size, the mid-end will
1105 create a fresh drawstate.
1107 \S{backend-colours} \cw{colours()}
1109 \c float *(*colours)(frontend *fe, int *ncolours);
1111 This function is responsible for telling the front end what colours
1112 the puzzle will need to draw itself.
1114 It returns the number of colours required in \c{*ncolours}, and the
1115 return value from the function itself is a dynamically allocated
1116 array of three times that many \c{float}s, containing the red, green
1117 and blue components of each colour respectively as numbers in the
1120 The second parameter passed to this function is a front end handle.
1121 The only things it is permitted to do with this handle are to call
1122 the front-end function called \cw{frontend_default_colour()} (see
1123 \k{frontend-default-colour}) or the utility function called
1124 \cw{game_mkhighlight()} (see \k{utils-game-mkhighlight}). (The
1125 latter is a wrapper on the former, so front end implementors only
1126 need to provide \cw{frontend_default_colour()}.) This allows
1127 \cw{colours()} to take local configuration into account when
1128 deciding on its own colour allocations. Most games use the front
1129 end's default colour as their background, apart from a few which
1130 depend on drawing relief highlights so they adjust the background
1131 colour if it's too light for highlights to show up against it.
1133 Note that the colours returned from this function are for
1134 \e{drawing}, not for printing. Printing has an entirely different
1135 colour allocation policy.
1137 \S{backend-anim-length} \cw{anim_length()}
1139 \c float (*anim_length)(game_state *oldstate, game_state *newstate,
1140 \c int dir, game_ui *ui);
1142 This function is called when a move is made, undone or redone. It is
1143 given the old and the new \c{game_state}, and its job is to decide
1144 whether the transition between the two needs to be animated or can
1147 \c{oldstate} is the state that was current until this call;
1148 \c{newstate} is the state that will be current after it. \c{dir}
1149 specifies the chronological order of those states: if it is
1150 positive, then the transition is the result of a move or a redo (and
1151 so \c{newstate} is the later of the two moves), whereas if it is
1152 negative then the transition is the result of an undo (so that
1153 \c{newstate} is the \e{earlier} move).
1155 If this function decides the transition should be animated, it
1156 returns the desired length of the animation in seconds. If not, it
1159 State changes as a result of a Restart operation are never animated;
1160 the mid-end will handle them internally and never consult this
1161 function at all. State changes as a result of Solve operations are
1162 also not animated by default, although you can change this for a
1163 particular game by setting a flag in \c{flags} (\k{backend-flags}).
1165 The function is also passed a pointer to the local \c{game_ui}. It
1166 may refer to information in here to help with its decision (see
1167 \k{writing-conditional-anim} for an example of this), and/or it may
1168 \e{write} information about the nature of the animation which will
1169 be read later by \cw{redraw()}.
1171 When this function is called, it may rely on \cw{changed_state()}
1172 having been called previously, so if \cw{anim_length()} needs to
1173 refer to information in the \c{game_ui}, then \cw{changed_state()}
1174 is a reliable place to have set that information up.
1176 Move animations do not inhibit further input events. If the user
1177 continues playing before a move animation is complete, the animation
1178 will be abandoned and the display will jump straight to the final
1181 \S{backend-flash-length} \cw{flash_length()}
1183 \c float (*flash_length)(game_state *oldstate, game_state *newstate,
1184 \c int dir, game_ui *ui);
1186 This function is called when a move is completed. (\q{Completed}
1187 means that not only has the move been made, but any animation which
1188 accompanied it has finished.) It decides whether the transition from
1189 \c{oldstate} to \c{newstate} merits a \q{flash}.
1191 A flash is much like a move animation, but it is \e{not} interrupted
1192 by further user interface activity; it runs to completion in
1193 parallel with whatever else might be going on on the display. The
1194 only thing which will rush a flash to completion is another flash.
1196 The purpose of flashes is to indicate that the game has been
1197 completed. They were introduced as a separate concept from move
1198 animations because of Net: the habit of most Net players (and
1199 certainly me) is to rotate a tile into place and immediately lock
1200 it, then move on to another tile. When you make your last move, at
1201 the instant the final tile is rotated into place the screen starts
1202 to flash to indicate victory \dash but if you then press the lock
1203 button out of habit, then the move animation is cancelled, and the
1204 victory flash does not complete. (And if you \e{don't} press the
1205 lock button, the completed grid will look untidy because there will
1206 be one unlocked square.) Therefore, I introduced a specific concept
1207 of a \q{flash} which is separate from a move animation and can
1208 proceed in parallel with move animations and any other display
1209 activity, so that the victory flash in Net is not cancelled by that
1212 The input parameters to \cw{flash_length()} are exactly the same as
1213 the ones to \cw{anim_length()}.
1215 Just like \cw{anim_length()}, when this function is called, it may
1216 rely on \cw{changed_state()} having been called previously, so if it
1217 needs to refer to information in the \c{game_ui} then
1218 \cw{changed_state()} is a reliable place to have set that
1221 (Some games use flashes to indicate defeat as well as victory;
1222 Mines, for example, flashes in a different colour when you tread on
1223 a mine from the colour it uses when you complete the game. In order
1224 to achieve this, its \cw{flash_length()} function has to store a
1225 flag in the \c{game_ui} to indicate which flash type is required.)
1227 \S{backend-redraw} \cw{redraw()}
1229 \c void (*redraw)(drawing *dr, game_drawstate *ds,
1230 \c game_state *oldstate, game_state *newstate, int dir,
1231 \c game_ui *ui, float anim_time, float flash_time);
1233 This function is responsible for actually drawing the contents of
1234 the game window, and for redrawing every time the game state or the
1235 \c{game_ui} changes.
1237 The parameter \c{dr} is a drawing object which may be passed to the
1238 drawing API functions (see \k{drawing} for documentation of the
1239 drawing API). This function may not save \c{dr} and use it
1240 elsewhere; it must only use it for calling back to the drawing API
1241 functions within its own lifetime.
1243 \c{ds} is the local \c{game_drawstate}, of course, and \c{ui} is the
1246 \c{newstate} is the semantically-current game state, and is always
1247 non-\cw{NULL}. If \c{oldstate} is also non-\cw{NULL}, it means that
1248 a move has recently been made and the game is still in the process
1249 of displaying an animation linking the old and new states; in this
1250 situation, \c{anim_time} will give the length of time (in seconds)
1251 that the animation has already been running. If \c{oldstate} is
1252 \cw{NULL}, then \c{anim_time} is unused (and will hopefully be set
1253 to zero to avoid confusion).
1255 \c{flash_time}, if it is is non-zero, denotes that the game is in
1256 the middle of a flash, and gives the time since the start of the
1257 flash. See \k{backend-flash-length} for general discussion of
1260 The very first time this function is called for a new
1261 \c{game_drawstate}, it is expected to redraw the \e{entire} drawing
1262 area. Since this often involves drawing visual furniture which is
1263 never subsequently altered, it is often simplest to arrange this by
1264 having a special \q{first time} flag in the draw state, and
1265 resetting it after the first redraw.
1267 When this function (or any subfunction) calls the drawing API, it is
1268 expected to pass colour indices which were previously defined by the
1269 \cw{colours()} function.
1271 \H{backend-printing} Printing functions
1273 This section discusses the back end functions that deal with
1274 printing puzzles out on paper.
1276 \S{backend-can-print} \c{can_print}
1280 This flag is set to \cw{TRUE} if the puzzle is capable of printing
1281 itself on paper. (This makes sense for some puzzles, such as Solo,
1282 which can be filled in with a pencil. Other puzzles, such as
1283 Twiddle, inherently involve moving things around and so would not
1284 make sense to print.)
1286 If this flag is \cw{FALSE}, then the functions \cw{print_size()}
1287 and \cw{print()} will never be called.
1289 \S{backend-can-print-in-colour} \c{can_print_in_colour}
1291 \c int can_print_in_colour;
1293 This flag is set to \cw{TRUE} if the puzzle is capable of printing
1294 itself differently when colour is available. For example, Map can
1295 actually print coloured regions in different \e{colours} rather than
1296 resorting to cross-hatching.
1298 If the \c{can_print} flag is \cw{FALSE}, then this flag will be
1301 \S{backend-print-size} \cw{print_size()}
1303 \c void (*print_size)(game_params *params, float *x, float *y);
1305 This function is passed a \c{game_params} structure and a tile size.
1306 It returns, in \c{*x} and \c{*y}, the preferred size in
1307 \e{millimetres} of that puzzle if it were to be printed out on paper.
1309 If the \c{can_print} flag is \cw{FALSE}, this function will never be
1312 \S{backend-print} \cw{print()}
1314 \c void (*print)(drawing *dr, game_state *state, int tilesize);
1316 This function is called when a puzzle is to be printed out on paper.
1317 It should use the drawing API functions (see \k{drawing}) to print
1320 This function is separate from \cw{redraw()} because it is often
1323 \b The printing function may not depend on pixel accuracy, since
1324 printer resolution is variable. Draw as if your canvas had infinite
1327 \b The printing function sometimes needs to display things in a
1328 completely different style. Net, for example, is very different as
1329 an on-screen puzzle and as a printed one.
1331 \b The printing function is often much simpler since it has no need
1332 to deal with repeated partial redraws.
1334 However, there's no reason the printing and redraw functions can't
1335 share some code if they want to.
1337 When this function (or any subfunction) calls the drawing API, the
1338 colour indices it passes should be colours which have been allocated
1339 by the \cw{print_*_colour()} functions within this execution of
1340 \cw{print()}. This is very different from the fixed small number of
1341 colours used in \cw{redraw()}, because printers do not have a
1342 limitation on the total number of colours that may be used. Some
1343 puzzles' printing functions might wish to allocate only one \q{ink}
1344 colour and use it for all drawing; others might wish to allocate
1345 \e{more} colours than are used on screen.
1347 One possible colour policy worth mentioning specifically is that a
1348 puzzle's printing function might want to allocate the \e{same}
1349 colour indices as are used by the redraw function, so that code
1350 shared between drawing and printing does not have to keep switching
1351 its colour indices. In order to do this, the simplest thing is to
1352 make use of the fact that colour indices returned from
1353 \cw{print_*_colour()} are guaranteed to be in increasing order from
1354 zero. So if you have declared an \c{enum} defining three colours
1355 \cw{COL_BACKGROUND}, \cw{COL_THIS} and \cw{COL_THAT}, you might then
1359 \c c = print_mono_colour(dr, 1); assert(c == COL_BACKGROUND);
1360 \c c = print_mono_colour(dr, 0); assert(c == COL_THIS);
1361 \c c = print_mono_colour(dr, 0); assert(c == COL_THAT);
1363 If the \c{can_print} flag is \cw{FALSE}, this function will never be
1366 \H{backend-misc} Miscellaneous
1368 \S{backend-can-format-as-text} \c{can_format_as_text}
1370 \c int can_format_as_text;
1372 This boolean field is \cw{TRUE} if the game supports formatting a
1373 game state as ASCII text (typically ASCII art) for copying to the
1374 clipboard and pasting into other applications. If it is \cw{FALSE},
1375 front ends will not offer the \q{Copy} command at all.
1377 If this field is \cw{FALSE}, the function \cw{text_format()}
1378 (\k{backend-text-format}) is not expected to do anything at all.
1380 \S{backend-text-format} \cw{text_format()}
1382 \c char *(*text_format)(game_state *state);
1384 This function is passed a \c{game_state}, and returns a newly
1385 allocated C string containing an ASCII representation of that game
1386 state. It is used to implement the \q{Copy} operation in many front
1389 This function should only be called if the back end field
1390 \c{can_format_as_text} (\k{backend-can-format-as-text}) is
1393 The returned string may contain line endings (and will probably want
1394 to), using the normal C internal \cq{\\n} convention. For
1395 consistency between puzzles, all multi-line textual puzzle
1396 representations should \e{end} with a newline as well as containing
1397 them internally. (There are currently no puzzles which have a
1398 one-line ASCII representation, so there's no precedent yet for
1399 whether that should come with a newline or not.)
1401 \S{backend-wants-statusbar} \cw{wants_statusbar()}
1403 \c int wants_statusbar;
1405 This boolean field is set to \cw{TRUE} if the puzzle has a use for a
1406 textual status line (to display score, completion status, currently
1409 \S{backend-is-timed} \c{is_timed}
1413 This boolean field is \cw{TRUE} if the puzzle is time-critical. If
1414 so, the mid-end will maintain a game timer while the user plays.
1416 If this field is \cw{FALSE}, then \cw{timing_state()} will never be
1417 called and need not do anything.
1419 \S{backend-timing-state} \cw{timing_state()}
1421 \c int (*timing_state)(game_state *state, game_ui *ui);
1423 This function is passed the current \c{game_state} and the local
1424 \c{game_ui}; it returns \cw{TRUE} if the game timer should currently
1427 A typical use for the \c{game_ui} in this function is to note when
1428 the game was first completed (by setting a flag in
1429 \cw{changed_state()} \dash see \k{backend-changed-state}), and
1430 freeze the timer thereafter so that the user can undo back through
1431 their solution process without altering their time.
1433 \S{backend-flags} \c{flags}
1437 This field contains miscellaneous per-backend flags. It consists of
1438 the bitwise OR of some combination of the following:
1440 \dt \cw{BUTTON_BEATS(x,y)}
1442 \dd Given any \cw{x} and \cw{y} from the set \{\cw{LEFT_BUTTON},
1443 \cw{MIDDLE_BUTTON}, \cw{RIGHT_BUTTON}\}, this macro evaluates to a
1444 bit flag which indicates that when buttons \cw{x} and \cw{y} are
1445 both pressed simultaneously, the mid-end should consider \cw{x} to
1446 have priority. (In the absence of any such flags, the mid-end will
1447 always consider the most recently pressed button to have priority.)
1449 \dt \cw{SOLVE_ANIMATES}
1451 \dd This flag indicates that moves generated by \cw{solve()}
1452 (\k{backend-solve}) are candidates for animation just like any other
1453 move. For most games, solve moves should not be animated, so the
1454 mid-end doesn't even bother calling \cw{anim_length()}
1455 (\k{backend-anim-length}), thus saving some special-case code in
1456 each game. On the rare occasion that animated solve moves are
1457 actually required, you can set this flag.
1459 \H{backend-initiative} Things a back end may do on its own initiative
1461 This section describes a couple of things that a back end may choose
1462 to do by calling functions elsewhere in the program, which would not
1463 otherwise be obvious.
1465 \S{backend-newrs} Create a random state
1467 If a back end needs random numbers at some point during normal play,
1468 it can create a fresh \c{random_state} by first calling
1469 \c{get_random_seed} (\k{frontend-get-random-seed}) and then passing
1470 the returned seed data to \cw{random_new()}.
1472 This is likely not to be what you want. If a puzzle needs randomness
1473 in the middle of play, it's likely to be more sensible to store some
1474 sort of random state within the \c{game_state}, so that the random
1475 numbers are tied to the particular game state and hence the player
1476 can't simply keep undoing their move until they get numbers they
1479 This facility is currently used only in Net, to implement the
1480 \q{jumble} command, which sets every unlocked tile to a new random
1481 orientation. This randomness \e{is} a reasonable use of the feature,
1482 because it's non-adversarial \dash there's no advantage to the user
1483 in getting different random numbers.
1485 \S{backend-supersede} Supersede its own game description
1487 In response to a move, a back end is (reluctantly) permitted to call
1488 \cw{midend_supersede_game_desc()}:
1490 \c void midend_supersede_game_desc(midend *me,
1491 \c char *desc, char *privdesc);
1493 When the user selects \q{New Game}, the mid-end calls
1494 \cw{new_desc()} (\k{backend-new-desc}) to get a new game
1495 description, and (as well as using that to generate an initial game
1496 state) stores it for the save file and for telling to the user. The
1497 function above overwrites that game description, and also splits it
1498 in two. \c{desc} becomes the new game description which is provided
1499 to the user on request, and is also the one used to construct a new
1500 initial game state if the user selects \q{Restart}. \c{privdesc} is
1501 a \q{private} game description, used to reconstruct the game's
1502 initial state when reloading.
