chiark / gitweb /
sgt-puzzles.git
18 years agoD'oh, there's always one. Remove first-click stuff from the todo list.
Simon Tatham [Mon, 30 May 2005 13:11:05 +0000 (13:11 +0000)]
D'oh, there's always one. Remove first-click stuff from the todo list.

[originally from svn r5863]

18 years agoMines now follows the conventional approach of offering a completely
Simon Tatham [Mon, 30 May 2005 13:10:37 +0000 (13:10 +0000)]
Mines now follows the conventional approach of offering a completely
blank grid until you make the first click; to ensure solubility, it
does not generate the mine layout until that click, and then ensures
it is solvable starting from that position.

This has involved three infrastructure changes:

 - random.c now offers functions to encode and decode an entire
   random_state as a string
 - each puzzle's new_game() is now passed a pointer to the midend
   itself, which most of them ignore
 - there's a function in the midend which a game can call back to
   _rewrite_ its current game description.

So Mines now has two entirely separate forms of game ID. One
contains the generation-time parameters (n and unique) plus an
encoding of a random_state; the other actually encodes the grid once
it's been generated, and also contains the initial click position.
When called with the latter, new_game() does plausibly normal stuff.
When called with the former, it notes down all the details and waits
until the first square is opened, and _then_ does the grid
generation and updates the game description in the midend. So if,
_after_ your first click, you decide you want to share this
particular puzzle with someone else, you can do that fine.

Also in this checkin, the mine layout is no longer _copied_ between
all the game_states on the undo chain. Instead, it's in a separate
structure and all game_states share a pointer to it - and the
structure is reference-counted to ensure deallocation.

[originally from svn r5862]

18 years agoNeat idea from Gareth: if you put a % on the end of the mine count
Simon Tatham [Mon, 30 May 2005 12:24:31 +0000 (12:24 +0000)]
Neat idea from Gareth: if you put a % on the end of the mine count
in the Custom dialog box, it'll treat it as a mine density.

[originally from svn r5861]

18 years agoNit: "warning: extra tokens at end of #endif directive"
Jacob Nevins [Mon, 30 May 2005 11:49:08 +0000 (11:49 +0000)]
Nit: "warning: extra tokens at end of #endif directive"

[originally from svn r5860]

18 years agoInitial checkin of my Minesweeper clone, which uses a solver during
Simon Tatham [Mon, 30 May 2005 10:08:27 +0000 (10:08 +0000)]
Initial checkin of my Minesweeper clone, which uses a solver during
grid generation to arrange a mine layout that never requires guessing.

[originally from svn r5859]

18 years agoInfrastructure change: game_anim_length and game_flash_length now
Simon Tatham [Mon, 30 May 2005 07:55:27 +0000 (07:55 +0000)]
Infrastructure change: game_anim_length and game_flash_length now
both get passed a pointer to the game_ui. This means that if they
need to note down information for the redraw function about what
_type_ of flash or animation is required, they now have somewhere to
do so.

[originally from svn r5858]

18 years agoAdd a limited-shuffle mode like that added to Sixteen and Twiddle in r5769,
Jacob Nevins [Sat, 28 May 2005 23:45:43 +0000 (23:45 +0000)]
Add a limited-shuffle mode like that added to Sixteen and Twiddle in r5769,
for completeness.

[originally from svn r5854]
[r5769 == 8f1c713735316422cfe041400ccc49999d563d8b]

18 years agoAdd the ability to use the Rectangles solver for actually solving
Simon Tatham [Sat, 28 May 2005 08:04:29 +0000 (08:04 +0000)]
Add the ability to use the Rectangles solver for actually solving
puzzles, rather than just doing its nondeterministic number
placement thing. This enables the use of the `Solve' menu option on
externally entered game IDs, provided of course that they aren't
_too_ difficult.

[originally from svn r5852]

18 years agoAdd support for Jacob's new cursor-key modifier flags in the OS X frontend.
Simon Tatham [Thu, 26 May 2005 17:12:04 +0000 (17:12 +0000)]
Add support for Jacob's new cursor-key modifier flags in the OS X frontend.

