chiark / gitweb /
sgt-puzzles.git
15 years agoKeyboard control patch for Pattern, from James H.
Simon Tatham [Tue, 27 Jan 2009 18:30:46 +0000 (18:30 +0000)]
Keyboard control patch for Pattern, from James H.

[originally from svn r8435]

15 years agoKeyboard control for Map, from James H.
Simon Tatham [Tue, 27 Jan 2009 18:26:10 +0000 (18:26 +0000)]
Keyboard control for Map, from James H.

[originally from svn r8434]

15 years agoReorder if statements in Unequal's interpret_move() so that presses
Simon Tatham [Mon, 26 Jan 2009 22:28:17 +0000 (22:28 +0000)]
Reorder if statements in Unequal's interpret_move() so that presses
of 'h' and 'm' are treated as digits if a square is selected, and
only treated as special commands otherwise. This renders very large
games (just about) playable.

Idea from Ben Hutchings's collection of Debian patches, although I
had to redo his (trivial) patch myself since the code has changed
recently.

(Addendum after committing: hmm, I see Jacob already applied the
original version of the patch a while back. Looks as if the recent
keyboard control change reintroduced the problem. Still, re-fixed
now.)

[originally from svn r8433]

15 years agoPatch from James H to add keyboard control in Sixteen and Netslide
Simon Tatham [Mon, 26 Jan 2009 19:14:44 +0000 (19:14 +0000)]
Patch from James H to add keyboard control in Sixteen and Netslide
(and also belatedly document the keyboard support in Unequal).

[originally from svn r8432]

15 years agoSwitch over to using the new-style GtkFileChooser in place of the
Simon Tatham [Mon, 26 Jan 2009 19:11:34 +0000 (19:11 +0000)]
Switch over to using the new-style GtkFileChooser in place of the
deprecated GtkFileSelection, at least when the latter is available.
Patch mostly due to Ori Avtalion.

[originally from svn r8431]

15 years agoAdd an automatic check for HAVE_SENSIBLE_ABSOLUTE_SIZE_FUNCTION by
Simon Tatham [Sat, 24 Jan 2009 17:58:42 +0000 (17:58 +0000)]
Add an automatic check for HAVE_SENSIBLE_ABSOLUTE_SIZE_FUNCTION by
using PANGO_CHECK_VERSION. (Irritatingly, the latter is not
available in all versions of Pango, so we have to be cautious about
using it at all. *sigh*)

[originally from svn r8430]

15 years agoKeyboard control patch for Pegs, from James H.
Simon Tatham [Tue, 20 Jan 2009 18:28:28 +0000 (18:28 +0000)]
Keyboard control patch for Pegs, from James H.

[originally from svn r8419]

15 years agoKeyboard control patch for Solo, from James H.
Simon Tatham [Mon, 19 Jan 2009 18:38:29 +0000 (18:38 +0000)]
Keyboard control patch for Solo, from James H.

[originally from svn r8418]

15 years agoKeyboard control patch for Slant, from James H.
Simon Tatham [Fri, 16 Jan 2009 18:52:18 +0000 (18:52 +0000)]
Keyboard control patch for Slant, from James H.

[originally from svn r8417]

15 years agoKeyboard cursor support for Tents, from James H.
Simon Tatham [Fri, 16 Jan 2009 18:50:22 +0000 (18:50 +0000)]
Keyboard cursor support for Tents, from James H.

[originally from svn r8416]

15 years agor8396 introduced midend_tilesize(). Document it.
Simon Tatham [Fri, 16 Jan 2009 18:45:01 +0000 (18:45 +0000)]
r8396 introduced midend_tilesize(). Document it.

[originally from svn r8415]
[r8396 == 7888d8db677e25d2401914e6aacb1be02301d15f]

15 years agoKeyboard cursor support for Rectangles, from James H.
Simon Tatham [Thu, 15 Jan 2009 18:20:02 +0000 (18:20 +0000)]
Keyboard cursor support for Rectangles, from James H.

[originally from svn r8414]

15 years agoKeyboard cursor support in Galaxies, by James H.
Simon Tatham [Thu, 15 Jan 2009 18:18:21 +0000 (18:18 +0000)]
Keyboard cursor support in Galaxies, by James H.

[originally from svn r8413]

15 years agoKeyboard interface for Filling, from James H.
Simon Tatham [Wed, 14 Jan 2009 20:44:25 +0000 (20:44 +0000)]
Keyboard interface for Filling, from James H.

[originally from svn r8412]

15 years agoPatch from James H to fix some off-by-one errors in Guess's click
Simon Tatham [Mon, 12 Jan 2009 20:30:12 +0000 (20:30 +0000)]
Patch from James H to fix some off-by-one errors in Guess's click
rectangle checking, preventing array bounds violation.