1504 The distinction between the two, as well as the need for this
1505 function at all, comes from Mines. Mines begins with a blank grid
1506 and no idea of where the mines actually are; \cw{new_desc()} does
1507 almost no work in interactive mode, and simply returns a string
1508 encoding the \c{random_state}. When the user first clicks to open a
1509 tile, \e{then} Mines generates the mine positions, in such a way
1510 that the game is soluble from that starting point. Then it uses this
1511 function to supersede the random-state game description with a
1512 proper one. But it needs two: one containing the initial click
1513 location (because that's what you want to happen if you restart the
1514 game, and also what you want to send to a friend so that they play
1515 \e{the same game} as you), and one without the initial click
1516 location (because when you save and reload the game, you expect to
1517 see the same blank initial state as you had before saving).
1519 I should stress again that this function is a horrid hack. Nobody
1520 should use it if they're not Mines; if you think you need to use it,
1521 think again repeatedly in the hope of finding a better way to do
1522 whatever it was you needed to do.
1524 \C{drawing} The drawing API
1526 The back end function \cw{redraw()} (\k{backend-redraw}) is required
1527 to draw the puzzle's graphics on the window's drawing area, or on
1528 paper if the puzzle is printable. To do this portably, it is
1529 provided with a drawing API allowing it to talk directly to the
1530 front end. In this chapter I document that API, both for the benefit
1531 of back end authors trying to use it and for front end authors
1532 trying to implement it.
1534 The drawing API as seen by the back end is a collection of global
1535 functions, each of which takes a pointer to a \c{drawing} structure
1536 (a \q{drawing object}). These objects are supplied as parameters to
1537 the back end's \cw{redraw()} and \cw{print()} functions.
1539 In fact these global functions are not implemented directly by the
1540 front end; instead, they are implemented centrally in \c{drawing.c}
1541 and form a small piece of middleware. The drawing API as supplied by
1542 the front end is a structure containing a set of function pointers,
1543 plus a \cq{void *} handle which is passed to each of those
1544 functions. This enables a single front end to switch between
1545 multiple implementations of the drawing API if necessary. For
1546 example, the Windows API supplies a printing mechanism integrated
1547 into the same GDI which deals with drawing in windows, and therefore
1548 the same API implementation can handle both drawing and printing;
1549 but on Unix, the most common way for applications to print is by
1550 producing PostScript output directly, and although it would be
1551 \e{possible} to write a single (say) \cw{draw_rect()} function which
1552 checked a global flag to decide whether to do GTK drawing operations
1553 or output PostScript to a file, it's much nicer to have two separate
1554 functions and switch between them as appropriate.
1556 When drawing, the puzzle window is indexed by pixel coordinates,
1557 with the top left pixel defined as \cw{(0,0)} and the bottom right
1558 pixel \cw{(w-1,h-1)}, where \c{w} and \c{h} are the width and height
1559 values returned by the back end function \cw{compute_size()}
1560 (\k{backend-compute-size}).
1562 When printing, the puzzle's print area is indexed in exactly the
1563 same way (with an arbitrary tile size provided by the printing
1564 module \c{printing.c}), to facilitate sharing of code between the
1565 drawing and printing routines. However, when printing, puzzles may
1566 no longer assume that the coordinate unit has any relationship to a
1567 pixel; the printer's actual resolution might very well not even be
1568 known at print time, so the coordinate unit might be smaller or
1569 larger than a pixel. Puzzles' print functions should restrict
1570 themselves to drawing geometric shapes rather than fiddly pixel
1573 \e{Puzzles' redraw functions may assume that the surface they draw
1574 on is persistent}. It is the responsibility of every front end to
1575 preserve the puzzle's window contents in the face of GUI window
1576 expose issues and similar. It is not permissible to request that the
1577 back end redraw any part of a window that it has already drawn,
1578 unless something has actually changed as a result of making moves in
1581 Most front ends accomplish this by having the drawing routines draw
1582 on a stored bitmap rather than directly on the window, and copying
1583 the bitmap to the window every time a part of the window needs to be
1584 redrawn. Therefore, it is vitally important that whenever the back
1585 end does any drawing it informs the front end of which parts of the
1586 window it has accessed, and hence which parts need repainting. This
1587 is done by calling \cw{draw_update()} (\k{drawing-draw-update}).
1589 In the following sections I first discuss the drawing API as seen by
1590 the back end, and then the \e{almost} identical function-pointer
1591 form seen by the front end.
1593 \H{drawing-backend} Drawing API as seen by the back end
1595 This section documents the back-end drawing API, in the form of
1596 functions which take a \c{drawing} object as an argument.
1598 \S{drawing-draw-rect} \cw{draw_rect()}
1600 \c void draw_rect(drawing *dr, int x, int y, int w, int h,
1603 Draws a filled rectangle in the puzzle window.
1605 \c{x} and \c{y} give the coordinates of the top left pixel of the
1606 rectangle. \c{w} and \c{h} give its width and height. Thus, the
1607 horizontal extent of the rectangle runs from \c{x} to \c{x+w-1}
1608 inclusive, and the vertical extent from \c{y} to \c{y+h-1}
1611 \c{colour} is an integer index into the colours array returned by
1612 the back end function \cw{colours()} (\k{backend-colours}).
1614 There is no separate pixel-plotting function. If you want to plot a
1615 single pixel, the approved method is to use \cw{draw_rect()} with
1616 width and height set to 1.
1618 Unlike many of the other drawing functions, this function is
1619 guaranteed to be pixel-perfect: the rectangle will be sharply
1620 defined and not anti-aliased or anything like that.
1622 This function may be used for both drawing and printing.
1624 \S{drawing-draw-rect-outline} \cw{draw_rect_outline()}
1626 \c void draw_rect_outline(drawing *dr, int x, int y, int w, int h,
1629 Draws an outline rectangle in the puzzle window.
1631 \c{x} and \c{y} give the coordinates of the top left pixel of the
1632 rectangle. \c{w} and \c{h} give its width and height. Thus, the
1633 horizontal extent of the rectangle runs from \c{x} to \c{x+w-1}
1634 inclusive, and the vertical extent from \c{y} to \c{y+h-1}
1637 \c{colour} is an integer index into the colours array returned by
1638 the back end function \cw{colours()} (\k{backend-colours}).
1640 From a back end perspective, this function may be considered to be
1641 part of the drawing API. However, front ends are not required to
1642 implement it, since it is actually implemented centrally (in
1643 \cw{misc.c}) as a wrapper on \cw{draw_polygon()}.
1645 This function may be used for both drawing and printing.
1647 \S{drawing-draw-line} \cw{draw_line()}
1649 \c void draw_line(drawing *dr, int x1, int y1, int x2, int y2,
1652 Draws a straight line in the puzzle window.
1654 \c{x1} and \c{y1} give the coordinates of one end of the line.
1655 \c{x2} and \c{y2} give the coordinates of the other end. The line
1656 drawn includes both those points.
1658 \c{colour} is an integer index into the colours array returned by
1659 the back end function \cw{colours()} (\k{backend-colours}).
1661 Some platforms may perform anti-aliasing on this function.
1662 Therefore, do not assume that you can erase a line by drawing the
1663 same line over it in the background colour; anti-aliasing might
1664 lead to perceptible ghost artefacts around the vanished line.
1666 This function may be used for both drawing and printing.
1668 \S{drawing-draw-polygon} \cw{draw_polygon()}
1670 \c void draw_polygon(drawing *dr, int *coords, int npoints,
1671 \c int fillcolour, int outlinecolour);
1673 Draws an outlined or filled polygon in the puzzle window.
1675 \c{coords} is an array of \cw{(2*npoints)} integers, containing the
1676 \c{x} and \c{y} coordinates of \c{npoints} vertices.
1678 \c{fillcolour} and \c{outlinecolour} are integer indices into the
1679 colours array returned by the back end function \cw{colours()}
1680 (\k{backend-colours}). \c{fillcolour} may also be \cw{-1} to
1681 indicate that the polygon should be outlined only.
1683 The polygon defined by the specified list of vertices is first
1684 filled in \c{fillcolour}, if specified, and then outlined in
1687 \c{outlinecolour} may \e{not} be \cw{-1}; it must be a valid colour
1688 (and front ends are permitted to enforce this by assertion). This is
1689 because different platforms disagree on whether a filled polygon
1690 should include its boundary line or not, so drawing \e{only} a
1691 filled polygon would have non-portable effects. If you want your
1692 filled polygon not to have a visible outline, you must set
1693 \c{outlinecolour} to the same as \c{fillcolour}.
1695 Some platforms may perform anti-aliasing on this function.
1696 Therefore, do not assume that you can erase a polygon by drawing the
1697 same polygon over it in the background colour. Also, be prepared for
1698 the polygon to extend a pixel beyond its obvious bounding box as a
1699 result of this; if you really need it not to do this to avoid
1700 interfering with other delicate graphics, you should probably use
1701 \cw{clip()} (\k{drawing-clip}).
1703 This function may be used for both drawing and printing.
1705 \S{drawing-draw-circle} \cw{draw_circle()}
1707 \c void draw_circle(drawing *dr, int cx, int cy, int radius,
1708 \c int fillcolour, int outlinecolour);
1710 Draws an outlined or filled circle in the puzzle window.
1712 \c{cx} and \c{cy} give the coordinates of the centre of the circle.
1713 \c{radius} gives its radius. The total horizontal pixel extent of
1714 the circle is from \c{cx-radius+1} to \c{cx+radius-1} inclusive, and
1715 the vertical extent similarly around \c{cy}.
1717 \c{fillcolour} and \c{outlinecolour} are integer indices into the
1718 colours array returned by the back end function \cw{colours()}
1719 (\k{backend-colours}). \c{fillcolour} may also be \cw{-1} to
1720 indicate that the circle should be outlined only.
1722 The circle is first filled in \c{fillcolour}, if specified, and then
1723 outlined in \c{outlinecolour}.
1725 \c{outlinecolour} may \e{not} be \cw{-1}; it must be a valid colour
1726 (and front ends are permitted to enforce this by assertion). This is
1727 because different platforms disagree on whether a filled circle
1728 should include its boundary line or not, so drawing \e{only} a
1729 filled circle would have non-portable effects. If you want your
1730 filled circle not to have a visible outline, you must set
1731 \c{outlinecolour} to the same as \c{fillcolour}.
1733 Some platforms may perform anti-aliasing on this function.
1734 Therefore, do not assume that you can erase a circle by drawing the
1735 same circle over it in the background colour. Also, be prepared for
1736 the circle to extend a pixel beyond its obvious bounding box as a
1737 result of this; if you really need it not to do this to avoid
1738 interfering with other delicate graphics, you should probably use
1739 \cw{clip()} (\k{drawing-clip}).
1741 This function may be used for both drawing and printing.
1743 \S{drawing-draw-text} \cw{draw_text()}
1745 \c void draw_text(drawing *dr, int x, int y, int fonttype,
1746 \c int fontsize, int align, int colour, char *text);
1748 Draws text in the puzzle window.
1750 \c{x} and \c{y} give the coordinates of a point. The relation of
1751 this point to the location of the text is specified by \c{align},
1752 which is a bitwise OR of horizontal and vertical alignment flags:
1754 \dt \cw{ALIGN_VNORMAL}
1756 \dd Indicates that \c{y} is aligned with the baseline of the text.
1758 \dt \cw{ALIGN_VCENTRE}
1760 \dd Indicates that \c{y} is aligned with the vertical centre of the
1761 text. (In fact, it's aligned with the vertical centre of normal
1762 \e{capitalised} text: displaying two pieces of text with
1763 \cw{ALIGN_VCENTRE} at the same \cw{y}-coordinate will cause their
1764 baselines to be aligned with one another, even if one is an ascender
1765 and the other a descender.)
1767 \dt \cw{ALIGN_HLEFT}
1769 \dd Indicates that \c{x} is aligned with the left-hand end of the
1772 \dt \cw{ALIGN_HCENTRE}
1774 \dd Indicates that \c{x} is aligned with the horizontal centre of
1777 \dt \cw{ALIGN_HRIGHT}
1779 \dd Indicates that \c{x} is aligned with the right-hand end of the
1782 \c{fonttype} is either \cw{FONT_FIXED} or \cw{FONT_VARIABLE}, for a
1783 monospaced or proportional font respectively. (No more detail than
1784 that may be specified; it would only lead to portability issues
1785 between different platforms.)
1787 \c{fontsize} is the desired size, in pixels, of the text. This size
1788 corresponds to the overall point size of the text, not to any
1789 internal dimension such as the cap-height.
1791 \c{colour} is an integer index into the colours array returned by
1792 the back end function \cw{colours()} (\k{backend-colours}).
1794 This function may be used for both drawing and printing.
1796 \S{drawing-clip} \cw{clip()}
1798 \c void clip(drawing *dr, int x, int y, int w, int h);
1800 Establishes a clipping rectangle in the puzzle window.
1802 \c{x} and \c{y} give the coordinates of the top left pixel of the
1803 clipping rectangle. \c{w} and \c{h} give its width and height. Thus,
1804 the horizontal extent of the rectangle runs from \c{x} to \c{x+w-1}
1805 inclusive, and the vertical extent from \c{y} to \c{y+h-1}
1806 inclusive. (These are exactly the same semantics as
1809 After this call, no drawing operation will affect anything outside
1810 the specified rectangle. The effect can be reversed by calling
1811 \cw{unclip()} (\k{drawing-unclip}).
1813 Back ends should not assume that a clipping rectangle will be
1814 automatically cleared up by the front end if it's left lying around;
1815 that might work on current front ends, but shouldn't be relied upon.
1816 Always explicitly call \cw{unclip()}.
1818 This function may be used for both drawing and printing.
1820 \S{drawing-unclip} \cw{unclip()}
1822 \c void unclip(drawing *dr);
1824 Reverts the effect of a previous call to \cw{clip()}. After this
1825 call, all drawing operations will be able to affect the entire
1826 puzzle window again.
1828 This function may be used for both drawing and printing.
1830 \S{drawing-draw-update} \cw{draw_update()}
1832 \c void draw_update(drawing *dr, int x, int y, int w, int h);
1834 Informs the front end that a rectangular portion of the puzzle
1835 window has been drawn on and needs to be updated.
1837 \c{x} and \c{y} give the coordinates of the top left pixel of the
1838 update rectangle. \c{w} and \c{h} give its width and height. Thus,
1839 the horizontal extent of the rectangle runs from \c{x} to \c{x+w-1}
1840 inclusive, and the vertical extent from \c{y} to \c{y+h-1}
1841 inclusive. (These are exactly the same semantics as
1844 The back end redraw function \e{must} call this function to report
1845 any changes it has made to the window. Otherwise, those changes may
1846 not become immediately visible, and may then appear at an
1847 unpredictable subsequent time such as the next time the window is
1848 covered and re-exposed.
1850 This function is only important when drawing. It may be called when
1851 printing as well, but doing so is not compulsory, and has no effect.
1852 (So if you have a shared piece of code between the drawing and
1853 printing routines, that code may safely call \cw{draw_update()}.)
1855 \S{drawing-status-bar} \cw{status_bar()}
1857 \c void status_bar(drawing *dr, char *text);
1859 Sets the text in the game's status bar to \c{text}. The text is copied
1860 from the supplied buffer, so the caller is free to deallocate or
1861 modify the buffer after use.
1863 (This function is not exactly a \e{drawing} function, but it shares
1864 with the drawing API the property that it may only be called from
1865 within the back end redraw function, so this is as good a place as
1866 any to document it.)
1868 The supplied text is filtered through the mid-end for optional
1869 rewriting before being passed on to the front end; the mid-end will
1870 prepend the current game time if the game is timed (and may in
1871 future perform other rewriting if it seems like a good idea).
1873 This function is for drawing only; it must never be called during
1876 \S{drawing-blitter} Blitter functions
1878 This section describes a group of related functions which save and
1879 restore a section of the puzzle window. This is most commonly used
1880 to implement user interfaces involving dragging a puzzle element
1881 around the window: at the end of each call to \cw{redraw()}, if an
1882 object is currently being dragged, the back end saves the window
1883 contents under that location and then draws the dragged object, and
1884 at the start of the next \cw{redraw()} the first thing it does is to
1885 restore the background.
1887 The front end defines an opaque type called a \c{blitter}, which is
1888 capable of storing a rectangular area of a specified size.