[originally from svn r5847]

18 years agoSorted out the barriers/corners trouble left by r5844. No wonder
Simon Tatham [Thu, 26 May 2005 17:03:51 +0000 (17:03 +0000)]
Sorted out the barriers/corners trouble left by r5844. No wonder
Jacob wasn't able to find a satisfactory fix: the whole area was a
horrid mess. Fortunately, the reason it was a horrid mess was
because the Net drawing routines predated the introduction of clip()
in the frontend interface, and it turns out that clip() makes it
possible to do all this more easily and better. So, a complete
rearchitecting of barrier corners: the corner flags in the
`barriers' array are now gone (and good riddance), and corner
information is computed on the fly so as to take into account the
moving grid edges. In the process I've also updated the corner
mechanism so that a barrier `corner' (really endpoint) is drawn at
the end of _every_ barrier, not just where two meet. This has
changed the appearance of a single isolated barrier, to what I would
have wanted it to look like in the first place but achieving it
without clip() was just too fiddly.

[originally from svn r5846]
[r5844 == 865e8ad6ca3d83ad2a585ceeb1809e9f68c18a20]

18 years agoSince the split into random and descriptive IDs, the section on game seeds has
Jacob Nevins [Thu, 26 May 2005 16:57:19 +0000 (16:57 +0000)]
Since the split into random and descriptive IDs, the section on game seeds has
been mostly covered by the main documentation or otherwise moot.

[originally from svn r5845]

18 years agoAdd origin-shifting (Shift+cursors) and source-shifting (Ctrl+cursors) to Net.
Jacob Nevins [Thu, 26 May 2005 13:40:38 +0000 (13:40 +0000)]
Add origin-shifting (Shift+cursors) and source-shifting (Ctrl+cursors) to Net.
(Adding modifier+cursors handling has had minor knock-on effects on the other
puzzles, so that they can continue to ignore modifiers.)

(An unfortunate side effect of this is some artifacts in exterior barrier
drawing; notably, a disconnected corner can now appear at the corner of the
grid under some circumstances. I haven't found a satisfactory way round
this yet.)

[originally from svn r5844]

19 years agoSupport for `pencil marks' in Solo, by right-clicking and typing a
Simon Tatham [Wed, 25 May 2005 11:09:43 +0000 (11:09 +0000)]
Support for `pencil marks' in Solo, by right-clicking and typing a
number. Many thanks to Chris Thomas, for helping with the detailed
UI design by means of testing an endless series of prototypes.

[originally from svn r5842]

19 years agoMention NetWalk and update comment
Jacob Nevins [Tue, 24 May 2005 20:28:38 +0000 (20:28 +0000)]
Mention NetWalk and update comment

[originally from svn r5841]

19 years agoAvoid leading zeroes on internally generated random seeds, _just_ in
Simon Tatham [Mon, 23 May 2005 12:02:37 +0000 (12:02 +0000)]
Avoid leading zeroes on internally generated random seeds, _just_ in
case they confuse anyone who expects the same seed without the
leading zero to be equivalent.

[originally from svn r5838]

19 years agoI'm sick of `--generate' giving the same answers if you run it twice
Simon Tatham [Mon, 23 May 2005 11:15:39 +0000 (11:15 +0000)]
I'm sick of `--generate' giving the same answers if you run it twice
in one second. Switched the Unix get_random_seed() to using
gettimeofday() rather than time().

[originally from svn r5837]

19 years agoThe GTK `--generate' option didn't validate the game parameters. It
Simon Tatham [Mon, 23 May 2005 11:13:39 +0000 (11:13 +0000)]
The GTK `--generate' option didn't validate the game parameters. It
does now.

[originally from svn r5836]

19 years agoNet hangs if you ask it for a 2xn or nx2 wrapping puzzle with a
Simon Tatham [Mon, 23 May 2005 11:03:52 +0000 (11:03 +0000)]
Net hangs if you ask it for a 2xn or nx2 wrapping puzzle with a
unique solution. This, it turns out, is because there is literally
no such thing. Protective constraint added to validate_params(),
with a proof in a comment alongside.

If you really want a 2xn or nx2 wrapping puzzle, you can still have
one if you turn uniqueness off.