[originally from svn r8409]

15 years agoPatch from James H which adds keyboard cursor support to Unequal,
Simon Tatham [Mon, 12 Jan 2009 20:23:56 +0000 (20:23 +0000)]
Patch from James H which adds keyboard cursor support to Unequal,
and also updates the docs for both that and the Mines cursor support
in r8402.

[originally from svn r8408]
[r8402 == f20847354cb6335fd349204f16021a72e2956cce]

15 years agoPatches from James H to add or improve arrow-key-driven cursors for
Simon Tatham [Thu, 8 Jan 2009 18:28:32 +0000 (18:28 +0000)]
Patches from James H to add or improve arrow-key-driven cursors for
some puzzles. (Light Up's and Net's are merely polished a bit, but
Mines acquires a new one.)

[originally from svn r8402]

15 years agoPatch from Lambros to make the Normal difficulty level easier, since
Simon Tatham [Wed, 7 Jan 2009 23:07:11 +0000 (23:07 +0000)]
Patch from Lambros to make the Normal difficulty level easier, since
people have generally seemed to think Loopy is one of the more
difficult puzzles in the collection. There's a new level called
Tricky, between Normal and Hard, which is equivalent to the old
Normal.

[originally from svn r8398]

15 years agoStandalone solver for Loopy. Bit half-hearted, since the solver
Simon Tatham [Wed, 7 Jan 2009 21:55:21 +0000 (21:55 +0000)]
Standalone solver for Loopy. Bit half-hearted, since the solver
doesn't have diagnostics embedded and the ASCII formatter can't
print non-square puzzles anyway; but it can grade difficulty, which
is what I most immediately want it for.

[originally from svn r8397]

15 years agoPatch from James H to enable a single monolithic binary to be built
Simon Tatham [Tue, 6 Jan 2009 23:26:18 +0000 (23:26 +0000)]
Patch from James H to enable a single monolithic binary to be built
alongside the individual puzzle binaries, on Windows only. (MacOS
already has it, of course; Unix would require about as much work
again.)

[originally from svn r8396]

15 years agoBring the unfinished .R files into line with the current conventions.
Simon Tatham [Tue, 6 Jan 2009 23:21:42 +0000 (23:21 +0000)]
Bring the unfinished .R files into line with the current conventions.

[originally from svn r8395]

15 years agoDon't ignore the return code of fwrite(). (Slight modification of a
Simon Tatham [Fri, 12 Dec 2008 19:49:05 +0000 (19:49 +0000)]
Don't ignore the return code of fwrite(). (Slight modification of a
patch from Phil Bordelon.)

[originally from svn r8373]

15 years agoAdd missing call to gtk_selection_clear_targets(), without which the
Simon Tatham [Mon, 1 Dec 2008 23:04:17 +0000 (23:04 +0000)]
Add missing call to gtk_selection_clear_targets(), without which the
list of selection targets offered by GTK Puzzles grows an extra copy
of each of the three supported text formats every time the user
makes a selection!

[originally from svn r8365]

15 years agoStand-alone command-line interface to the obfuscate_bitmap()
Simon Tatham [Sat, 29 Nov 2008 13:51:36 +0000 (13:51 +0000)]
Stand-alone command-line interface to the obfuscate_bitmap()
function. Useful in conjunction with the new --save option to
generate lots of games, extract the aux strings from the game
generator, and de-obfuscate them in order to measure statistical
properties of their solutions.

[originally from svn r8352]

15 years agoThe other day I found it useful for a (silly) special purpose to
Simon Tatham [Sat, 29 Nov 2008 08:50:50 +0000 (08:50 +0000)]
The other day I found it useful for a (silly) special purpose to
generate a thousand saved-game files of randomly generated puzzles.
On the general principle that if I find it useful someone else
conceivably might too, add a --save option to the Unix puzzle
binaries, for use with --generate in much the same way as --print.

[originally from svn r8351]

15 years agoFix parenthesis problem in Light Up solver. Should fix Debian bug
Simon Tatham [Fri, 28 Nov 2008 19:33:40 +0000 (19:33 +0000)]
Fix parenthesis problem in Light Up solver. Should fix Debian bug
#505359.

[originally from svn r8342]

15 years agoLet's have some more presets in here: 3x3 is much _too_ short and
Simon Tatham [Fri, 28 Nov 2008 18:33:39 +0000 (18:33 +0000)]
Let's have some more presets in here: 3x3 is much _too_ short and
easy compared to 6x6.

[originally from svn r8340]

15 years agoPatch from Lee Dowling to implement mouse control in Sokoban, along
Simon Tatham [Sun, 16 Nov 2008 15:47:55 +0000 (15:47 +0000)]
Patch from Lee Dowling to implement mouse control in Sokoban, along
pretty much the same lines as Cube and Inertia.

[originally from svn r8301]

15 years agoTake out some lurking "nullgame" holdovers from unfinished puzzle
Simon Tatham [Sun, 16 Nov 2008 15:42:32 +0000 (15:42 +0000)]
Take out some lurking "nullgame" holdovers from unfinished puzzle
source files, in case they cause trouble. Spotted by Lee Dowling.