1890 Blitter functions are for drawing only; they must never be called
1893 \S2{drawing-blitter-new} \cw{blitter_new()}
1895 \c blitter *blitter_new(drawing *dr, int w, int h);
1897 Creates a new blitter object which stores a rectangle of size \c{w}
1898 by \c{h} pixels. Returns a pointer to the blitter object.
1900 Blitter objects are best stored in the \c{game_drawstate}. A good
1901 time to create them is in the \cw{set_size()} function
1902 (\k{backend-set-size}), since it is at this point that you first
1903 know how big a rectangle they will need to save.
1905 \S2{drawing-blitter-free} \cw{blitter_free()}
1907 \c void blitter_free(drawing *dr, blitter *bl);
1909 Disposes of a blitter object. Best called in \cw{free_drawstate()}.
1910 (However, check that the blitter object is not \cw{NULL} before
1911 attempting to free it; it is possible that a draw state might be
1912 created and freed without ever having \cw{set_size()} called on it
1915 \S2{drawing-blitter-save} \cw{blitter_save()}
1917 \c void blitter_save(drawing *dr, blitter *bl, int x, int y);
1919 This is a true drawing API function, in that it may only be called
1920 from within the game redraw routine. It saves a rectangular portion
1921 of the puzzle window into the specified blitter object.
1923 \c{x} and \c{y} give the coordinates of the top left corner of the
1924 saved rectangle. The rectangle's width and height are the ones
1925 specified when the blitter object was created.
1927 This function is required to cope and do the right thing if \c{x}
1928 and \c{y} are out of range. (The right thing probably means saving
1929 whatever part of the blitter rectangle overlaps with the visible
1930 area of the puzzle window.)
1932 \S2{drawing-blitter-load} \cw{blitter_load()}
1934 \c void blitter_load(drawing *dr, blitter *bl, int x, int y);
1936 This is a true drawing API function, in that it may only be called
1937 from within the game redraw routine. It restores a rectangular
1938 portion of the puzzle window from the specified blitter object.
1940 \c{x} and \c{y} give the coordinates of the top left corner of the
1941 rectangle to be restored. The rectangle's width and height are the
1942 ones specified when the blitter object was created.
1944 Alternatively, you can specify both \c{x} and \c{y} as the special
1945 value \cw{BLITTER_FROMSAVED}, in which case the rectangle will be
1946 restored to exactly where it was saved from. (This is probably what
1947 you want to do almost all the time, if you're using blitters to
1948 implement draggable puzzle elements.)
1950 This function is required to cope and do the right thing if \c{x}
1951 and \c{y} (or the equivalent ones saved in the blitter) are out of
1952 range. (The right thing probably means restoring whatever part of
1953 the blitter rectangle overlaps with the visible area of the puzzle
1956 If this function is called on a blitter which had previously been
1957 saved from a partially out-of-range rectangle, then the parts of the
1958 saved bitmap which were not visible at save time are undefined. If
1959 the blitter is restored to a different position so as to make those
1960 parts visible, the effect on the drawing area is undefined.
1962 \S{print-mono-colour} \cw{print_mono_colour()}
1964 \c int print_mono_colour(drawing *dr, int grey);
1966 This function allocates a colour index for a simple monochrome
1967 colour during printing.
1969 \c{grey} must be 0 or 1. If \c{grey} is 0, the colour returned is
1970 black; if \c{grey} is 1, the colour is white.
1972 \S{print-grey-colour} \cw{print_grey_colour()}
1974 \c int print_grey_colour(drawing *dr, int hatch, float grey);
1976 This function allocates a colour index for a grey-scale colour
1979 \c{grey} may be any number between 0 (black) and 1 (white); for
1980 example, 0.5 indicates a medium grey.
1982 If printing in black and white only, the \c{grey} value will not be
1983 used; instead, regions shaded in this colour will be hatched with
1984 parallel lines. The \c{hatch} parameter defines what type of
1985 hatching should be used in place of this colour:
1987 \dt \cw{HATCH_SOLID}
1989 \dd In black and white, this colour will be replaced by solid black.
1991 \dt \cw{HATCH_CLEAR}
1993 \dd In black and white, this colour will be replaced by solid white.
1995 \dt \cw{HATCH_SLASH}
1997 \dd This colour will be hatched by lines slanting to the right at 45
2000 \dt \cw{HATCH_BACKSLASH}
2002 \dd This colour will be hatched by lines slanting to the left at 45
2005 \dt \cw{HATCH_HORIZ}
2007 \dd This colour will be hatched by horizontal lines.
2011 \dd This colour will be hatched by vertical lines.
2015 \dd This colour will be hatched by criss-crossing horizontal and
2020 \dd This colour will be hatched by criss-crossing diagonal lines.
2022 Colours defined to use hatching may not be used for drawing lines;
2023 they may only be used for filling areas. That is, they may be used
2024 as the \c{fillcolour} parameter to \cw{draw_circle()} and
2025 \cw{draw_polygon()}, and as the colour parameter to
2026 \cw{draw_rect()}, but may not be used as the \c{outlinecolour}
2027 parameter to \cw{draw_circle()} or \cw{draw_polygon()}, or with
2030 \S{print-rgb-colour} \cw{print_rgb_colour()}
2032 \c int print_rgb_colour(drawing *dr, int hatch,
2033 \c float r, float g, float b);
2035 This function allocates a colour index for a fully specified RGB
2036 colour during printing.
2038 \c{r}, \c{g} and \c{b} may each be anywhere in the range from 0 to 1.
2040 If printing in black and white only, these values will not be used;
2041 instead, regions shaded in this colour will be hatched with parallel
2042 lines. The \c{hatch} parameter defines what type of hatching should
2043 be used in place of this colour; see \k{print-grey-colour} for its
2046 \S{print-line-width} \cw{print_line_width()}
2048 \c void print_line_width(drawing *dr, int width);
2050 This function is called to set the thickness of lines drawn during
2051 printing. It is meaningless in drawing: all lines drawn by
2052 \cw{draw_line()}, \cw{draw_circle} and \cw{draw_polygon()} are one
2053 pixel in thickness. However, in printing there is no clear
2054 definition of a pixel and so line widths must be explicitly
2057 The line width is specified in the usual coordinate system. Note,
2058 however, that it is a hint only: the central printing system may
2059 choose to vary line thicknesses at user request or due to printer
2062 \H{drawing-frontend} The drawing API as implemented by the front end
2064 This section describes the drawing API in the function-pointer form
2065 in which it is implemented by a front end.
2067 (It isn't only platform-specific front ends which implement this
2068 API; the platform-independent module \c{ps.c} also provides an
2069 implementation of it which outputs PostScript. Thus, any platform
2070 which wants to do PS printing can do so with minimum fuss.)
2072 The following entries all describe function pointer fields in a
2073 structure called \c{drawing_api}. Each of the functions takes a
2074 \cq{void *} context pointer, which it should internally cast back to
2075 a more useful type. Thus, a drawing \e{object} (\c{drawing *)}
2076 suitable for passing to the back end redraw or printing functions
2077 is constructed by passing a \c{drawing_api} and a \cq{void *} to the
2078 function \cw{drawing_new()} (see \k{drawing-new}).
2080 \S{drawingapi-draw-text} \cw{draw_text()}
2082 \c void (*draw_text)(void *handle, int x, int y, int fonttype,
2083 \c int fontsize, int align, int colour, char *text);
2085 This function behaves exactly like the back end \cw{draw_text()}
2086 function; see \k{drawing-draw-text}.
2088 \S{drawingapi-draw-rect} \cw{draw_rect()}
2090 \c void (*draw_rect)(void *handle, int x, int y, int w, int h,
2093 This function behaves exactly like the back end \cw{draw_rect()}
2094 function; see \k{drawing-draw-rect}.
2096 \S{drawingapi-draw-line} \cw{draw_line()}
2098 \c void (*draw_line)(void *handle, int x1, int y1, int x2, int y2,
2101 This function behaves exactly like the back end \cw{draw_line()}
2102 function; see \k{drawing-draw-line}.
2104 \S{drawingapi-draw-polygon} \cw{draw_polygon()}
2106 \c void (*draw_polygon)(void *handle, int *coords, int npoints,
2107 \c int fillcolour, int outlinecolour);
2109 This function behaves exactly like the back end \cw{draw_polygon()}
2110 function; see \k{drawing-draw-polygon}.
2112 \S{drawingapi-draw-circle} \cw{draw_circle()}
2114 \c void (*draw_circle)(void *handle, int cx, int cy, int radius,
2115 \c int fillcolour, int outlinecolour);
2117 This function behaves exactly like the back end \cw{draw_circle()}
2118 function; see \k{drawing-draw-circle}.
2120 \S{drawingapi-draw-update} \cw{draw_update()}
2122 \c void (*draw_update)(void *handle, int x, int y, int w, int h);
2124 This function behaves exactly like the back end \cw{draw_text()}
2125 function; see \k{drawing-draw-text}.
2127 An implementation of this API which only supports printing is
2128 permitted to define this function pointer to be \cw{NULL} rather
2129 than bothering to define an empty function. The middleware in
2130 \cw{drawing.c} will notice and avoid calling it.
2132 \S{drawingapi-clip} \cw{clip()}
2134 \c void (*clip)(void *handle, int x, int y, int w, int h);
2136 This function behaves exactly like the back end \cw{clip()}
2137 function; see \k{drawing-clip}.
2139 \S{drawingapi-unclip} \cw{unclip()}
2141 \c void (*unclip)(void *handle);
2143 This function behaves exactly like the back end \cw{unclip()}
2144 function; see \k{drawing-unclip}.
2146 \S{drawingapi-start-draw} \cw{start_draw()}
2148 \c void (*start_draw)(void *handle);
2150 This function is called at the start of drawing. It allows the front
2151 end to initialise any temporary data required to draw with, such as
2154 Implementations of this API which do not provide drawing services
2155 may define this function pointer to be \cw{NULL}; it will never be
2156 called unless drawing is attempted.
2158 \S{drawingapi-end-draw} \cw{end_draw()}
2160 \c void (*end_draw)(void *handle);
2162 This function is called at the end of drawing. It allows the front
2163 end to do cleanup tasks such as deallocating device contexts and
2164 scheduling appropriate GUI redraw events.
2166 Implementations of this API which do not provide drawing services
2167 may define this function pointer to be \cw{NULL}; it will never be
2168 called unless drawing is attempted.
2170 \S{drawingapi-status-bar} \cw{status_bar()}
2172 \c void (*status_bar)(void *handle, char *text);
2174 This function behaves exactly like the back end \cw{status_bar()}
2175 function; see \k{drawing-status-bar}.
2177 Front ends implementing this function need not worry about it being
2178 called repeatedly with the same text; the middleware code in
2179 \cw{status_bar()} will take care of this.
2181 Implementations of this API which do not provide drawing services
2182 may define this function pointer to be \cw{NULL}; it will never be
2183 called unless drawing is attempted.
2185 \S{drawingapi-blitter-new} \cw{blitter_new()}
2187 \c blitter *(*blitter_new)(void *handle, int w, int h);
2189 This function behaves exactly like the back end \cw{blitter_new()}
2190 function; see \k{drawing-blitter-new}.
2192 Implementations of this API which do not provide drawing services
2193 may define this function pointer to be \cw{NULL}; it will never be
2194 called unless drawing is attempted.
2196 \S{drawingapi-blitter-free} \cw{blitter_free()}
2198 \c void (*blitter_free)(void *handle, blitter *bl);
2200 This function behaves exactly like the back end \cw{blitter_free()}
2201 function; see \k{drawing-blitter-free}.
2203 Implementations of this API which do not provide drawing services
2204 may define this function pointer to be \cw{NULL}; it will never be
2205 called unless drawing is attempted.
2207 \S{drawingapi-blitter-save} \cw{blitter_save()}
2209 \c void (*blitter_save)(void *handle, blitter *bl, int x, int y);
2211 This function behaves exactly like the back end \cw{blitter_save()}
2212 function; see \k{drawing-blitter-save}.
2214 Implementations of this API which do not provide drawing services
2215 may define this function pointer to be \cw{NULL}; it will never be
2216 called unless drawing is attempted.
2218 \S{drawingapi-blitter-load} \cw{blitter_load()}
2220 \c void (*blitter_load)(void *handle, blitter *bl, int x, int y);
2222 This function behaves exactly like the back end \cw{blitter_load()}
2223 function; see \k{drawing-blitter-load}.
2225 Implementations of this API which do not provide drawing services
2226 may define this function pointer to be \cw{NULL}; it will never be
2227 called unless drawing is attempted.
2229 \S{drawingapi-begin-doc} \cw{begin_doc()}
2231 \c void (*begin_doc)(void *handle, int pages);
2233 This function is called at the beginning of a printing run. It gives
2234 the front end an opportunity to initialise any required printing
2235 subsystem. It also provides the number of pages in advance.
2237 Implementations of this API which do not provide printing services
2238 may define this function pointer to be \cw{NULL}; it will never be
2239 called unless printing is attempted.
2241 \S{drawingapi-begin-page} \cw{begin_page()}
2243 \c void (*begin_page)(void *handle, int number);
2245 This function is called during printing, at the beginning of each
2246 page. It gives the page number (numbered from 1 rather than 0, so
2247 suitable for use in user-visible contexts).
2249 Implementations of this API which do not provide printing services
2250 may define this function pointer to be \cw{NULL}; it will never be
2251 called unless printing is attempted.
2253 \S{drawingapi-begin-puzzle} \cw{begin_puzzle()}
2255 \c void (*begin_puzzle)(void *handle, float xm, float xc,
2256 \c float ym, float yc, int pw, int ph, float wmm);
2258 This function is called during printing, just before printing a
2259 single puzzle on a page. It specifies the size and location of the
2262 \c{xm} and \c{xc} specify the horizontal position of the puzzle on
2263 the page, as a linear function of the page width. The front end is
2264 expected to multiply the page width by \c{xm}, add \c{xc} (measured
2265 in millimetres), and use the resulting x-coordinate as the left edge
2268 Similarly, \c{ym} and \c{yc} specify the vertical position of the
2269 puzzle as a function of the page height: the page height times
2270 \c{xm}, plus \c{xc} millimetres, equals the desired distance from
2271 the top of the page to the top of the puzzle.
2273 (This unwieldy mechanism is required because not all printing
2274 systems can communicate the page size back to the software. The
2275 PostScript back end, for example, writes out PS which determines the
2276 page size at print time by means of calling \cq{clippath}, and
2277 centres the puzzles within that. Thus, exactly the same PS file
2278 works on A4 or on US Letter paper without needing local
2279 configuration, which simplifies matters.)
2281 \cw{pw} and \cw{ph} give the size of the puzzle in drawing API
2282 coordinates. The printing system will subsequently call the puzzle's
2283 own print function, which will in turn call drawing API functions in
2284 the expectation that an area \cw{pw} by \cw{ph} units is available
2285 to draw the puzzle on.
2287 Finally, \cw{wmm} gives the desired width of the puzzle in
2288 millimetres. (The aspect ratio is expected to be preserved, so if
2289 the desired puzzle height is also needed then it can be computed as
2292 Implementations of this API which do not provide printing services
2293 may define this function pointer to be \cw{NULL}; it will never be
2294 called unless printing is attempted.
2296 \S{drawingapi-end-puzzle} \cw{end_puzzle()}
2298 \c void (*end_puzzle)(void *handle);
2300 This function is called after the printing of a specific puzzle is
2303 Implementations of this API which do not provide printing services
2304 may define this function pointer to be \cw{NULL}; it will never be
2305 called unless printing is attempted.
2307 \S{drawingapi-end-page} \cw{end_page()}
2309 \c void (*end_page)(void *handle, int number);
2311 This function is called after the printing of a page is finished.
2313 Implementations of this API which do not provide printing services
2314 may define this function pointer to be \cw{NULL}; it will never be
2315 called unless printing is attempted.
2317 \S{drawingapi-end-doc} \cw{end_doc()}
2319 \c void (*end_doc)(void *handle);
2321 This function is called after the printing of the entire document is
2322 finished. This is the moment to close files, send things to the
2323 print spooler, or whatever the local convention is.
2325 Implementations of this API which do not provide printing services
2326 may define this function pointer to be \cw{NULL}; it will never be
2327 called unless printing is attempted.