[originally from svn r5835]

19 years agoAesthetic improvements, mostly suggested by Gareth.
Simon Tatham [Sun, 22 May 2005 13:35:13 +0000 (13:35 +0000)]
Aesthetic improvements, mostly suggested by Gareth.

[originally from svn r5833]

19 years agoForgot to add print.py to the source archive. Ahem.
Simon Tatham [Sun, 22 May 2005 12:45:39 +0000 (12:45 +0000)]
Forgot to add print.py to the source archive. Ahem.

[originally from svn r5830]

19 years agoAnother tweak to the solver to make it handle blank tiles correctly.
Simon Tatham [Sun, 22 May 2005 12:14:39 +0000 (12:14 +0000)]
Another tweak to the solver to make it handle blank tiles correctly.
The previous checkin stopped it choking on them, but it didn't
actually manage to _deduce_ that all the edges bordering them had to
be closed. Now it does better.

[originally from svn r5829]

19 years agoSpecial case in dead-end checking which prevents the solver falling
Simon Tatham [Sun, 22 May 2005 11:45:20 +0000 (11:45 +0000)]
Special case in dead-end checking which prevents the solver falling
over on a grid containing a 0 (completely blank) tile. This can't
happen in self-generated grids, but can happen if you type in a grid
from another Net implementation. Previously, the solver would notice
(technically correctly!) that a completely blank tile connects to no
other tiles and thus forms an isolated subgraph, and would therefore
complain that no orientation of that tile could possibly yield a
valid solution...

[originally from svn r5828]

19 years agoThe Net solver now makes use of barrier information when applied to
Simon Tatham [Sun, 22 May 2005 11:15:03 +0000 (11:15 +0000)]
The Net solver now makes use of barrier information when applied to
a typed-in grid.

[originally from svn r5827]

19 years agoIt occurred to me yesterday that Net could perfectly well be played
Simon Tatham [Sun, 22 May 2005 11:06:25 +0000 (11:06 +0000)]
It occurred to me yesterday that Net could perfectly well be played
on paper, so here's a routine to print it out.

[originally from svn r5826]

19 years agoEnsure that an old random seed isn't left around for the user to see when a
Jacob Nevins [Sat, 21 May 2005 22:07:48 +0000 (22:07 +0000)]
Ensure that an old random seed isn't left around for the user to see when a
descriptive ID has been specified.
Fix tiny memory leak.

[originally from svn r5825]

19 years agoCleanups:
Simon Tatham [Sat, 21 May 2005 13:39:23 +0000 (13:39 +0000)]
Cleanups:
 - fix documentation of Net's unique solution option (should have
   tested before last checkin)
 - make unique solutions optional in Rectangles too (same reasons)
 - tidy up various issues in parameter encoding in both games.

[originally from svn r5818]

19 years agoSolution uniqueness for Net. Can be disabled on request (but is
Simon Tatham [Sat, 21 May 2005 13:23:26 +0000 (13:23 +0000)]
Solution uniqueness for Net. Can be disabled on request (but is
enabled by default), since ambiguous sections in grids can present
additional interesting challenges. I think uniqueness is a better
default, though.

[originally from svn r5816]

19 years agoMove IDM_ABOUT so that it doesn't overlap the presets space! Ahem.
Simon Tatham [Fri, 20 May 2005 12:30:37 +0000 (12:30 +0000)]
Move IDM_ABOUT so that it doesn't overlap the presets space! Ahem.

[originally from svn r5814]

19 years agoCunning way to ensure unique solutions in generated Rectangles
Simon Tatham [Thu, 19 May 2005 16:17:03 +0000 (16:17 +0000)]
Cunning way to ensure unique solutions in generated Rectangles
puzzles. I generate the grid of rectangles as normal, but before I
place the numbers I run it through a non-deterministic solver
algorithm which tries to do as much as it can with as little
information about where the numbers are going to be. The solver
itself narrows down the number placement when it runs out of steam,
but does so as little as possible. Once it reaches a state where it
has ensured solubility, and then the generation algorithm chooses
random number placement from whatever's left.