[originally from svn r8300]

15 years agoPatches from Lee Dowling to make Light Up and Net use the
Simon Tatham [Sun, 16 Nov 2008 15:37:58 +0000 (15:37 +0000)]
Patches from Lee Dowling to make Light Up and Net use the
CURSOR_SELECT2 button (to, respectively, toggle a "definitely not
light" dot and to rotate in the opposite direction from
CURSOR_SELECT).

[originally from svn r8299]

15 years agoCouple of solving-related mid-end tweaks. Firstly, when we generate
Simon Tatham [Sun, 16 Nov 2008 15:28:28 +0000 (15:28 +0000)]
Couple of solving-related mid-end tweaks. Firstly, when we generate
a game which comes with an aux string, we immediately self-test that
string by passing it to solve() and test by assertion that it
succeeded. So a bug in a back end which intermittently generates
malformed aux strings will be detected as soon as it occurs, instead
of only if the user happens to use the Solve operation on a
particular game in which it happened.

Secondly, Ctrl-S now (undocumentedly) triggers the Solve operation,
on the general principle that keyboard shortcuts tend to come in
handy, and on the specific principle that if you want to look at
lots of solved grids in quick succession (say, when observing their
general shape and nature to see if your generation algorithm was
good or not) it's handy to have a quick way of getting to them.

[originally from svn r8298]

15 years agoCheck return values from fwrite when saving files.
Simon Tatham [Tue, 4 Nov 2008 23:02:07 +0000 (23:02 +0000)]
Check return values from fwrite when saving files.

[originally from svn r8278]

15 years agoPatch from Lambros to improve the generality of path-generation. In
Simon Tatham [Tue, 4 Nov 2008 21:39:59 +0000 (21:39 +0000)]
Patch from Lambros to improve the generality of path-generation. In
particular, Great Hexagonal tilings previously had virtually every
(if not _actually_ every) hexagon on the inside of the path, and now
don't.

[originally from svn r8277]

15 years agoPatch from Chris Boyle to prevent Solo's inter-block dividing lines
Simon Tatham [Sun, 2 Nov 2008 14:29:41 +0000 (14:29 +0000)]
Patch from Chris Boyle to prevent Solo's inter-block dividing lines
from becoming indistinguishable from the intra-block ones at low
tile sizes.

[originally from svn r8259]

15 years agoApply "103_fix-unequal-digit-h.diff" from the Debian package:
Jacob Nevins [Sun, 5 Oct 2008 12:22:37 +0000 (12:22 +0000)]
Apply "103_fix-unequal-digit-h.diff" from the Debian package:
Unequal 18x18 or above was unplayable due to a clash with the undocumented
"H" (hint) key. Resolve the clash by making the hint function only fire
when no square is selected.

[originally from svn r8200]

15 years agoRemove strange punctuation.
Jacob Nevins [Sat, 4 Oct 2008 21:49:44 +0000 (21:49 +0000)]
Remove strange punctuation.

[originally from svn r8199]

15 years agoCut-and-paste error which was preventing any drop-down list in the
Simon Tatham [Fri, 19 Sep 2008 07:31:52 +0000 (07:31 +0000)]
Cut-and-paste error which was preventing any drop-down list in the
custom game configuration code from working in the Java applets.

[originally from svn r8192]

15 years agoLambros points out that trying to generate a 3x3 Cairo Easy grid
Simon Tatham [Thu, 18 Sep 2008 18:19:55 +0000 (18:19 +0000)]
Lambros points out that trying to generate a 3x3 Cairo Easy grid
spins forever, but observes that raising the limit to 4x4 across all
grid types is not good for the complex grids like great-hexagonal.
Introduce per-grid minimum sizes using mad macro trickery.

(In fact, for each grid type I've put in a minimum size which _both_
dimensions must equal or exceed, plus another minimum size which _at
least one_ must equal or exceed; that permits both 3x4 and 4x3 Cairo
while disallowing 3x3.)

[originally from svn r8191]

15 years agoPatch from Lambros implementing error highlighting in Loopy.
Simon Tatham [Thu, 18 Sep 2008 15:33:13 +0000 (15:33 +0000)]
Patch from Lambros implementing error highlighting in Loopy.

[originally from svn r8190]

15 years agoYet another complete rewrite of Slant's loop detection during
Simon Tatham [Wed, 17 Sep 2008 16:43:36 +0000 (16:43 +0000)]
Yet another complete rewrite of Slant's loop detection during
gameplay. Having tried methods based on using the slashes to define
a dsf on grid vertices, and also methods based on tracing round the
loops using conventional (non-dsf-based) graph theory, it occurred
to me the other day that there's a far simpler technique involving
connectivity. A loop is precisely that which causes the playing area
to become disconnected; so what we do now is to go through and build
a dsf describing connectedness of the _area_ of the grid rather than
the vertices. That divides the area into its maximal connected
components, and then we can trivially identify every edge that's
part of a loop by noticing that it separates two nonequivalent
pieces of space. The resulting algorithm is half the size of the old
one, and it's much easier to be confident of its correctness.