2329 \S{drawingapi-line-width} \cw{line_width()}
2331 \c void (*line_width)(void *handle, float width);
2333 This function is called to set the line thickness, during printing
2334 only. Note that the width is a \cw{float} here, where it was an
2335 \cw{int} as seen by the back end. This is because \cw{drawing.c} may
2336 have scaled it on the way past.
2338 However, the width is still specified in the same coordinate system
2339 as the rest of the drawing.
2341 Implementations of this API which do not provide printing services
2342 may define this function pointer to be \cw{NULL}; it will never be
2343 called unless printing is attempted.
2345 \H{drawingapi-frontend} The drawing API as called by the front end
2347 There are a small number of functions provided in \cw{drawing.c}
2348 which the front end needs to \e{call}, rather than helping to
2349 implement. They are described in this section.
2351 \S{drawing-new} \cw{drawing_new()}
2353 \c drawing *drawing_new(const drawing_api *api, midend *me,
2356 This function creates a drawing object. It is passed a
2357 \c{drawing_api}, which is a structure containing nothing but
2358 function pointers; and also a \cq{void *} handle. The handle is
2359 passed back to each function pointer when it is called.
2361 The \c{midend} parameter is used for rewriting the status bar
2362 contents: \cw{status_bar()} (see \k{drawing-status-bar}) has to call
2363 a function in the mid-end which might rewrite the status bar text.
2364 If the drawing object is to be used only for printing, or if the
2365 game is known not to call \cw{status_bar()}, this parameter may be
2368 \S{drawing-free} \cw{drawing_free()}
2370 \c void drawing_free(drawing *dr);
2372 This function frees a drawing object. Note that the \cq{void *}
2373 handle is not freed; if that needs cleaning up it must be done by
2376 \S{drawing-print-get-colour} \cw{print_get_colour()}
2378 \c void print_get_colour(drawing *dr, int colour, int *hatch,
2379 \c float *r, float *g, float *b)
2381 This function is called by the implementations of the drawing API
2382 functions when they are called in a printing context. It takes a
2383 colour index as input, and returns the description of the colour as
2384 requested by the back end.
2386 \c{*r}, \c{*g} and \c{*b} are filled with the RGB values of the
2387 desired colour if printing in colour.
2389 \c{*hatch} is filled with the type of hatching (or not) desired if
2390 printing in black and white. See \k{print-grey-colour} for details
2391 of the values this integer can take.
2393 \C{midend} The API provided by the mid-end
2395 This chapter documents the API provided by the mid-end to be called
2396 by the front end. You probably only need to read this if you are a
2397 front end implementor, i.e. you are porting Puzzles to a new
2398 platform. If you're only interested in writing new puzzles, you can
2399 safely skip this chapter.
2401 All the persistent state in the mid-end is encapsulated within a
2402 \c{midend} structure, to facilitate having multiple mid-ends in any
2403 port which supports multiple puzzle windows open simultaneously.
2404 Each \c{midend} is intended to handle the contents of a single
2407 \H{midend-new} \cw{midend_new()}
2409 \c midend *midend_new(frontend *fe, const game *ourgame,
2410 \c const drawing_api *drapi, void *drhandle)
2412 Allocates and returns a new mid-end structure.
2414 The \c{fe} argument is stored in the mid-end. It will be used when
2415 calling back to functions such as \cw{activate_timer()}
2416 (\k{frontend-activate-timer}), and will be passed on to the back end
2417 function \cw{colours()} (\k{backend-colours}).
2419 The parameters \c{drapi} and \c{drhandle} are passed to
2420 \cw{drawing_new()} (\k{drawing-new}) to construct a drawing object
2421 which will be passed to the back end function \cw{redraw()}
2422 (\k{backend-redraw}). Hence, all drawing-related function pointers
2423 defined in \c{drapi} can expect to be called with \c{drhandle} as
2424 their first argument.
2426 The \c{ourgame} argument points to a container structure describing
2427 a game back end. The mid-end thus created will only be capable of
2428 handling that one game. (So even in a monolithic front end
2429 containing all the games, this imposes the constraint that any
2430 individual puzzle window is tied to a single game. Unless, of
2431 course, you feel brave enough to change the mid-end for the window
2432 without closing the window...)
2434 \H{midend-free} \cw{midend_free()}
2436 \c void midend_free(midend *me);
2438 Frees a mid-end structure and all its associated data.
2440 \H{midend-set-params} \cw{midend_set_params()}
2442 \c void midend_set_params(midend *me, game_params *params);
2444 Sets the current game parameters for a mid-end. Subsequent games
2445 generated by \cw{midend_new_game()} (\k{midend-new-game}) will use
2446 these parameters until further notice.
2448 The usual way in which the front end will have an actual
2449 \c{game_params} structure to pass to this function is if it had
2450 previously got it from \cw{midend_fetch_preset()}
2451 (\k{midend-fetch-preset}). Thus, this function is usually called in
2452 response to the user making a selection from the presets menu.
2454 \H{midend-get-params} \cw{midend_get_params()}
2456 \c game_params *midend_get_params(midend *me);
2458 Returns the current game parameters stored in this mid-end.
2460 The returned value is dynamically allocated, and should be freed
2461 when finished with by passing it to the game's own
2462 \cw{free_params()} function (see \k{backend-free-params}).
2464 \H{midend-size} \cw{midend_size()}
2466 \c void midend_size(midend *me, int *x, int *y, int expand);
2468 Tells the mid-end to figure out its window size.
2470 On input, \c{*x} and \c{*y} should contain the maximum or requested
2471 size for the window. (Typically this will be the size of the screen
2472 that the window has to fit on, or similar.) The mid-end will
2473 repeatedly call the back end function \cw{compute_size()}
2474 (\k{backend-compute-size}), searching for a tile size that best
2475 satisfies the requirements. On exit, \c{*x} and \c{*y} will contain
2476 the size needed for the puzzle window's drawing area. (It is of
2477 course up to the front end to adjust this for any additional window
2478 furniture such as menu bars and window borders, if necessary. The
2479 status bar is also not included in this size.)
2481 If \c{expand} is set to \cw{FALSE}, then the game's tile size will
2482 never go over its preferred one. This is the recommended approach
2483 when opening a new window at default size: the game will use its
2484 preferred size unless it has to use a smaller one to fit on the
2487 If \c{expand} is set to \cw{TRUE}, the mid-end will pick a tile size
2488 which approximates the input size \e{as closely as possible}, and
2489 will go over the game's preferred tile size if necessary to achieve
2490 this. Use this option if you want your front end to support dynamic
2491 resizing of the puzzle window with automatic scaling of the puzzle
2494 The mid-end will try as hard as it can to return a size which is
2495 less than or equal to the input size, in both dimensions. In extreme
2496 circumstances it may fail (if even the lowest possible tile size
2497 gives window dimensions greater than the input), in which case it
2498 will return a size greater than the input size. Front ends should be
2499 prepared for this to happen (i.e. don't crash or fail an assertion),
2500 but may handle it in any way they see fit: by rejecting the game
2501 parameters which caused the problem, by opening a window larger than
2502 the screen regardless of inconvenience, by introducing scroll bars
2503 on the window, by drawing on a large bitmap and scaling it into a
2504 smaller window, or by any other means you can think of. It is likely
2505 that when the tile size is that small the game will be unplayable
2506 anyway, so don't put \e{too} much effort into handling it
2509 If your platform has no limit on window size (or if you're planning
2510 to use scroll bars for large puzzles), you can pass dimensions of
2511 \cw{INT_MAX} as input to this function. You should probably not do
2512 that \e{and} set the \c{expand} flag, though!
2514 \H{midend-new-game} \cw{midend_new_game()}
2516 \c void midend_new_game(midend *me);
2518 Causes the mid-end to begin a new game. Normally the game will be a
2519 new randomly generated puzzle. However, if you have previously
2520 called \cw{midend_game_id()} or \cw{midend_set_config()}, the game
2521 generated might be dictated by the results of those functions. (In
2522 particular, you \e{must} call \cw{midend_new_game()} after calling
2523 either of those functions, or else no immediate effect will be
2526 You will probably need to call \cw{midend_size()} after calling this
2527 function, because if the game parameters have been changed since the
2528 last new game then the window size might need to change. (If you
2529 know the parameters \e{haven't} changed, you don't need to do this.)
2531 This function will create a new \c{game_drawstate}, but does not
2532 actually perform a redraw (since you often need to call
2533 \cw{midend_size()} before the redraw can be done). So after calling
2534 this function and after calling \cw{midend_size()}, you should then
2535 call \cw{midend_redraw()}. (It is not necessary to call
2536 \cw{midend_force_redraw()}; that will discard the draw state and
2537 create a fresh one, which is unnecessary in this case since there's
2538 a fresh one already. It would work, but it's usually excessive.)
2540 \H{midend-restart-game} \cw{midend_restart_game()}
2542 \c void midend_restart_game(midend *me);
2544 This function causes the current game to be restarted. This is done
2545 by placing a new copy of the original game state on the end of the
2546 undo list (so that an accidental restart can be undone).
2548 This function automatically causes a redraw, i.e. the front end can
2549 expect its drawing API to be called from \e{within} a call to this
2552 \H{midend-force-redraw} \cw{midend_force_redraw()}
2554 \c void midend_force_redraw(midend *me);
2556 Forces a complete redraw of the puzzle window, by means of
2557 discarding the current \c{game_drawstate} and creating a new one
2558 from scratch before calling the game's \cw{redraw()} function.
2560 The front end can expect its drawing API to be called from within a
2561 call to this function.
2563 \H{midend-redraw} \cw{midend_redraw()}
2565 \c void midend_redraw(midend *me);
2567 Causes a partial redraw of the puzzle window, by means of simply
2568 calling the game's \cw{redraw()} function. (That is, the only things
2569 redrawn will be things that have changed since the last redraw.)
2571 The front end can expect its drawing API to be called from within a
2572 call to this function.
2574 \H{midend-process-key} \cw{midend_process_key()}
2576 \c int midend_process_key(midend *me, int x, int y, int button);
2578 The front end calls this function to report a mouse or keyboard
2579 event. The parameters \c{x}, \c{y} and \c{button} are almost
2580 identical to the ones passed to the back end function
2581 \cw{interpret_move()} (\k{backend-interpret-move}), except that the
2582 front end is \e{not} required to provide the guarantees about mouse
2583 event ordering. The mid-end will sort out multiple simultaneous
2584 button presses and changes of button; the front end's responsibility
2585 is simply to pass on the mouse events it receives as accurately as
2588 (Some platforms may need to emulate absent mouse buttons by means of
2589 using a modifier key such as Shift with another mouse button. This
2590 tends to mean that if Shift is pressed or released in the middle of
2591 a mouse drag, the mid-end will suddenly stop receiving, say,
2592 \cw{LEFT_DRAG} events and start receiving \cw{RIGHT_DRAG}s, with no
2593 intervening button release or press events. This too is something
2594 which the mid-end will sort out for you; the front end has no
2595 obligation to maintain sanity in this area.)
2597 The front end \e{should}, however, always eventually send some kind
2598 of button release. On some platforms this requires special effort:
2599 Windows, for example, requires a call to the system API function
2600 \cw{SetCapture()} in order to ensure that your window receives a
2601 mouse-up event even if the pointer has left the window by the time
2602 the mouse button is released. On any platform that requires this
2603 sort of thing, the front end \e{is} responsible for doing it.
2605 Calling this function is very likely to result in calls back to the
2606 front end's drawing API and/or \cw{activate_timer()}
2607 (\k{frontend-activate-timer}).
2609 \H{midend-colours} \cw{midend_colours()}
2611 \c float *midend_colours(midend *me, int *ncolours);
2613 Returns an array of the colours required by the game, in exactly the
2614 same format as that returned by the back end function \cw{colours()}
2615 (\k{backend-colours}). Front ends should call this function rather
2616 than calling the back end's version directly, since the mid-end adds
2617 standard customisation facilities. (At the time of writing, those
2618 customisation facilities are implemented hackily by means of
2619 environment variables, but it's not impossible that they may become
2620 more full and formal in future.)
2622 \H{midend-timer} \cw{midend_timer()}
2624 \c void midend_timer(midend *me, float tplus);
2626 If the mid-end has called \cw{activate_timer()}
2627 (\k{frontend-activate-timer}) to request regular callbacks for
2628 purposes of animation or timing, this is the function the front end
2629 should call on a regular basis. The argument \c{tplus} gives the
2630 time, in seconds, since the last time either this function was
2631 called or \cw{activate_timer()} was invoked.
2633 One of the major purposes of timing in the mid-end is to perform
2634 move animation. Therefore, calling this function is very likely to
2635 result in calls back to the front end's drawing API.
2637 \H{midend-num-presets} \cw{midend_num_presets()}
2639 \c int midend_num_presets(midend *me);
2641 Returns the number of game parameter presets supplied by this game.
2642 Front ends should use this function and \cw{midend_fetch_preset()}
2643 to configure their presets menu rather than calling the back end
2644 directly, since the mid-end adds standard customisation facilities.
2645 (At the time of writing, those customisation facilities are
2646 implemented hackily by means of environment variables, but it's not
2647 impossible that they may become more full and formal in future.)
2649 \H{midend-fetch-preset} \cw{midend_fetch_preset()}
2651 \c void midend_fetch_preset(midend *me, int n,
2652 \c char **name, game_params **params);
2654 Returns one of the preset game parameter structures for the game. On
2655 input \c{n} must be a non-negative integer and less than the value
2656 returned from \cw{midend_num_presets()}. On output, \c{*name} is set
2657 to an ASCII string suitable for entering in the game's presets menu,
2658 and \c{*params} is set to the corresponding \c{game_params}
2661 Both of the two output values are dynamically allocated, but they
2662 are owned by the mid-end structure: the front end should not ever
2663 free them directly, because they will be freed automatically during
2666 \H{midend-wants-statusbar} \cw{midend_wants_statusbar()}
2668 \c int midend_wants_statusbar(midend *me);
2670 This function returns \cw{TRUE} if the puzzle has a use for a
2671 textual status line (to display score, completion status, currently
2672 active tiles, time, or anything else).
2674 Front ends should call this function rather than talking directly to
2677 \H{midend-get-config} \cw{midend_get_config()}
2679 \c config_item *midend_get_config(midend *me, int which,
2680 \c char **wintitle);
2682 Returns a dialog box description for user configuration.
2684 On input, \cw{which} should be set to one of three values, which
2685 select which of the various dialog box descriptions is returned:
2687 \dt \cw{CFG_SETTINGS}
2689 \dd Requests the GUI parameter configuration box generated by the
2690 puzzle itself. This should be used when the user selects \q{Custom}
2691 from the game types menu (or equivalent). The mid-end passes this
2692 request on to the back end function \cw{configure()}
2693 (\k{backend-configure}).
2697 \dd Requests a box suitable for entering a descriptive game ID (and
2698 viewing the existing one). The mid-end generates this dialog box
2699 description itself. This should be used when the user selects
2700 \q{Specific} from the game menu (or equivalent).
2704 \dd Requests a box suitable for entering a random-seed game ID (and
2705 viewing the existing one). The mid-end generates this dialog box
2706 description itself. This should be used when the user selects
2707 \q{Random Seed} from the game menu (or equivalent).
2709 The returned value is an array of \cw{config_item}s, exactly as
2710 described in \k{backend-configure}. Another returned value is an
2711 ASCII string giving a suitable title for the configuration window,
2714 Both returned values are dynamically allocated and will need to be
2715 freed. The window title can be freed in the obvious way; the
2716 \cw{config_item} array is a slightly complex structure, so a utility
2717 function \cw{free_cfg()} is provided to free it for you. See
2720 (Of course, you will probably not want to free the \cw{config_item}
2721 array until the dialog box is dismissed, because before then you
2722 will probably need to pass it to \cw{midend_set_config}.)
2724 \H{midend-set-config} \cw{midend_set_config()}
2726 \c char *midend_set_config(midend *me, int which,
2727 \c config_item *cfg);
2729 Passes the mid-end the results of a configuration dialog box.
2730 \c{which} should have the same value which it had when
2731 \cw{midend_get_config()} was called; \c{cfg} should be the array of
2732 \c{config_item}s returned from \cw{midend_get_config()}, modified to
2733 contain the results of the user's editing operations.
2735 This function returns \cw{NULL} on success, or otherwise (if the
2736 configuration data was in some way invalid) an ASCII string
2737 containing an error message suitable for showing to the user.