(Having said which, there will doubtless turn out to be an
embarrassing bug in it, but I haven't found it yet.)

[originally from svn r8187]

15 years agoCosmetic: fix mismatch between game_compute_size() and game_redraw()
Jacob Nevins [Tue, 16 Sep 2008 23:51:57 +0000 (23:51 +0000)]
Cosmetic: fix mismatch between game_compute_size() and game_redraw()
(was causing unwanted "drop-shadow" type effect).

[originally from svn r8186]

15 years agoLambros provides this workaround for a compiler warning on his
Simon Tatham [Sun, 14 Sep 2008 08:52:59 +0000 (08:52 +0000)]
Lambros provides this workaround for a compiler warning on his
Ubuntu system. I'm inclined to think the real problem is in his gtk
headers, but this is a harmless enough change to avoid hassle.

[originally from svn r8181]

15 years agoTypo spotted by James H.
Simon Tatham [Sat, 13 Sep 2008 19:21:53 +0000 (19:21 +0000)]
Typo spotted by James H.

[originally from svn r8180]

15 years agoOops, left this out of r8178: having defined COMBINED everywhere in
Simon Tatham [Sat, 13 Sep 2008 19:18:42 +0000 (19:18 +0000)]
Oops, left this out of r8178: having defined COMBINED everywhere in
the puzzles, we can now remove it from the OS X makefile section.

[originally from svn r8179]
[r8178 == 43eafe1fdf356c0c1c88936ffa79c83291973b5d]

15 years agoChange to the handling of -DCOMBINED in the makefiles. Instead of
Simon Tatham [Sat, 13 Sep 2008 19:17:26 +0000 (19:17 +0000)]
Change to the handling of -DCOMBINED in the makefiles. Instead of
defining it centrally per port, I think it's neater to define it for
each puzzle when adding that puzzle to the ALL list - because those
front ends which take -DCOMBINED are precisely those which use ALL.
In particular, this change opens up the possibility of compiling
both individual puzzles _and_ a combined monolith within the same
makefile.

[originally from svn r8178]

15 years agoPatch from James H providing lots more paranoid casting. Also one
Simon Tatham [Sat, 13 Sep 2008 18:29:20 +0000 (18:29 +0000)]
Patch from James H providing lots more paranoid casting. Also one
actual behaviour change: Untangle now permits dragging with the
right mouse button, which has exactly the same effect as it does
with the left. (Harmless on desktop platforms, but helpful when
"right-click" is achieved by press-and-hold; now the drag takes
place even if you hesitate first.)

[originally from svn r8177]

15 years agoPatch from James H to centralise some generally useful cursor-
Simon Tatham [Sat, 13 Sep 2008 18:26:53 +0000 (18:26 +0000)]
Patch from James H to centralise some generally useful cursor-
handling functionality into misc.c.

[originally from svn r8176]

15 years agoSince the lack of this has caused portability issues in the past:
Simon Tatham [Sat, 13 Sep 2008 18:25:19 +0000 (18:25 +0000)]
Since the lack of this has caused portability issues in the past:
add "-ansi -pedantic" to the main Unix makefile, and clean up a few
minor problems pointed out thereby.

[originally from svn r8175]

15 years agoPatch from James H to make new-Loopy port more easily.
Simon Tatham [Wed, 10 Sep 2008 21:44:23 +0000 (21:44 +0000)]
Patch from James H to make new-Loopy port more easily.

[originally from svn r8174]

15 years agoHaving played new-Loopy a bit recently, I've had occasion to think a
Simon Tatham [Sun, 7 Sep 2008 10:02:40 +0000 (10:02 +0000)]
Having played new-Loopy a bit recently, I've had occasion to think a
bit harder about advanced solver techniques. Expand the comment at
the top of the file.

[originally from svn r8167]

15 years agoHow did I manage to check this in without actually trying to build
Simon Tatham [Sun, 7 Sep 2008 08:35:52 +0000 (08:35 +0000)]
How did I manage to check this in without actually trying to build
on Windows at all?! Fix some departures from the C standard, mostly
declaring variables after a statement has already been issued in the
same block. MSVC is picky about this where gcc is forgiving, and TBH
I'd change the latter given the choice.

[originally from svn r8166]

15 years agoNew Loopy save file, compatible with the new Loopy.
Simon Tatham [Sat, 6 Sep 2008 21:24:21 +0000 (21:24 +0000)]
New Loopy save file, compatible with the new Loopy.