2739 If the function succeeds, it is likely that the game parameters will
2740 have been changed and it is certain that a new game will be
2741 requested. The front end should therefore call
2742 \cw{midend_new_game()}, and probably also re-think the window size
2743 using \cw{midend_size()} and eventually perform a refresh using
2744 \cw{midend_redraw()}.
2746 \H{midend-game-id} \cw{midend_game_id()}
2748 \c char *midend_game_id(midend *me, char *id);
2750 Passes the mid-end a string game ID (of any of the valid forms
2751 \cq{params}, \cq{params:description} or \cq{params#seed}) which the
2752 mid-end will process and use for the next generated game.
2754 This function returns \cw{NULL} on success, or otherwise (if the
2755 configuration data was in some way invalid) an ASCII string
2756 containing an error message (not dynamically allocated) suitable for
2757 showing to the user. In the event of an error, the mid-end's
2758 internal state will be left exactly as it was before the call.
2760 If the function succeeds, it is likely that the game parameters will
2761 have been changed and it is certain that a new game will be
2762 requested. The front end should therefore call
2763 \cw{midend_new_game()}, and probably also re-think the window size
2764 using \cw{midend_size()} and eventually case a refresh using
2765 \cw{midend_redraw()}.
2767 \H{midend-get-game-id} \cw{midend_get_game_id()}
2769 \c char *midend_get_game_id(midend *me)
2771 Returns a descriptive game ID (i.e. one in the form
2772 \cq{params:description}) describing the game currently active in the
2773 mid-end. The returned string is dynamically allocated.
2775 \H{midend-text-format} \cw{midend_text_format()}
2777 \c char *midend_text_format(midend *me);
2779 Formats the current game's current state as ASCII text suitable for
2780 copying to the clipboard. The returned string is dynamically
2783 You should not call this function if the game's
2784 \c{can_format_as_text} flag is \cw{FALSE}.
2786 If the returned string contains multiple lines (which is likely), it
2787 will use the normal C line ending convention (\cw{\\n} only). On
2788 platforms which use a different line ending convention for data in
2789 the clipboard, it is the front end's responsibility to perform the
2792 \H{midend-solve} \cw{midend_solve()}
2794 \c char *midend_solve(midend *me);
2796 Requests the mid-end to perform a Solve operation.
2798 On success, \cw{NULL} is returned. On failure, an error message (not
2799 dynamically allocated) is returned, suitable for showing to the
2802 The front end can expect its drawing API and/or
2803 \cw{activate_timer()} to be called from within a call to this
2806 \H{midend-serialise} \cw{midend_serialise()}
2808 \c void midend_serialise(midend *me,
2809 \c void (*write)(void *ctx, void *buf, int len),
2812 Calling this function causes the mid-end to convert its entire
2813 internal state into a long ASCII text string, and to pass that
2814 string (piece by piece) to the supplied \c{write} function.
2816 Desktop implementations can use this function to save a game in any
2817 state (including half-finished) to a disk file, by supplying a
2818 \c{write} function which is a wrapper on \cw{fwrite()} (or local
2819 equivalent). Other implementations may find other uses for it, such
2820 as compressing the large and sprawling mid-end state into a
2821 manageable amount of memory when a palmtop application is suspended
2822 so that another one can run; in this case \cw{write} might want to
2823 write to a memory buffer rather than a file. There may be other uses
2826 This function will call back to the supplied \c{write} function a
2827 number of times, with the first parameter (\c{ctx}) equal to
2828 \c{wctx}, and the other two parameters pointing at a piece of the
2831 \H{midend-deserialise} \cw{midend_deserialise()}
2833 \c char *midend_deserialise(midend *me,
2834 \c int (*read)(void *ctx, void *buf, int len),
2837 This function is the counterpart to \cw{midend_serialise()}. It
2838 calls the supplied \cw{read} function repeatedly to read a quantity
2839 of data, and attempts to interpret that data as a serialised mid-end
2840 as output by \cw{midend_serialise()}.
2842 The \cw{read} function is called with the first parameter (\c{ctx})
2843 equal to \c{rctx}, and should attempt to read \c{len} bytes of data
2844 into the buffer pointed to by \c{buf}. It should return \cw{FALSE}
2845 on failure or \cw{TRUE} on success. It should not report success
2846 unless it has filled the entire buffer; on platforms which might be
2847 reading from a pipe or other blocking data source, \c{read} is
2848 responsible for looping until the whole buffer has been filled.
2850 If the de-serialisation operation is successful, the mid-end's
2851 internal data structures will be replaced by the results of the
2852 load, and \cw{NULL} will be returned. Otherwise, the mid-end's state
2853 will be completely unchanged and an error message (typically some
2854 variation on \q{save file is corrupt}) will be returned. As usual,
2855 the error message string is not dynamically allocated.
2857 If this function succeeds, it is likely that the game parameters
2858 will have been changed. The front end should therefore probably
2859 re-think the window size using \cw{midend_size()}, and probably
2860 cause a refresh using \cw{midend_redraw()}.
2862 Because each mid-end is tied to a specific game back end, this
2863 function will fail if you attempt to read in a save file generated
2864 by a different game from the one configured in this mid-end, even if
2865 your application is a monolithic one containing all the puzzles. (It
2866 would be pretty easy to write a function which would look at a save
2867 file and determine which game it was for; any front end implementor
2868 who needs such a function can probably be accommodated.)
2870 \H{frontend-backend} Direct reference to the back end structure by
2873 Although \e{most} things the front end needs done should be done by
2874 calling the mid-end, there are a few situations in which the front
2875 end needs to refer directly to the game back end structure.
2877 The most obvious of these is
2879 \b passing the game back end as a parameter to \cw{midend_new()}.
2881 There are a few other back end features which are not wrapped by the
2882 mid-end because there didn't seem much point in doing so:
2884 \b fetching the \c{name} field to use in window titles and similar
2886 \b reading the \c{can_configure}, \c{can_solve} and
2887 \c{can_format_as_text} fields to decide whether to add those items
2888 to the menu bar or equivalent
2890 \b reading the \c{winhelp_topic} field (Windows only)
2892 \b the GTK front end provides a \cq{--generate} command-line option
2893 which directly calls the back end to do most of its work. This is
2894 not really part of the main front end code, though, and I'm not sure
2897 In order to find the game back end structure, the front end does one
2900 \b If the particular front end is compiling a separate binary per
2901 game, then the back end structure is a global variable with the
2902 standard name \cq{thegame}:
2906 \c extern const game thegame;
2910 \b If the front end is compiled as a monolithic application
2911 containing all the puzzles together (in which case the preprocessor
2912 symbol \cw{COMBINED} must be defined when compiling most of the code
2913 base), then there will be two global variables defined:
2917 \c extern const game *gamelist[];
2918 \c extern const int gamecount;
2920 \c{gamelist} will be an array of \c{gamecount} game structures,
2921 declared in the automatically constructed source module \c{list.c}.
2922 The application should search that array for the game it wants,
2923 probably by reaching into each game structure and looking at its
2928 \H{frontend-api} Mid-end to front-end calls
2930 This section describes the small number of functions which a front
2931 end must provide to be called by the mid-end or other standard
2934 \H{frontend-get-random-seed} \cw{get_random_seed()}
2936 \c void get_random_seed(void **randseed, int *randseedsize);
2938 This function is called by a new mid-end, and also occasionally by
2939 game back ends. Its job is to return a piece of data suitable for
2940 using as a seed for initialisation of a new \c{random_state}.
2942 On exit, \c{*randseed} should be set to point at a newly allocated
2943 piece of memory containing some seed data, and \c{*randseedsize}
2944 should be set to the length of that data.
2946 A simple and entirely adequate implementation is to return a piece
2947 of data containing the current system time at the highest
2948 conveniently available resolution.
2950 \H{frontend-activate-timer} \cw{activate_timer()}
2952 \c void activate_timer(frontend *fe);
2954 This is called by the mid-end to request that the front end begin
2955 calling it back at regular intervals.
2957 The timeout interval is left up to the front end; the finer it is,
2958 the smoother move animations will be, but the more CPU time will be
2959 used. Current front ends use values around 20ms (i.e. 50Hz).
2961 After this function is called, the mid-end will expect to receive
2962 calls to \cw{midend_timer()} on a regular basis.
2964 \H{frontend-deactivate-timer} \cw{deactivate_timer()}
2966 \c void deactivate_timer(frontend *fe);
2968 This is called by the mid-end to request that the front end stop
2969 calling \cw{midend_timer()}.
2971 \H{frontend-fatal} \cw{fatal()}
2973 \c void fatal(char *fmt, ...);
2975 This is called by some utility functions if they encounter a
2976 genuinely fatal error such as running out of memory. It is a
2977 variadic function in the style of \cw{printf()}, and is expected to
2978 show the formatted error message to the user any way it can and then
2979 terminate the application. It must not return.
2981 \H{frontend-default-colour} \cw{frontend_default_colour()}
2983 \c void frontend_default_colour(frontend *fe, float *output);
2985 This function expects to be passed a pointer to an array of three
2986 \cw{float}s. It returns the platform's local preferred background
2987 colour in those three floats, as red, green and blue values (in that
2988 order) ranging from \cw{0.0} to \cw{1.0}.
2990 This function should only ever be called by the back end function
2991 \cw{colours()} (\k{backend-colours}). (Thus, it isn't a
2992 \e{midend}-to-frontend function as such, but there didn't seem to be
2993 anywhere else particularly good to put it. Sorry.)
2995 \C{utils} Utility APIs
2997 This chapter documents a variety of utility APIs provided for the
2998 general use of the rest of the Puzzles code.
3000 \H{utils-random} Random number generation
3002 Platforms' local random number generators vary widely in quality and
3003 seed size. Puzzles therefore supplies its own high-quality random
3004 number generator, with the additional advantage of giving the same
3005 results if fed the same seed data on different platforms. This
3006 allows game random seeds to be exchanged between different ports of
3007 Puzzles and still generate the same games.
3009 Unlike the ANSI C \cw{rand()} function, the Puzzles random number
3010 generator has an \e{explicit} state object called a
3011 \c{random_state}. One of these is managed by each mid-end, for
3012 example, and passed to the back end to generate a game with.
3014 \S{utils-random-init} \cw{random_new()}
3016 \c random_state *random_new(char *seed, int len);
3018 Allocates, initialises and returns a new \c{random_state}. The input
3019 data is used as the seed for the random number stream (i.e. using
3020 the same seed at a later time will generate the same stream).
3022 The seed data can be any data at all; there is no requirement to use
3023 printable ASCII, or NUL-terminated strings, or anything like that.
3025 \S{utils-random-copy} \cw{random_copy()}
3027 \c random_state *random_copy(random_state *tocopy);
3029 Allocates a new \c{random_state}, copies the contents of another
3030 \c{random_state} into it, and returns the new state. If exactly the
3031 same sequence of functions is subseqently called on both the copy and
3032 the original, the results will be identical. This may be useful for
3033 speculatively performing some operation using a given random state,
3034 and later replaying that operation precisely.
3036 \S{utils-random-free} \cw{random_free()}
3038 \c void random_free(random_state *state);
3040 Frees a \c{random_state}.
3042 \S{utils-random-bits} \cw{random_bits()}
3044 \c unsigned long random_bits(random_state *state, int bits);
3046 Returns a random number from 0 to \cw{2^bits-1} inclusive. \c{bits}
3047 should be between 1 and 32 inclusive.
3049 \S{utils-random-upto} \cw{random_upto()}
3051 \c unsigned long random_upto(random_state *state, unsigned long limit);
3053 Returns a random number from 0 to \cw{limit-1} inclusive.
3055 \S{utils-random-state-encode} \cw{random_state_encode()}
3057 \c char *random_state_encode(random_state *state);
3059 Encodes the entire contents of a \c{random_state} in printable
3060 ASCII. Returns a dynamically allocated string containing that
3061 encoding. This can subsequently be passed to
3062 \cw{random_state_decode()} to reconstruct the same \c{random_state}.
3064 \S{utils-random-state-decode} \cw{random_state_decode()}
3066 \c random_state *random_state_decode(char *input);
3068 Decodes a string generated by \cw{random_state_encode()} and
3069 reconstructs an equivalent \c{random_state} to the one encoded, i.e.
3070 it should produce the same stream of random numbers.
3072 This function has no error reporting; if you pass it an invalid
3073 string it will simply generate an arbitrary random state, which may
3074 turn out to be noticeably non-random.
3076 \S{utils-shuffle} \cw{shuffle()}
3078 \c void shuffle(void *array, int nelts, int eltsize, random_state *rs);
3080 Shuffles an array into a random order. The interface is much like
3081 ANSI C \cw{qsort()}, except that there's no need for a compare
3084 \c{array} is a pointer to the first element of the array. \c{nelts}
3085 is the number of elements in the array; \c{eltsize} is the size of a
3086 single element (typically measured using \c{sizeof}). \c{rs} is a
3087 \c{random_state} used to generate all the random numbers for the
3090 \H{utils-alloc} Memory allocation
3092 Puzzles has some central wrappers on the standard memory allocation
3093 functions, which provide compile-time type checking, and run-time
3094 error checking by means of quitting the application if it runs out
3095 of memory. This doesn't provide the best possible recovery from
3096 memory shortage, but on the other hand it greatly simplifies the
3097 rest of the code, because nothing else anywhere needs to worry about
3098 \cw{NULL} returns from allocation.
3100 \S{utils-snew} \cw{snew()}
3102 \c var = snew(type);
3105 This macro takes a single argument which is a \e{type name}. It
3106 allocates space for one object of that type. If allocation fails it
3107 will call \cw{fatal()} and not return; so if it does return, you can
3108 be confident that its return value is non-\cw{NULL}.
3110 The return value is cast to the specified type, so that the compiler
3111 will type-check it against the variable you assign it into. Thus,
3112 this ensures you don't accidentally allocate memory the size of the
3113 wrong type and assign it into a variable of the right one (or vice
3116 \S{utils-snewn} \cw{snewn()}
3118 \c var = snewn(n, type);
3121 This macro is the array form of \cw{snew()}. It takes two arguments;
3122 the first is a number, and the second is a type name. It allocates
3123 space for that many objects of that type, and returns a type-checked
3124 non-\cw{NULL} pointer just as \cw{snew()} does.
3126 \S{utils-sresize} \cw{sresize()}
3128 \c var = sresize(var, n, type);
3131 This macro is a type-checked form of \cw{realloc()}. It takes three
3132 arguments: an input memory block, a new size in elements, and a
3133 type. It re-sizes the input memory block to a size sufficient to
3134 contain that many elements of that type. It returns a type-checked
3135 non-\cw{NULL} pointer, like \cw{snew()} and \cw{snewn()}.
3137 The input memory block can be \cw{NULL}, in which case this function
3138 will behave exactly like \cw{snewn()}. (In principle any
3139 ANSI-compliant \cw{realloc()} implementation ought to cope with
3140 this, but I've never quite trusted it to work everywhere.)
3142 \S{utils-sfree} \cw{sfree()}
3144 \c void sfree(void *p);
3146 This function is pretty much equivalent to \cw{free()}. It is
3147 provided with a dynamically allocated block, and frees it.
3149 The input memory block can be \cw{NULL}, in which case this function
3150 will do nothing. (In principle any ANSI-compliant \cw{free()}
3151 implementation ought to cope with this, but I've never quite trusted
3152 it to work everywhere.)
3154 \S{utils-dupstr} \cw{dupstr()}
3156 \c char *dupstr(const char *s);
3158 This function dynamically allocates a duplicate of a C string. Like
3159 the \cw{snew()} functions, it guarantees to return non-\cw{NULL} or
3162 (Many platforms provide the function \cw{strdup()}. As well as
3163 guaranteeing never to return \cw{NULL}, my version has the advantage
3164 of being defined \e{everywhere}, rather than inconveniently not
3167 \S{utils-free-cfg} \cw{free_cfg()}
3169 \c void free_cfg(config_item *cfg);
3171 This function correctly frees an array of \c{config_item}s,
3172 including walking the array until it gets to the end and freeing
3173 precisely those \c{sval} fields which are expected to be dynamically
3176 (See \k{backend-configure} for details of the \c{config_item}
3179 \H{utils-tree234} Sorted and counted tree functions
3181 Many games require complex algorithms for generating random puzzles,
3182 and some require moderately complex algorithms even during play. A
3183 common requirement during these algorithms is for a means of
3184 maintaining sorted or unsorted lists of items, such that items can
3185 be removed and added conveniently.