[originally from svn r8165]

15 years agoNearly forgot: Lambros definitely deserves a place in the copyright
Simon Tatham [Sat, 6 Sep 2008 17:38:43 +0000 (17:38 +0000)]
Nearly forgot: Lambros definitely deserves a place in the copyright
statement!

[originally from svn r8164]

15 years agoDon't call changed_preset() until after we've initialised
Simon Tatham [Sat, 6 Sep 2008 17:33:04 +0000 (17:33 +0000)]
Don't call changed_preset() until after we've initialised
fe->copy_menu_item.

[originally from svn r8163]

15 years agoCompletely re-engineered version of Loopy, courtesy of Lambros
Simon Tatham [Sat, 6 Sep 2008 15:19:47 +0000 (15:19 +0000)]
Completely re-engineered version of Loopy, courtesy of Lambros
Lambrou. Now capable of handling triangular and hexagonal grids as
well as square ones, and then a number of semiregular plane tilings
and duals of semiregular ones. In fact, most of the solver code
supports an _arbitrary_ planar graph (well, provided both the graph
and its dual have no self-edges), so it could easily be extended
further with only a little more effort.

[originally from svn r8162]

15 years agoNew infrastructure feature. Games are now permitted to be
Simon Tatham [Sat, 6 Sep 2008 09:27:56 +0000 (09:27 +0000)]
New infrastructure feature. Games are now permitted to be
_conditionally_ able to format the current puzzle as text to be sent
to the clipboard. For instance, if a game were to support playing on
a square grid and on other kinds of grid such as hexagonal, then it
might reasonably feel that only the former could be sensibly
rendered in ASCII art; so it can now arrange for the "Copy" menu
item to be greyed out depending on the game_params.

To do this I've introduced a new backend function
(can_format_as_text_now()), and renamed the existing static backend
field "can_format_as_text" to "can_format_as_text_ever". The latter
will cause compile errors for anyone maintaining a third-party front
end; if any such person is reading this, I apologise to them for the
inconvenience, but I did do it deliberately so that they'd know to
update their front end.

As yet, no checked-in game actually uses this feature; all current
games can still either copy always or copy never.

[originally from svn r8161]

15 years agoCorrection from James H: sqrt(0) shouldn't occur any more than
Simon Tatham [Sat, 5 Jul 2008 22:07:35 +0000 (22:07 +0000)]
Correction from James H: sqrt(0) shouldn't occur any more than
sqrt(1) should.

[originally from svn r8108]

15 years agoMore operations and bug fixes from James H.
Simon Tatham [Sat, 5 Jul 2008 15:40:43 +0000 (15:40 +0000)]
More operations and bug fixes from James H.

[originally from svn r8107]

15 years agoRemove rogue diagnostic.
Simon Tatham [Sat, 5 Jul 2008 13:32:28 +0000 (13:32 +0000)]
Remove rogue diagnostic.

[originally from svn r8106]

15 years agoAdd a build version designation to the NestedVM build, after Jacob
Simon Tatham [Sat, 5 Jul 2008 13:31:59 +0000 (13:31 +0000)]
Add a build version designation to the NestedVM build, after Jacob
pointed out that Help > About in the Java applets on my website
currently reports "Unidentified build".

[originally from svn r8105]

15 years agoBuild the Java versions of the puzzles automatically as part of the
Simon Tatham [Fri, 27 Jun 2008 17:28:32 +0000 (17:28 +0000)]
Build the Java versions of the puzzles automatically as part of the
build process. Also update the new-puzzle checklist to make sure I
set up and test the Java applet for any new game I add.

[originally from svn r8096]

15 years agoHandle a <param name="game_id"> by passing it in to the C side as
Simon Tatham [Thu, 26 Jun 2008 19:09:07 +0000 (19:09 +0000)]
Handle a <param name="game_id"> by passing it in to the C side as
argv[1], which in turn feeds it into the midend as a game ID. This
can of course take any of the forms supported by the native C
puzzles: a pure game parameter string, a params:description specific
game ID, or a params#seed random game ID.

[originally from svn r8095]

15 years agoThe Java console keeps showing up error reports due to being asked
Simon Tatham [Thu, 26 Jun 2008 19:07:44 +0000 (19:07 +0000)]
The Java console keeps showing up error reports due to being asked
to resize the puzzle to zero size. Ignore all such requests, in the
assumption that a more sensible resize will be along soon enough
(which does seem to happen, though I haven't debugged the NestedVM
front end hard enough to figure out why the bogus resizes happen in
the first place).

[originally from svn r8094]

15 years agoJames H has helpfully provided yet more silly operators for the -A
Simon Tatham [Tue, 24 Jun 2008 20:58:35 +0000 (20:58 +0000)]
James H has helpfully provided yet more silly operators for the -A
mode. I think some user-defined ruleset configuration options are
now required...

[originally from svn r8092]

15 years agoAn option to enable a debugging mode for the BFS.
Simon Tatham [Mon, 23 Jun 2008 17:34:56 +0000 (17:34 +0000)]
An option to enable a debugging mode for the BFS.