3187 For general use, Puzzles provides the following set of functions
3188 which maintain 2-3-4 trees in memory. (A 2-3-4 tree is a balanced
3189 tree structure, with the property that all lookups, insertions,
3190 deletions, splits and joins can be done in \cw{O(log N)} time.)
3192 All these functions expect you to be storing a tree of \c{void *}
3193 pointers. You can put anything you like in those pointers.
3195 By the use of per-node element counts, these tree structures have
3196 the slightly unusual ability to look elements up by their numeric
3197 index within the list represented by the tree. This means that they
3198 can be used to store an unsorted list (in which case, every time you
3199 insert a new element, you must explicitly specify the position where
3200 you wish to insert it). They can also do numeric lookups in a sorted
3201 tree, which might be useful for (for example) tracking the median of
3202 a changing data set.
3204 As well as storing sorted lists, these functions can be used for
3205 storing \q{maps} (associative arrays), by defining each element of a
3206 tree to be a (key, value) pair.
3208 \S{utils-newtree234} \cw{newtree234()}
3210 \c tree234 *newtree234(cmpfn234 cmp);
3212 Creates a new empty tree, and returns a pointer to it.
3214 The parameter \c{cmp} determines the sorting criterion on the tree.
3217 \c typedef int (*cmpfn234)(void *, void *);
3219 If you want a sorted tree, you should provide a function matching
3220 this prototype, which returns like \cw{strcmp()} does (negative if
3221 the first argument is smaller than the second, positive if it is
3222 bigger, zero if they compare equal). In this case, the function
3223 \cw{addpos234()} will not be usable on your tree (because all
3224 insertions must respect the sorting order).
3226 If you want an unsorted tree, pass \cw{NULL}. In this case you will
3227 not be able to use either \cw{add234()} or \cw{del234()}, or any
3228 other function such as \cw{find234()} which depends on a sorting
3229 order. Your tree will become something more like an array, except
3230 that it will efficiently support insertion and deletion as well as
3231 lookups by numeric index.
3233 \S{utils-freetree234} \cw{freetree234()}
3235 \c void freetree234(tree234 *t);
3237 Frees a tree. This function will not free the \e{elements} of the
3238 tree (because they might not be dynamically allocated, or you might
3239 be storing the same set of elements in more than one tree); it will
3240 just free the tree structure itself. If you want to free all the
3241 elements of a tree, you should empty it before passing it to
3242 \cw{freetree234()}, by means of code along the lines of
3244 \c while ((element = delpos234(tree, 0)) != NULL)
3245 \c sfree(element); /* or some more complicated free function */
3246 \e iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
3248 \S{utils-add234} \cw{add234()}
3250 \c void *add234(tree234 *t, void *e);
3252 Inserts a new element \c{e} into the tree \c{t}. This function
3253 expects the tree to be sorted; the new element is inserted according
3256 If an element comparing equal to \c{e} is already in the tree, then
3257 the insertion will fail, and the return value will be the existing
3258 element. Otherwise, the insertion succeeds, and \c{e} is returned.
3260 \S{utils-addpos234} \cw{addpos234()}
3262 \c void *addpos234(tree234 *t, void *e, int index);
3264 Inserts a new element into an unsorted tree. Since there is no
3265 sorting order to dictate where the new element goes, you must
3266 specify where you want it to go. Setting \c{index} to zero puts the
3267 new element right at the start of the list; setting \c{index} to the
3268 current number of elements in the tree puts the new element at the
3271 Return value is \c{e}, in line with \cw{add234()} (although this
3272 function cannot fail except by running out of memory, in which case
3273 it will bomb out and die rather than returning an error indication).
3275 \S{utils-index234} \cw{index234()}
3277 \c void *index234(tree234 *t, int index);
3279 Returns a pointer to the \c{index}th element of the tree, or
3280 \cw{NULL} if \c{index} is out of range. Elements of the tree are
3283 \S{utils-find234} \cw{find234()}
3285 \c void *find234(tree234 *t, void *e, cmpfn234 cmp);
3287 Searches for an element comparing equal to \c{e} in a sorted tree.
3289 If \c{cmp} is \cw{NULL}, the tree's ordinary comparison function
3290 will be used to perform the search. However, sometimes you don't
3291 want that; suppose, for example, each of your elements is a big
3292 structure containing a \c{char *} name field, and you want to find
3293 the element with a given name. You \e{could} achieve this by
3294 constructing a fake element structure, setting its name field
3295 appropriately, and passing it to \cw{find234()}, but you might find
3296 it more convenient to pass \e{just} a name string to \cw{find234()},
3297 supplying an alternative comparison function which expects one of
3298 its arguments to be a bare name and the other to be a large
3299 structure containing a name field.
3301 Therefore, if \c{cmp} is not \cw{NULL}, then it will be used to
3302 compare \c{e} to elements of the tree. The first argument passed to
3303 \c{cmp} will always be \c{e}; the second will be an element of the
3306 (See \k{utils-newtree234} for the definition of the \c{cmpfn234}
3307 function pointer type.)
3309 The returned value is the element found, or \cw{NULL} if the search
3312 \S{utils-findrel234} \cw{findrel234()}
3314 \c void *findrel234(tree234 *t, void *e, cmpfn234 cmp, int relation);
3316 This function is like \cw{find234()}, but has the additional ability
3317 to do a \e{relative} search. The additional parameter \c{relation}
3318 can be one of the following values:
3322 \dd Find only an element that compares equal to \c{e}. This is
3323 exactly the behaviour of \cw{find234()}.
3327 \dd Find the greatest element that compares strictly less than
3328 \c{e}. \c{e} may be \cw{NULL}, in which case it finds the greatest
3329 element in the whole tree (which could also be done by
3330 \cw{index234(t, count234(t)-1)}).
3334 \dd Find the greatest element that compares less than or equal to
3335 \c{e}. (That is, find an element that compares equal to \c{e} if
3336 possible, but failing that settle for something just less than it.)
3340 \dd Find the smallest element that compares strictly greater than
3341 \c{e}. \c{e} may be \cw{NULL}, in which case it finds the smallest
3342 element in the whole tree (which could also be done by
3343 \cw{index234(t, 0)}).
3347 \dd Find the smallest element that compares greater than or equal to
3348 \c{e}. (That is, find an element that compares equal to \c{e} if
3349 possible, but failing that settle for something just bigger than
3352 Return value, as before, is the element found or \cw{NULL} if no
3353 element satisfied the search criterion.
3355 \S{utils-findpos234} \cw{findpos234()}
3357 \c void *findpos234(tree234 *t, void *e, cmpfn234 cmp, int *index);
3359 This function is like \cw{find234()}, but has the additional feature
3360 of returning the index of the element found in the tree; that index
3361 is written to \c{*index} in the event of a successful search (a
3362 non-\cw{NULL} return value).
3364 \c{index} may be \cw{NULL}, in which case this function behaves
3365 exactly like \cw{find234()}.
3367 \S{utils-findrelpos234} \cw{findrelpos234()}
3369 \c void *findrelpos234(tree234 *t, void *e, cmpfn234 cmp, int relation,
3372 This function combines all the features of \cw{findrel234()} and
3375 \S{utils-del234} \cw{del234()}
3377 \c void *del234(tree234 *t, void *e);
3379 Finds an element comparing equal to \c{e} in the tree, deletes it,
3382 The input tree must be sorted.
3384 The element found might be \c{e} itself, or might merely compare
3387 Return value is \cw{NULL} if no such element is found.
3389 \S{utils-delpos234} \cw{delpos234()}
3391 \c void *delpos234(tree234 *t, int index);
3393 Deletes the element at position \c{index} in the tree, and returns
3396 Return value is \cw{NULL} if the index is out of range.
3398 \S{utils-count234} \cw{count234()}
3400 \c int count234(tree234 *t);
3402 Returns the number of elements currently in the tree.
3404 \S{utils-splitpos234} \cw{splitpos234()}
3406 \c tree234 *splitpos234(tree234 *t, int index, int before);
3408 Splits the input tree into two pieces at a given position, and
3409 creates a new tree containing all the elements on one side of that
3412 If \c{before} is \cw{TRUE}, then all the items at or after position
3413 \c{index} are left in the input tree, and the items before that
3414 point are returned in the new tree. Otherwise, the reverse happens:
3415 all the items at or after \c{index} are moved into the new tree, and
3416 those before that point are left in the old one.
3418 If \c{index} is equal to 0 or to the number of elements in the input
3419 tree, then one of the two trees will end up empty (and this is not
3420 an error condition). If \c{index} is further out of range in either
3421 direction, the operation will fail completely and return \cw{NULL}.
3423 This operation completes in \cw{O(log N)} time, no matter how large
3424 the tree or how balanced or unbalanced the split.
3426 \S{utils-split234} \cw{split234()}
3428 \c tree234 *split234(tree234 *t, void *e, cmpfn234 cmp, int rel);
3430 Splits a sorted tree according to its sort order.
3432 \c{rel} can be any of the relation constants described in
3433 \k{utils-findrel234}, \e{except} for \cw{REL234_EQ}. All the
3434 elements having that relation to \c{e} will be transferred into the
3435 new tree; the rest will be left in the old one.
3437 The parameter \c{cmp} has the same semantics as it does in
3438 \cw{find234()}: if it is not \cw{NULL}, it will be used in place of
3439 the tree's own comparison function when comparing elements to \c{e},
3440 in such a way that \c{e} itself is always the first of its two
3443 Again, this operation completes in \cw{O(log N)} time, no matter how
3444 large the tree or how balanced or unbalanced the split.
3446 \S{utils-join234} \cw{join234()}
3448 \c tree234 *join234(tree234 *t1, tree234 *t2);
3450 Joins two trees together by concatenating the lists they represent.
3451 All the elements of \c{t2} are moved into \c{t1}, in such a way that
3452 they appear \e{after} the elements of \c{t1}. The tree \c{t2} is
3453 freed; the return value is \c{t1}.
3455 If you apply this function to a sorted tree and it violates the sort
3456 order (i.e. the smallest element in \c{t2} is smaller than or equal
3457 to the largest element in \c{t1}), the operation will fail and
3460 This operation completes in \cw{O(log N)} time, no matter how large
3461 the trees being joined together.
3463 \S{utils-join234r} \cw{join234r()}
3465 \c tree234 *join234r(tree234 *t1, tree234 *t2);
3467 Joins two trees together in exactly the same way as \cw{join234()},
3468 but this time the combined tree is returned in \c{t2}, and \c{t1} is
3469 destroyed. The elements in \c{t1} still appear before those in
3472 Again, this operation completes in \cw{O(log N)} time, no matter how
3473 large the trees being joined together.
3475 \S{utils-copytree234} \cw{copytree234()}
3477 \c tree234 *copytree234(tree234 *t, copyfn234 copyfn,
3478 \c void *copyfnstate);
3480 Makes a copy of an entire tree.
3482 If \c{copyfn} is \cw{NULL}, the tree will be copied but the elements
3483 will not be; i.e. the new tree will contain pointers to exactly the
3484 same physical elements as the old one.
3486 If you want to copy each actual element during the operation, you
3487 can instead pass a function in \c{copyfn} which makes a copy of each
3488 element. That function has the prototype
3490 \c typedef void *(*copyfn234)(void *state, void *element);
3492 and every time it is called, the \c{state} parameter will be set to
3493 the value you passed in as \c{copyfnstate}.
3495 \H{utils-misc} Miscellaneous utility functions and macros
3497 This section contains all the utility functions which didn't
3498 sensibly fit anywhere else.
3500 \S{utils-truefalse} \cw{TRUE} and \cw{FALSE}
3502 The main Puzzles header file defines the macros \cw{TRUE} and
3503 \cw{FALSE}, which are used throughout the code in place of 1 and 0
3504 (respectively) to indicate that the values are in a boolean context.
3505 For code base consistency, I'd prefer it if submissions of new code
3506 followed this convention as well.
3508 \S{utils-maxmin} \cw{max()} and \cw{min()}
3510 The main Puzzles header file defines the pretty standard macros
3511 \cw{max()} and \cw{min()}, each of which is given two arguments and
3512 returns the one which compares greater or less respectively.
3514 These macros may evaluate their arguments multiple times. Avoid side
3517 \S{utils-pi} \cw{PI}
3519 The main Puzzles header file defines a macro \cw{PI} which expands
3520 to a floating-point constant representing pi.
3522 (I've never understood why ANSI's \cw{<math.h>} doesn't define this.
3525 \S{utils-obfuscate-bitmap} \cw{obfuscate_bitmap()}
3527 \c void obfuscate_bitmap(unsigned char *bmp, int bits, int decode);
3529 This function obscures the contents of a piece of data, by
3530 cryptographic methods. It is useful for games of hidden information
3531 (such as Mines, Guess or Black Box), in which the game ID
3532 theoretically reveals all the information the player is supposed to
3533 be trying to guess. So in order that players should be able to send
3534 game IDs to one another without accidentally spoiling the resulting
3535 game by looking at them, these games obfuscate their game IDs using
3538 Although the obfuscation function is cryptographic, it cannot
3539 properly be called encryption because it has no key. Therefore,
3540 anybody motivated enough can re-implement it, or hack it out of the
3541 Puzzles source, and strip the obfuscation off one of these game IDs
3542 to see what lies beneath. (Indeed, they could usually do it much
3543 more easily than that, by entering the game ID into their own copy
3544 of the puzzle and hitting Solve.) The aim is not to protect against
3545 a determined attacker; the aim is simply to protect people who
3546 wanted to play the game honestly from \e{accidentally} spoiling
3549 The input argument \c{bmp} points at a piece of memory to be
3550 obfuscated. \c{bits} gives the length of the data. Note that that
3551 length is in \e{bits} rather than bytes: if you ask for obfuscation
3552 of a partial number of bytes, then you will get it. Bytes are
3553 considered to be used from the top down: thus, for example, setting
3554 \c{bits} to 10 will cover the whole of \cw{bmp[0]} and the \e{top
3555 two} bits of \cw{bmp[1]}. The remainder of a partially used byte is
3556 undefined (i.e. it may be corrupted by the function).
3558 The parameter \c{decode} is \cw{FALSE} for an encoding operation,
3559 and \cw{TRUE} for a decoding operation. Each is the inverse of the
3560 other. (There's no particular reason you shouldn't obfuscate by
3561 decoding and restore cleartext by encoding, if you really wanted to;
3562 it should still work.)
3564 The input bitmap is processed in place.
3566 \S{utils-bin2hex} \cw{bin2hex()}
3568 \c char *bin2hex(const unsigned char *in, int inlen);
3570 This function takes an input byte array and converts it into an
3571 ASCII string encoding those bytes in (lower-case) hex. It returns a
3572 dynamically allocated string containing that encoding.
3574 This function is useful for encoding the result of
3575 \cw{obfuscate_bitmap()} in printable ASCII for use in game IDs.
3577 \S{utils-hex2bin} \cw{hex2bin()}
3579 \c unsigned char *hex2bin(const char *in, int outlen);
3581 This function takes an ASCII string containing hex digits, and
3582 converts it back into a byte array of length \c{outlen}. If there
3583 aren't enough hex digits in the string, the contents of the
3584 resulting array will be undefined.
3586 This function is the inverse of \cw{bin2hex()}.
3588 \S{utils-game-mkhighlight} \cw{game_mkhighlight()}
3590 \c void game_mkhighlight(frontend *fe, float *ret,
3591 \c int background, int highlight, int lowlight);
3593 It's reasonably common for a puzzle game's graphics to use
3594 highlights and lowlights to indicate \q{raised} or \q{lowered}
3595 sections. Fifteen, Sixteen and Twiddle are good examples of this.
3597 Puzzles using this graphical style are running a risk if they just
3598 use whatever background colour is supplied to them by the front end,
3599 because that background colour might be too light to see any
3600 highlights on at all. (In particular, it's not unheard of for the
3601 front end to specify a default background colour of white.)
3603 Therefore, such puzzles can call this utility function from their
3604 \cw{colours()} routine (\k{backend-colours}). You pass it your front
3605 end handle, a pointer to the start of your return array, and three
3606 colour indices. It will:
3608 \b call \cw{frontend_default_colour()} (\k{frontend-default-colour})
3609 to fetch the front end's default background colour
3611 \b alter the brightness of that colour if it's unsuitable
3613 \b define brighter and darker variants of the colour to be used as
3614 highlights and lowlights
3616 \b write those results into the relevant positions in the \c{ret}
3619 Thus, \cw{ret[background*3]} to \cw{ret[background*3+2]} will be set
3620 to RGB values defining a sensible background colour, and similary
3621 \c{highlight} and \c{lowlight} will be set to sensible colours.
3623 \C{writing} How to write a new puzzle
3625 This chapter gives a guide to how to actually write a new puzzle:
3626 where to start, what to do first, how to solve common problems.
3628 The previous chapters have been largely composed of facts. This one
3631 \H{writing-editorial} Choosing a puzzle
3633 Before you start writing a puzzle, you have to choose one. Your
3634 taste in puzzle games is up to you, of course; and, in fact, you're
3635 probably reading this guide because you've \e{already} thought of a
3636 game you want to write. But if you want to get it accepted into the
3637 official Puzzles distribution, then there's a criterion it has to
3640 The current Puzzles editorial policy is that all games should be
3641 \e{fair}. A fair game is one which a player can only fail to
3642 complete through demonstrable lack of skill \dash that is, such that
3643 a better player in the same situation would have \e{known} to do
3644 something different.
3646 For a start, that means every game presented to the user must have
3647 \e{at least one solution}. Giving the unsuspecting user a puzzle
3648 which is actually impossible is not acceptable. (There is an
3649 exception: if the user has selected some non-default option which is
3650 clearly labelled as potentially unfair, \e{then} you're allowed to
3651 generate possibly insoluble puzzles, because the user isn't
3652 unsuspecting any more. Same Game and Mines both have options of this
3655 Also, this actually \e{rules out} games such as Klondike, or the
3656 normal form of Mahjong Solitaire. Those games have the property that
3657 even if there is a solution (i.e. some sequence of moves which will
3658 get from the start state to the solved state), the player doesn't
3659 necessarily have enough information to \e{find} that solution. In
3660 both games, it is possible to reach a dead end because you had an
3661 arbitrary choice to make and made it the wrong way. This violates
3662 the fairness criterion, because a better player couldn't have known
3663 they needed to make the other choice.
3665 (GNOME has a variant on Mahjong Solitaire which makes it fair: there
3666 is a Shuffle operation which randomly permutes all the remaining
3667 tiles without changing their positions, which allows you to get out
3668 of a sticky situation. Using this operation adds a 60-second penalty
3669 to your solution time, so it's to the player's advantage to try to
3670 minimise the chance of having to use it. It's still possible to
3671 render the game uncompletable if you end up with only two tiles
3672 vertically stacked, but that's easy to foresee and avoid using a
3673 shuffle operation. This form of the game \e{is} fair. Implementing
3674 it in Puzzles would require an infrastructure change so that the
3675 back end could communicate time penalties to the mid-end, but that
3676 would be easy enough.)
3678 Providing a \e{unique} solution is a little more negotiable; it
3679 depends on the puzzle. Solo would have been of unacceptably low
3680 quality if it didn't always have a unique solution, whereas Twiddle
3681 inherently has multiple solutions by its very nature and it would
3682 have been meaningless to even \e{suggest} making it uniquely
3683 soluble. Somewhere in between, Flip could reasonably be made to have
3684 unique solutions (by enforcing a zero-dimension kernel in every
3685 generated matrix) but it doesn't seem like a serious quality problem
3688 Of course, you don't \e{have} to care about all this. There's
3689 nothing stopping you implementing any puzzle you want to if you're
3690 happy to maintain your puzzle yourself, distribute it from your own
3691 web site, fork the Puzzles code completely, or anything like that.
3692 It's free software; you can do what you like with it. But any game
3693 that you want to be accepted into \e{my} Puzzles code base has to
3694 satisfy the fairness criterion, which means all randomly generated
3695 puzzles must have a solution (unless the user has deliberately
3696 chosen otherwise) and it must be possible \e{in theory} to find that
3697 solution without having to guess.
3699 \H{writing-gs} Getting started
3701 The simplest way to start writing a new puzzle is to copy
3702 \c{nullgame.c}. This is a template puzzle source file which does
3703 almost nothing, but which contains all the back end function
3704 prototypes and declares the back end data structure correctly. It is
3705 built every time the rest of Puzzles is built, to ensure that it
3706 doesn't get out of sync with the code and remains buildable.
3708 So start by copying \c{nullgame.c} into your new source file. Then
3709 you'll gradually add functionality until the very boring Null Game
3710 turns into your real game.
3712 Next you'll need to add your puzzle to the Makefiles, in order to
3713 compile it conveniently. \e{Do not edit the Makefiles}: they are
3714 created automatically by the script \c{mkfiles.pl}, from the file
3715 called \c{Recipe}. Edit \c{Recipe}, and then re-run \c{mkfiles.pl}.
3717 Also, don't forget to add your puzzle to \c{list.c}: if you don't,
3718 then it will still run fine on platforms which build each puzzle
3719 separately, but Mac OS X and other monolithic platforms will not
3720 include your new puzzle in their single binary.
3722 Once your source file is building, you can move on to the fun bit.
3724 \S{writing-generation} Puzzle generation
3726 Randomly generating instances of your puzzle is almost certain to be
3727 the most difficult part of the code, and also the task with the
3728 highest chance of turning out to be completely infeasible. Therefore
3729 I strongly recommend doing it \e{first}, so that if it all goes
3730 horribly wrong you haven't wasted any more time than you absolutely
3731 had to. What I usually do is to take an unmodified \c{nullgame.c},
3732 and start adding code to \cw{new_game_desc()} which tries to
3733 generate a puzzle instance and print it out using \cw{printf()}.
3734 Once that's working, \e{then} I start connecting it up to the return
3735 value of \cw{new_game_desc()}, populating other structures like
3736 \c{game_params}, and generally writing the rest of the source file.
3738 There are many ways to generate a puzzle which is known to be
3739 soluble. In this section I list all the methods I currently know of,
3740 in case any of them can be applied to your puzzle. (Not all of these
3741 methods will work, or in some cases even make sense, for all
3744 Some puzzles are mathematically tractable, meaning you can work out
3745 in advance which instances are soluble. Sixteen, for example, has a
3746 parity constraint in some settings which renders exactly half the
3747 game space unreachable, but it can be mathematically proved that any
3748 position not in that half \e{is} reachable. Therefore, Sixteen's
3749 grid generation simply consists of selecting at random from a well
3750 defined subset of the game space. Cube in its default state is even
3751 easier: \e{every} possible arrangement of the blue squares and the
3752 cube's starting position is soluble!
3754 Another option is to redefine what you mean by \q{soluble}. Black
3755 Box takes this approach. There are layouts of balls in the box which
3756 are completely indistinguishable from one another no matter how many
3757 beams you fire into the box from which angles, which would normally
3758 be grounds for declaring those layouts unfair; but fortunately,
3759 detecting that indistinguishability is computationally easy. So
3760 Black Box doesn't demand that your ball placements match its own; it
3761 merely demands that your ball placements be \e{indistinguishable}
3762 from the ones it was thinking of. If you have an ambiguous puzzle,
3763 then any of the possible answers is considered to be a solution.
3764 Having redefined the rules in that way, any puzzle is soluble again.
3766 Those are the simple techniques. If they don't work, you have to get
3769 One way to generate a soluble puzzle is to start from the solved
3770 state and make inverse moves until you reach a starting state. Then
3771 you know there's a solution, because you can just list the inverse
3772 moves you made and make them in the opposite order to return to the
3775 This method can be simple and effective for puzzles where you get to
3776 decide what's a starting state and what's not. In Pegs, for example,
3777 the generator begins with one peg in the centre of the board and
3778 makes inverse moves until it gets bored; in this puzzle, valid
3779 inverse moves are easy to detect, and \e{any} state that's reachable
3780 from the solved state by inverse moves is a reasonable starting
3781 position. So Pegs just continues making inverse moves until the
3782 board satisfies some criteria about extent and density, and then
3783 stops and declares itself done.
3785 For other puzzles, it can be a lot more difficult. Same Game uses
3786 this strategy too, and it's lucky to get away with it at all: valid
3787 inverse moves aren't easy to find (because although it's easy to
3788 insert additional squares in a Same Game position, it's difficult to
3789 arrange that \e{after} the insertion they aren't adjacent to any
3790 other squares of the same colour), so you're constantly at risk of
3791 running out of options and having to backtrack or start again. Also,
3792 Same Game grids never start off half-empty, which means you can't
3793 just stop when you run out of moves \dash you have to find a way to
3794 fill the grid up \e{completely}.
3796 The other way to generate a puzzle that's soluble is to start from
3797 the other end, and actually write a \e{solver}. This tends to ensure
3798 that a puzzle has a \e{unique} solution over and above having a
3799 solution at all, so it's a good technique to apply to puzzles for
3800 which that's important.
3802 One theoretical drawback of generating soluble puzzles by using a
3803 solver is that your puzzles are restricted in difficulty to those
3804 which the solver can handle. (Most solvers are not fully general:
3805 many sets of puzzle rules are NP-complete or otherwise nasty, so
3806 most solvers can only handle a subset of the theoretically soluble
3807 puzzles.) It's been my experience in practice, however, that this
3808 usually isn't a problem; computers are good at very different things
3809 from humans, and what the computer thinks is nice and easy might
3810 still be pleasantly challenging for a human. For example, when
3811 solving Dominosa puzzles I frequently find myself using a variety of
3812 reasoning techniques that my solver doesn't know about; in
3813 principle, therefore, I should be able to solve the puzzle using
3814 only those techniques it \e{does} know about, but this would involve
3815 repeatedly searching the entire grid for the one simple deduction I
3816 can make. Computers are good at this sort of exhaustive search, but
3817 it's been my experience that human solvers prefer to do more complex
3818 deductions than to spend ages searching for simple ones. So in many
3819 cases I don't find my own playing experience to be limited by the
3820 restrictions on the solver.
3822 (This isn't \e{always} the case. Solo is a counter-example;
3823 generating Solo puzzles using a simple solver does lead to
3824 qualitatively easier puzzles. Therefore I had to make the Solo
3825 solver rather more advanced than most of them.)
3827 There are several different ways to apply a solver to the problem of
3828 generating a soluble puzzle. I list a few of them below.
3830 The simplest approach is brute force: randomly generate a puzzle,
3831 use the solver to see if it's soluble, and if not, throw it away and
3832 try again until you get lucky. This is often a viable technique if
3833 all else fails, but it tends not to scale well: for many puzzle
3834 types, the probability of finding a uniquely soluble instance
3835 decreases sharply as puzzle size goes up, so this technique might
3836 work reasonably fast for small puzzles but take (almost) forever at
3837 larger sizes. Still, if there's no other alternative it can be
3838 usable: Pattern and Dominosa both use this technique. (However,
3839 Dominosa has a means of tweaking the randomly generated grids to
3840 increase the \e{probability} of them being soluble, by ruling out
3841 one of the most common ambiguous cases. This improved generation
3842 speed by over a factor of 10 on the highest preset!)
3844 An approach which can be more scalable involves generating a grid
3845 and then tweaking it to make it soluble. This is the technique used
3846 by Mines and also by Net: first a random puzzle is generated, and
3847 then the solver is run to see how far it gets. Sometimes the solver
3848 will get stuck; when that happens, examine the area it's having
3849 trouble with, and make a small random change in that area to allow
3850 it to make more progress. Continue solving (possibly even without
3851 restarting the solver), tweaking as necessary, until the solver
3852 finishes. Then restart the solver from the beginning to ensure that
3853 the tweaks haven't caused new problems in the process of solving old
3854 ones (which can sometimes happen).
3856 This strategy works well in situations where the usual solver
3857 failure mode is to get stuck in an easily localised spot. Thus it
3858 works well for Net and Mines, whose most common failure mode tends
3859 to be that most of the grid is fine but there are a few widely
3860 separated ambiguous sections; but it would work less well for
3861 Dominosa, in which the way you get stuck is to have scoured the
3862 whole grid and not found anything you can deduce \e{anywhere}. Also,
3863 it relies on there being a low probability that tweaking the grid
3864 introduces a new problem at the same time as solving the old one;
3865 Mines and Net also have the property that most of their deductions
3866 are local, so that it's very unlikely for a tweak to affect
3867 something half way across the grid from the location where it was
3868 applied. In Dominosa, by contrast, a lot of deductions use
3869 information about half the grid (\q{out of all the sixes, only one
3870 is next to a three}, which can depend on the values of up to 32 of
3871 the 56 squares in the default setting!), so this tweaking strategy
3872 would be rather less likely to work well.
3874 A more specialised strategy is that used in Solo and Slant. These
3875 puzzles have the property that they derive their difficulty from not
3876 presenting all the available clues. (In Solo's case, if all the
3877 possible clues were provided then the puzzle would already be
3878 solved; in Slant it would still require user action to fill in the
3879 lines, but it would present no challenge at all). Therefore, a
3880 simple generation technique is to leave the decision of which clues
3881 to provide until the last minute. In other words, first generate a
3882 random \e{filled} grid with all possible clues present, and then
3883 gradually remove clues for as long as the solver reports that it's
3884 still soluble. Unlike the methods described above, this technique
3885 \e{cannot} fail \dash once you've got a filled grid, nothing can
3886 stop you from being able to convert it into a viable puzzle.
3887 However, it wouldn't even be meaningful to apply this technique to
3888 (say) Pattern, in which clues can never be left out, so the only way
3889 to affect the set of clues is by altering the solution.
3891 (Unfortunately, Solo is complicated by the need to provide puzzles
3892 at varying difficulty levels. It's easy enough to generate a puzzle
3893 of \e{at most} a given level of difficulty; you just have a solver
3894 with configurable intelligence, and you set it to a given level and
3895 apply the above technique, thus guaranteeing that the resulting grid
3896 is solvable by someone with at most that much intelligence. However,
3897 generating a puzzle of \e{at least} a given level of difficulty is
3898 rather harder; if you go for \e{at most} Intermediate level, you're
3899 likely to find that you've accidentally generated a Trivial grid a
3900 lot of the time, because removing just one number is sufficient to
3901 take the puzzle from Trivial straight to Ambiguous. In that
3902 situation Solo has no remaining options but to throw the puzzle away
3905 A final strategy is to use the solver \e{during} puzzle
3906 construction: lay out a bit of the grid, run the solver to see what
3907 it allows you to deduce, and then lay out a bit more to allow the
3908 solver to make more progress. There are articles on the web that
3909 recommend constructing Sudoku puzzles by this method (which is
3910 completely the opposite way round to how Solo does it); for Sudoku
3911 it has the advantage that you get to specify your clue squares in
3912 advance (so you can have them make pretty patterns).
3914 Rectangles uses a strategy along these lines. First it generates a
3915 grid by placing the actual rectangles; then it has to decide where
3916 in each rectangle to place a number. It uses a solver to help it
3917 place the numbers in such a way as to ensure a unique solution. It
3918 does this by means of running a test solver, but it runs the solver
3919 \e{before} it's placed any of the numbers \dash which means the
3920 solver must be capable of coping with uncertainty about exactly
3921 where the numbers are! It runs the solver as far as it can until it
3922 gets stuck; then it narrows down the possible positions of a number
3923 in order to allow the solver to make more progress, and so on. Most
3924 of the time this process terminates with the grid fully solved, at
3925 which point any remaining number-placement decisions can be made at
3926 random from the options not so far ruled out. Note that unlike the
3927 Net/Mines tweaking strategy described above, this algorithm does not
3928 require a checking run after it completes: if it finishes
3929 successfully at all, then it has definitely produced a uniquely
3932 Most of the strategies described above are not 100% reliable. Each
3933 one has a failure rate: every so often it has to throw out the whole
3934 grid and generate a fresh one from scratch. (Solo's strategy would
3935 be the exception, if it weren't for the need to provide configurable
3936 difficulty levels.) Occasional failures are not a fundamental
3937 problem in this sort of work, however: it's just a question of
3938 dividing the grid generation time by the success rate (if it takes
3939 10ms to generate a candidate grid and 1/5 of them work, then it will
3940 take 50ms on average to generate a viable one), and seeing whether
3941 the expected time taken to \e{successfully} generate a puzzle is
3942 unacceptably slow. Dominosa's generator has a very low success rate
3943 (about 1 out of 20 candidate grids turn out to be usable, and if you
3944 think \e{that's} bad then go and look at the source code and find
3945 the comment showing what the figures were before the generation-time
3946 tweaks!), but the generator itself is very fast so this doesn't
3947 matter. Rectangles has a slower generator, but fails well under 50%
3950 So don't be discouraged if you have an algorithm that doesn't always
3951 work: if it \e{nearly} always works, that's probably good enough.