[originally from svn r8091]

15 years agoMichael Schierl's patch to compile the puzzles as Java applets using
Simon Tatham [Tue, 10 Jun 2008 20:35:17 +0000 (20:35 +0000)]
Michael Schierl's patch to compile the puzzles as Java applets using
NestedVM. Wow!

[originally from svn r8064]

15 years agoJust noticed that the return value of midend_process_key() wasn't
Simon Tatham [Tue, 10 Jun 2008 17:24:09 +0000 (17:24 +0000)]
Just noticed that the return value of midend_process_key() wasn't
documented.

[originally from svn r8062]

15 years agoNew -A mode permitting even madder operators, and also -m to try to
Simon Tatham [Mon, 9 Jun 2008 18:28:03 +0000 (18:28 +0000)]
New -A mode permitting even madder operators, and also -m to try to
print all possible paths to a value. The latter has a lot of
de-duplication left to be done, due to multiple evaluation orders.

[originally from svn r8061]

16 years agoShamelessly pander to compilers whose data flow warning systems
Simon Tatham [Mon, 14 Apr 2008 11:32:06 +0000 (11:32 +0000)]
Shamelessly pander to compilers whose data flow warning systems
insist that a variable should be initialised in all branches of an
if, instead of just all the non-assertion-failing ones.

[originally from svn r7989]

16 years agoNow that we're highlighting the currently selected preset in the
Simon Tatham [Thu, 10 Apr 2008 11:11:33 +0000 (11:11 +0000)]
Now that we're highlighting the currently selected preset in the
Type menu, it looks faintly silly that Fifteen doesn't have any
presets other than Custom: you open a Fifteen window in its default
state, and the Type menu appears to be telling you it has a custom
size! Fixed by adding a preset for the default parameters.

I'd quite like to fix this properly by revamping the presets
mechanism in a way that _enforces_ that there must always be a
preset which matches the default parameters, but that's more fiddly
than it sounds. For the moment, this change fixes the only
externally visible infelicity in the current game set.

[originally from svn r7983]

16 years agoImplement tick marks in the Type menu on Windows. Now all my front
Simon Tatham [Wed, 9 Apr 2008 14:57:20 +0000 (14:57 +0000)]
Implement tick marks in the Type menu on Windows. Now all my front
ends have got them.

[originally from svn r7982]

16 years agoImplement tick marks in the Type menu on OS X.
Simon Tatham [Wed, 9 Apr 2008 14:36:08 +0000 (14:36 +0000)]
Implement tick marks in the Type menu on OS X.

[originally from svn r7981]

16 years agoNew feature in midend.c which allows us to ask for the number of the
Simon Tatham [Tue, 8 Apr 2008 16:25:39 +0000 (16:25 +0000)]
New feature in midend.c which allows us to ask for the number of the
currently selected preset, if any. I've used this in the GTK front
end to have the Type menu mark the currently selected menu item.
(After considerable beating of GTK with sticks, I might add. Grr.)
Currently the same UI feature is not yet supported on Windows or
MacOS, but I hope to do those too at some point if it's feasible.

[originally from svn r7980]

16 years agoHaving got Jigsaw mode generation working at reasonable speed, we
Simon Tatham [Tue, 8 Apr 2008 10:30:18 +0000 (10:30 +0000)]
Having got Jigsaw mode generation working at reasonable speed, we
can now productise it.

[originally from svn r7979]

16 years agoImprovements to filled-grid generation. Introduced a cunning idea
Simon Tatham [Tue, 8 Apr 2008 09:36:33 +0000 (09:36 +0000)]
Improvements to filled-grid generation. Introduced a cunning idea
suggested by IWJ last night: grid generation can immediately choose
an entire grid row randomly, since all that's doing is nailing down
the names of the numbers, and that gets the whole thing started more
efficiently. But the main difference is that now grid generation is
given only area^2 steps to come up with a filled grid, and then cut
off unceremoniously, causing grid generation to fail and be retried
from scratch. This seems to prevent hangups on jigsaw layouts that
admit few useful solutions, by changing layout constantly. 9j
puzzles now generate at a sensible rate, and as an added bonus so do
5x5 normal puzzles, which they never used to.

[originally from svn r7978]

16 years agoAhem. Apparently forgot to compile-test after that one last tiny
Simon Tatham [Mon, 7 Apr 2008 17:16:08 +0000 (17:16 +0000)]
Ahem. Apparently forgot to compile-test after that one last tiny
change.

[originally from svn r7977]

16 years agoRevise the printing colour framework so that we can explicitly
Simon Tatham [Mon, 7 Apr 2008 17:13:29 +0000 (17:13 +0000)]
Revise the printing colour framework so that we can explicitly
request either of hatching or halftoning, and also choose which to
supply as a fallback when printing in colour.