3952 The one place where reliability is important is that your algorithm
3953 must never produce false positives: it must not claim a puzzle is
3954 soluble when it isn't. It can produce false negatives (failing to
3955 notice that a puzzle is soluble), and it can fail to generate a
3956 puzzle at all, provided it doesn't do either so often as to become
3959 One last piece of advice: for grid-based puzzles, when writing and
3960 testing your generation algorithm, it's almost always a good idea
3961 \e{not} to test it initially on a grid that's square (i.e.
3962 \cw{w==h}), because if the grid is square then you won't notice if
3963 you mistakenly write \c{h} instead of \c{w} (or vice versa)
3964 somewhere in the code. Use a rectangular grid for testing, and any
3965 size of grid will be likely to work after that.
3967 \S{writing-textformats} Designing textual description formats
3969 Another aspect of writing a puzzle which is worth putting some
3970 thought into is the design of the various text description formats:
3971 the format of the game parameter encoding, the game description
3972 encoding, and the move encoding.
3974 The first two of these should be reasonably intuitive for a user to
3975 type in; so provide some flexibility where possible. Suppose, for
3976 example, your parameter format consists of two numbers separated by
3977 an \c{x} to specify the grid dimensions (\c{10x10} or \c{20x15}),
3978 and then has some suffixes to specify other aspects of the game
3979 type. It's almost always a good idea in this situation to arrange
3980 that \cw{decode_params()} can handle the suffixes appearing in any
3981 order, even if \cw{encode_params()} only ever generates them in one
3984 These formats will also be expected to be reasonably stable: users
3985 will expect to be able to exchange game IDs with other users who
3986 aren't running exactly the same version of your game. So make them
3987 robust and stable: don't build too many assumptions into the game ID
3988 format which will have to be changed every time something subtle
3989 changes in the puzzle code.
3991 \H{writing-howto} Common how-to questions
3993 This section lists some common things people want to do when writing
3994 a puzzle, and describes how to achieve them within the Puzzles
3997 \S{writing-howto-cursor} Drawing objects at only one position
3999 A common phenomenon is to have an object described in the
4000 \c{game_state} or the \c{game_ui} which can only be at one position.
4001 A cursor \dash probably specified in the \c{game_ui} \dash is a good
4004 In the \c{game_ui}, it would \e{obviously} be silly to have an array
4005 covering the whole game grid with a boolean flag stating whether the
4006 cursor was at each position. Doing that would waste space, would
4007 make it difficult to find the cursor in order to do anything with
4008 it, and would introduce the potential for synchronisation bugs in
4009 which you ended up with two cursors or none. The obviously sensible
4010 way to store a cursor in the \c{game_ui} is to have fields directly
4011 encoding the cursor's coordinates.
4013 However, it is a mistake to assume that the same logic applies to
4014 the \c{game_drawstate}. If you replicate the cursor position fields
4015 in the draw state, the redraw code will get very complicated. In the
4016 draw state, in fact, it \e{is} probably the right thing to have a
4017 cursor flag for every position in the grid. You probably have an
4018 array for the whole grid in the drawstate already (stating what is
4019 currently displayed in the window at each position); the sensible
4020 approach is to add a \q{cursor} flag to each element of that array.
4021 Then the main redraw loop will look something like this
4024 \c for (y = 0; y < h; y++) {
4025 \c for (x = 0; x < w; x++) {
4026 \c int value = state->symbol_at_position[y][x];
4027 \c if (x == ui->cursor_x && y == ui->cursor_y)
4029 \c if (ds->symbol_at_position[y][x] != value) {
4030 \c symbol_drawing_subroutine(dr, ds, x, y, value);
4031 \c ds->symbol_at_position[y][x] = value;
4036 This loop is very simple, pretty hard to get wrong, and
4037 \e{automatically} deals both with erasing the previous cursor and
4038 drawing the new one, with no special case code required.
4040 This type of loop is generally a sensible way to write a redraw
4041 function, in fact. The best thing is to ensure that the information
4042 stored in the draw state for each position tells you \e{everything}
4043 about what was drawn there. A good way to ensure that is to pass
4044 precisely the same information, and \e{only} that information, to a
4045 subroutine that does the actual drawing; then you know there's no
4046 additional information which affects the drawing but which you don't
4049 \S{writing-keyboard-cursor} Implementing a keyboard-controlled cursor
4051 It is often useful to provide a keyboard control method in a
4052 basically mouse-controlled game. A keyboard-controlled cursor is
4053 best implemented by storing its location in the \c{game_ui} (since
4054 if it were in the \c{game_state} then the user would have to
4055 separately undo every cursor move operation). So the procedure would
4058 \b Put cursor position fields in the \c{game_ui}.
4060 \b \cw{interpret_move()} responds to arrow keys by modifying the
4061 cursor position fields and returning \cw{""}.
4063 \b \cw{interpret_move()} responds to some sort of fire button by
4064 actually performing a move based on the current cursor location.
4066 \b You might want an additional \c{game_ui} field stating whether
4067 the cursor is currently visible, and having it disappear when a
4068 mouse action occurs (so that it doesn't clutter the display when not
4071 \b You might also want to automatically hide the cursor in
4072 \cw{changed_state()} when the current game state changes to one in
4073 which there is no move to make (which is the case in some types of
4076 \b \cw{redraw()} draws the cursor using the technique described in
4077 \k{writing-howto-cursor}.
4079 \S{writing-howto-dragging} Implementing draggable sprites
4081 Some games have a user interface which involves dragging some sort
4082 of game element around using the mouse. If you need to show a
4083 graphic moving smoothly over the top of other graphics, use a
4084 blitter (see \k{drawing-blitter} for the blitter API) to save the
4085 background underneath it. The typical scenario goes:
4087 \b Have a blitter field in the \c{game_drawstate}.
4089 \b Set the blitter field to \cw{NULL} in the game's
4090 \cw{new_drawstate()} function, since you don't yet know how big the
4091 piece of saved background needs to be.
4093 \b In the game's \cw{set_size()} function, once you know the size of
4094 the object you'll be dragging around the display and hence the
4095 required size of the blitter, actually allocate the blitter.
4097 \b In \cw{free_drawstate()}, free the blitter if it's not \cw{NULL}.
4099 \b In \cw{interpret_move()}, respond to mouse-down and mouse-drag
4100 events by updating some fields in the \cw{game_ui} which indicate
4101 that a drag is in progress.
4103 \b At the \e{very end} of \cw{redraw()}, after all other drawing has
4104 been done, draw the moving object if there is one. First save the
4105 background under the object in the blitter; then set a clip
4106 rectangle covering precisely the area you just saved (just in case
4107 anti-aliasing or some other error causes your drawing to go beyond
4108 the area you saved). Then draw the object, and call \cw{unclip()}.
4109 Finally, set a flag in the \cw{game_drawstate} that indicates that
4110 the blitter needs restoring.
4112 \b At the very start of \cw{redraw()}, before doing anything else at
4113 all, check the flag in the \cw{game_drawstate}, and if it says the
4114 blitter needs restoring then restore it. (Then clear the flag, so
4115 that this won't happen again in the next redraw if no moving object
4116 is drawn this time.)
4118 This way, you will be able to write the rest of the redraw function
4119 completely ignoring the dragged object, as if it were floating above
4120 your bitmap and being completely separate.
4122 \S{writing-ref-counting} Sharing large invariant data between all
4125 In some puzzles, there is a large amount of data which never changes
4126 between game states. The array of numbers in Dominosa is a good
4129 You \e{could} dynamically allocate a copy of that array in every
4130 \c{game_state}, and have \cw{dup_game()} make a fresh copy of it for
4131 every new \c{game_state}; but it would waste memory and time. A
4132 more efficient way is to use a reference-counted structure.
4134 \b Define a structure type containing the data in question, and also
4135 containing an integer reference count.
4137 \b Have a field in \c{game_state} which is a pointer to this
4140 \b In \cw{new_game()}, when creating a fresh game state at the start
4141 of a new game, create an instance of this structure, initialise it
4142 with the invariant data, and set its reference count to 1.
4144 \b In \cw{dup_game()}, rather than making a copy of the structure
4145 for the new game state, simply set the new game state to point at
4146 the same copy of the structure, and increment its reference count.
4148 \b In \cw{free_game()}, decrement the reference count in the
4149 structure pointed to by the game state; if the count reaches zero,
4152 This way, the invariant data will persist for only as long as it's
4153 genuinely needed; \e{as soon} as the last game state for a
4154 particular puzzle instance is freed, the invariant data for that
4155 puzzle will vanish as well. Reference counting is a very efficient
4156 form of garbage collection, when it works at all. (Which it does in
4157 this instance, of course, because there's no possibility of circular
4160 \S{writing-flash-types} Implementing multiple types of flash
4162 In some games you need to flash in more than one different way.
4163 Mines, for example, flashes white when you win, and flashes red when
4164 you tread on a mine and die.
4166 The simple way to do this is:
4168 \b Have a field in the \c{game_ui} which describes the type of flash.
4170 \b In \cw{flash_length()}, examine the old and new game states to
4171 decide whether a flash is required and what type. Write the type of
4172 flash to the \c{game_ui} field whenever you return non-zero.
4174 \b In \cw{redraw()}, when you detect that \c{flash_time} is
4175 non-zero, examine the field in \c{game_ui} to decide which type of
4178 \cw{redraw()} will never be called with \c{flash_time} non-zero
4179 unless \cw{flash_length()} was first called to tell the mid-end that
4180 a flash was required; so whenever \cw{redraw()} notices that
4181 \c{flash_time} is non-zero, you can be sure that the field in
4182 \c{game_ui} is correctly set.
4184 \S{writing-move-anim} Animating game moves
4186 A number of puzzle types benefit from a quick animation of each move
4189 For some games, such as Fifteen, this is particularly easy. Whenever
4190 \cw{redraw()} is called with \c{oldstate} non-\cw{NULL}, Fifteen
4191 simply compares the position of each tile in the two game states,
4192 and if the tile is not in the same place then it draws it some
4193 fraction of the way from its old position to its new position. This
4194 method copes automatically with undo.
4196 Other games are less obvious. In Sixteen, for example, you can't
4197 just draw each tile a fraction of the way from its old to its new
4198 position: if you did that, the end tile would zip very rapidly past
4199 all the others to get to the other end and that would look silly.
4200 (Worse, it would look inconsistent if the end tile was drawn on top
4201 going one way and on the bottom going the other way.)
4203 A useful trick here is to define a field or two in the game state
4204 that indicates what the last move was.
4206 \b Add a \q{last move} field to the \c{game_state} (or two or more
4207 fields if the move is complex enough to need them).
4209 \b \cw{new_game()} initialises this field to a null value for a new
4212 \b \cw{execute_move()} sets up the field to reflect the move it just
4215 \b \cw{redraw()} now needs to examine its \c{dir} parameter. If
4216 \c{dir} is positive, it determines the move being animated by
4217 looking at the last-move field in \c{newstate}; but if \c{dir} is
4218 negative, it has to look at the last-move field in \c{oldstate}, and
4219 invert whatever move it finds there.
4221 Note also that Sixteen needs to store the \e{direction} of the move,
4222 because you can't quite determine it by examining the row or column
4223 in question. You can in almost all cases, but when the row is
4224 precisely two squares long it doesn't work since a move in either
4225 direction looks the same. (You could argue that since moving a
4226 2-element row left and right has the same effect, it doesn't matter
4227 which one you animate; but in fact it's very disorienting to click
4228 the arrow left and find the row moving right, and almost as bad to
4229 undo a move to the right and find the game animating \e{another}
4232 \S{writing-conditional-anim} Animating drag operations
4234 In Untangle, moves are made by dragging a node from an old position
4235 to a new position. Therefore, at the time when the move is initially
4236 made, it should not be animated, because the node has already been
4237 dragged to the right place and doesn't need moving there. However,
4238 it's nice to animate the same move if it's later undone or redone.
4239 This requires a bit of fiddling.
4241 The obvious approach is to have a flag in the \c{game_ui} which
4242 inhibits move animation, and to set that flag in
4243 \cw{interpret_move()}. The question is, when would the flag be reset
4244 again? The obvious place to do so is \cw{changed_state()}, which
4245 will be called once per move. But it will be called \e{before}
4246 \cw{anim_length()}, so if it resets the flag then \cw{anim_length()}
4247 will never see the flag set at all.
4249 The solution is to have \e{two} flags in a queue.
4251 \b Define two flags in \c{game_ui}; let's call them \q{current} and
4254 \b Set both to \cw{FALSE} in \c{new_ui()}.
4256 \b When a drag operation completes in \cw{interpret_move()}, set the
4257 \q{next} flag to \cw{TRUE}.
4259 \b Every time \cw{changed_state()} is called, set the value of
4260 \q{current} to the value in \q{next}, and then set the value of
4261 \q{next} to \cw{FALSE}.
4263 \b That way, \q{current} will be \cw{TRUE} \e{after} a call to
4264 \cw{changed_state()} if and only if that call to
4265 \cw{changed_state()} was the result of a drag operation processed by
4266 \cw{interpret_move()}. Any other call to \cw{changed_state()}, due
4267 to an Undo or a Redo or a Restart or a Solve, will leave \q{current}
4270 \b So now \cw{anim_length()} can request a move animation if and
4271 only if the \q{current} flag is \e{not} set.
4273 \S{writing-cheating} Inhibiting the victory flash when Solve is used
4275 Many games flash when you complete them, as a visual congratulation
4276 for having got to the end of the puzzle. It often seems like a good
4277 idea to disable that flash when the puzzle is brought to a solved
4278 state by means of the Solve operation.
4280 This is easily done:
4282 \b Add a \q{cheated} flag to the \c{game_state}.
4284 \b Set this flag to \cw{FALSE} in \cw{new_game()}.
4286 \b Have \cw{solve()} return a move description string which clearly
4287 identifies the move as a solve operation.
4289 \b Have \cw{execute_move()} respond to that clear identification by
4290 setting the \q{cheated} flag in the returned \c{game_state}. The
4291 flag will then be propagated to all subsequent game states, even if
4292 the user continues fiddling with the game after it is solved.
4294 \b \cw{flash_length()} now returns non-zero if \c{oldstate} is not
4295 completed and \c{newstate} is, \e{and} neither state has the
4296 \q{cheated} flag set.
4298 \H{writing-testing} Things to test once your puzzle is written
4300 Puzzle implementations written in this framework are self-testing as
4301 far as I could make them.
4303 Textual game and move descriptions, for example, are generated and
4304 parsed as part of the normal process of play. Therefore, if you can
4305 make moves in the game \e{at all} you can be reasonably confident
4306 that the mid-end serialisation interface will function correctly and
4307 you will be able to save your game. (By contrast, if I'd stuck with
4308 a single \cw{make_move()} function performing the jobs of both
4309 \cw{interpret_move()} and \cw{execute_move()}, and had separate
4310 functions to encode and decode a game state in string form, then
4311 those functions would not be used during normal play; so they could
4312 have been completely broken, and you'd never know it until you tried
4313 to save the game \dash which would have meant you'd have to test
4314 game saving \e{extensively} and make sure to test every possible
4315 type of game state. As an added bonus, doing it the way I did leads
4316 to smaller save files.)
4318 There is one exception to this, which is the string encoding of the
4319 \c{game_ui}. Most games do not store anything permanent in the
4320 \c{game_ui}, and hence do not need to put anything in its encode and
4321 decode functions; but if there is anything in there, you do need to
4322 test game loading and saving to ensure those functions work
4325 It's also worth testing undo and redo of all operations, to ensure
4326 that the redraw and the animations (if any) work properly. Failing
4327 to animate undo properly seems to be a common error.
4329 Other than that, just use your common sense.