[originally from svn r7976]

16 years agoPedantic tweaks to allow successful compilation on Windows. (gcc
Simon Tatham [Mon, 7 Apr 2008 17:12:21 +0000 (17:12 +0000)]
Pedantic tweaks to allow successful compilation on Windows. (gcc
failed to point out a declaration after a statement, and gcc's
linker was clever enough to optimise the call to divvy_rectangle()
out of solosolver so that I didn't have to include divvy.c in that.)

[originally from svn r7975]

16 years agoSubstantial reworking of Solo so that it implements both Sudoku-X
Simon Tatham [Mon, 7 Apr 2008 15:56:42 +0000 (15:56 +0000)]
Substantial reworking of Solo so that it implements both Sudoku-X
(require both main diagonals to have one of every digit in addition
to all the usual constraints) and Jigsaw Sudoku (replace the array
of rectangular sub-blocks with the sub-blocks being random
polyominoes). To implement the latter, I've moved my `divvy.c'
library routine out of the `unfinished' subdirectory.

Jigsaw mode is currently an undocumented feature: you enable it by
setting the rows parameter to 1 (and the columns parameter to your
desired grid size, which unlike normal Sudoku can be anything you
like including a prime number). The reason it's undocumented is
because generation times are not yet reliably short: sometimes
generating a jigsaw-type puzzle can hang for hours and still get
nowhere. (The algorithm should terminate in principle, but not in
any time you're prepared to wait.) I _think_ I know how to solve
this, but have yet to try it. Until then, jigsaw mode will remain a
hidden feature.

Printing of X-type puzzles is also substandard at present, because
the current print-colour API replaces the desired light shading of
the X-cells with heavy diagonal hatching. I plan to adjust the API
imminently to address this.

[originally from svn r7974]

16 years agoHmm. Telling xvfb to default to a TrueColor visual did help, in that
Simon Tatham [Thu, 20 Mar 2008 09:18:26 +0000 (09:18 +0000)]
Hmm. Telling xvfb to default to a TrueColor visual did help, in that
it got rid of the bogus backgrounds on all the text; but on the
other hand it mysteriously caused all the images to become black and
white! Serves me right for testing with Bridges which was B&W to
start with. Instead, we'll just tell xvfb to use a 24-bit display
and let it sort out the visuals for itself; that seems to work better.

[originally from svn r7932]

16 years agoI _think_, after some fairly random experimentation, that this ought
Simon Tatham [Thu, 20 Mar 2008 00:19:27 +0000 (00:19 +0000)]
I _think_, after some fairly random experimentation, that this ought
to fix the weird blacked-out text in the xvfb-built screenshots.

[originally from svn r7931]

16 years agoUpdate the OS X Puzzles makefile so that it builds on Leopard and
Simon Tatham [Tue, 11 Mar 2008 17:59:38 +0000 (17:59 +0000)]
Update the OS X Puzzles makefile so that it builds on Leopard and
generates PPC/Intel dual-architecture binaries.

This turns out not to be too painful: you compile and link your
programs using `gcc -arch ppc' or `gcc -arch i386', then you use a
command of the form `lipo -create ppc-binary i386-binary -output
binary' to construct a universal binary. It works equally well on
command-line standalone executable files and the executables within
application directories. Also added the -mmacosx-version-min option,
since otherwise the OS X build tools appear to default to building
binaries which will crash (without anything resembling a
comprehensible error message) on any earlier release.

The handling of version.o in this checkin is somewhat grotty. I'd
prefer a method more cleverly intertwingled with mkfiles.pl so I
didn't have to maintain the OS X architecture list in both
mkfiles.pl and Recipe. (Not that I anticipate Apple switching
architectures again in the immediate future, but it's the principle
of the thing.)

[originally from svn r7916]

16 years agoUI change to Filling: allow multiple squares to be set at once.
Jacob Nevins [Sun, 10 Feb 2008 18:43:29 +0000 (18:43 +0000)]
UI change to Filling: allow multiple squares to be set at once.

(This change adds a new possibility to the save format, such that new save
files won't necessarily be loadable by old binaries. I think that's acceptable
-- it's certainly happened before -- but I couldn't find anything in the
developer docs explicitly blessing it.)

[originally from svn r7849]

16 years agoIt's another new year.
Jacob Nevins [Sun, 10 Feb 2008 17:12:49 +0000 (17:12 +0000)]
It's another new year.

[originally from svn r7848]

16 years agoSyntactic fixes by Zach Wily to make the OS X build work under 10.5.
Simon Tatham [Tue, 15 Jan 2008 18:04:13 +0000 (18:04 +0000)]
Syntactic fixes by Zach Wily to make the OS X build work under 10.5.

[originally from svn r7836]

16 years ago64-bit cleanliness: we were already carefully using a uint32 type in
Simon Tatham [Sat, 15 Dec 2007 21:09:45 +0000 (21:09 +0000)]
64-bit cleanliness: we were already carefully using a uint32 type in
the SHA code, but it wasn't correctly defined!

[originally from svn r7817]

16 years agoA rigorous proof. Totally unimportant to the code, but I didn't want
Simon Tatham [Sat, 25 Aug 2007 17:46:13 +0000 (17:46 +0000)]
A rigorous proof. Totally unimportant to the code, but I didn't want
to lose it :-)

[originally from svn r7703]

16 years agoFix an inaccurate comment.
Simon Tatham [Sat, 25 Aug 2007 15:50:33 +0000 (15:50 +0000)]
Fix an inaccurate comment.

[originally from svn r7702]

16 years agoI've just realised that my deliberate avoidance of non-simply
Simon Tatham [Sat, 25 Aug 2007 15:32:41 +0000 (15:32 +0000)]
I've just realised that my deliberate avoidance of non-simply
connected polyominoes actually causes a loss of generality for
sufficiently large k. I hadn't previously noticed, because you need
k to be (I think) at least 23 and none of my potential applications
require anything nearly that large. Add some discussion of this.

[originally from svn r7701]

16 years agoCommit my work so far on a generator for Nikoli's `Block Puzzle'. It
Simon Tatham [Sat, 25 Aug 2007 14:10:49 +0000 (14:10 +0000)]
Commit my work so far on a generator for Nikoli's `Block Puzzle'. It
works, but it's slow, and the puzzles are currently at a relatively
low level of difficulty. Also this is a generator only: no UI yet
(because I'm waiting to see if I can make the generator practical
before bothering to write the rest).

[originally from svn r7700]

16 years agoAhem. Finishing writing the comment _before_ checkin is generally sensible.
Simon Tatham [Sat, 18 Aug 2007 13:32:56 +0000 (13:32 +0000)]
Ahem. Finishing writing the comment _before_ checkin is generally sensible.

[originally from svn r7694]

16 years agoAllow a 1-omino to be completely destroyed and recreated in an
Simon Tatham [Sat, 18 Aug 2007 13:30:13 +0000 (13:30 +0000)]
Allow a 1-omino to be completely destroyed and recreated in an
arbitrary unclaimed square. This cures the most common cause of
generation failures (covering a large area in dominoes was the most
difficult case, and would fail even if the large area was 1xn!); the
failure rate is now sufficiently low under all circumstances I've
found that I'm willing to just loop until I get a success.

[originally from svn r7693]

16 years agoBetter test-mode diagnostics.
Simon Tatham [Sat, 18 Aug 2007 11:19:29 +0000 (11:19 +0000)]
Better test-mode diagnostics.

[originally from svn r7691]

16 years agoA piece of library code which constructs a random division of a
Simon Tatham [Sat, 18 Aug 2007 10:07:29 +0000 (10:07 +0000)]
A piece of library code which constructs a random division of a
rectangle into equally sized ominoes. I have a couple of potential
applications for this, but none I've actually implemented yet, so
for the moment it's living in `unfinished'.

[originally from svn r7690]

16 years agoExperimental UI tweak enabled by a hacky environment variable:
Simon Tatham [Tue, 31 Jul 2007 17:04:20 +0000 (17:04 +0000)]
Experimental UI tweak enabled by a hacky environment variable:
suppress the display of `this square can't be a light' blobs in a
lit square, on the grounds that we already know _lit_ squares can't
be lights. This makes the solved game look cleaner (I've always
thought the detritus of blobs on some but not all non-light squares
looked messy), but on the other hand it's slightly jarring during
play. So I'm checking it in, but as a configurable option which is
off by default.

[originally from svn r7656]

16 years agoA user points out that Loopy's solver relies on elements of the
Simon Tatham [Thu, 5 Jul 2007 18:57:50 +0000 (18:57 +0000)]
A user points out that Loopy's solver relies on elements of the
`clues' array being able to be -1, so we must explicitly declare it
as `signed char' or it will break on platforms whose default char is
unsigned.

[originally from svn r7636]

16 years agoOptimiser placation.
Simon Tatham [Fri, 29 Jun 2007 09:45:44 +0000 (09:45 +0000)]
Optimiser placation.

[originally from svn r7625]

17 years agoUpdates and improvements from Jonas Koelker.
Simon Tatham [Sun, 20 May 2007 14:28:48 +0000 (14:28 +0000)]
Updates and improvements from Jonas Koelker.

[originally from svn r7601]

17 years agoLeft puzzles.rc2 out of the tarball. Oops.
Simon Tatham [Sat, 19 May 2007 06:16:24 +0000 (06:16 +0000)]
Left puzzles.rc2 out of the tarball. Oops.

[originally from svn r7600]

17 years agoMore forgiving selection of dragging targets.
Simon Tatham [Sat, 12 May 2007 13:13:39 +0000 (13:13 +0000)]
More forgiving selection of dragging targets.

[originally from svn r7574]