Occasionally it paints itself into a corner and can't ensure a
unique solution no matter what happens; in that situation we just
have to give up, generate a fresh grid, and try again.

[originally from svn r5809]

19 years agoIf you paste in a random seed with different ephemeral parameters
Simon Tatham [Wed, 18 May 2005 17:41:53 +0000 (17:41 +0000)]
If you paste in a random seed with different ephemeral parameters
from the currently configured ones, and then bring the Random Seed
box back up, the wrong parameters get shown, and the resulting
random seed is incorrect for the current game.

At least, it was, until this checkin.

[originally from svn r5808]

19 years agoRectangles random seed IDs shouldn't bother stating the expansion
Simon Tatham [Wed, 18 May 2005 17:28:48 +0000 (17:28 +0000)]
Rectangles random seed IDs shouldn't bother stating the expansion
factor if it's zero.

[originally from svn r5807]

19 years ago--version.
Simon Tatham [Wed, 18 May 2005 17:27:16 +0000 (17:27 +0000)]
--version.

[originally from svn r5806]

19 years agoIt's a good idea to validate presets received from the environment.
Simon Tatham [Wed, 18 May 2005 17:25:30 +0000 (17:25 +0000)]
It's a good idea to validate presets received from the environment.

[originally from svn r5805]

19 years agoMove the colour configuration into midend.c so that it becomes
Simon Tatham [Wed, 18 May 2005 09:04:47 +0000 (09:04 +0000)]
Move the colour configuration into midend.c so that it becomes
cross-platform, and rename the environment variables so that they
follow the puzzle name. Should allow a static environment
configuration for each puzzle. Also introduced a <game>_PRESETS
variable for people whose favourite configuration isn't on the Type
menu by default.

[originally from svn r5801]

19 years agoStandalone solvers were broken by my recent API changes.
Simon Tatham [Wed, 18 May 2005 08:34:01 +0000 (08:34 +0000)]
Standalone solvers were broken by my recent API changes.

[originally from svn r5800]

19 years agoUpdate doc for recent changes in Restart behaviour.
Jacob Nevins [Tue, 17 May 2005 18:04:12 +0000 (18:04 +0000)]
Update doc for recent changes in Restart behaviour.

[originally from svn r5799]

19 years agoAfter much thought, I've decided that `Restart' on r is not a
Simon Tatham [Tue, 17 May 2005 17:20:08 +0000 (17:20 +0000)]
After much thought, I've decided that `Restart' on r is not a
particularly useful keypress, particularly given how easy it is to
confuse it with `Redo'. So both r and ^R are now Redo, and Restart
is relegated to being a menu-only option.

[originally from svn r5796]

19 years agoNull-terminate generated Net/Netslide descriptive game IDs.
Jacob Nevins [Tue, 17 May 2005 12:02:30 +0000 (12:02 +0000)]
Null-terminate generated Net/Netslide descriptive game IDs.

[originally from svn r5794]

19 years agoKeyboard shortcuts for Twiddle: abcdABCD in line with the notation
Simon Tatham [Tue, 17 May 2005 11:53:42 +0000 (11:53 +0000)]
Keyboard shortcuts for Twiddle: abcdABCD in line with the notation
Gareth and I have been using to analyse the game, and also the
number pad. They don't work sensibly for all sizes, but they'll be
handy for the most common ones.

[originally from svn r5793]

19 years agoJust for Gareth: a means of overriding individual game colour
Simon Tatham [Tue, 17 May 2005 11:47:33 +0000 (11:47 +0000)]
Just for Gareth: a means of overriding individual game colour
settings using environment variables. GTK frontend only, because
this is an unsupported (and unprincipled) hack.

[originally from svn r5792]

19 years ago`Restart' is now an undo-able action: it appends a move to the end
Simon Tatham [Tue, 17 May 2005 11:46:55 +0000 (11:46 +0000)]
`Restart' is now an undo-able action: it appends a move to the end
of the undo list rather than destroying it. Partly this is because
accidental restarts are a real pain, and partly because it allows
you to compare the initial to the current state by restart-then-undo
which is handy in some puzzles.

In order to do this, I've introduced an additional per-entry field
in the undo list in the midend, which tracks which states were
created by `unusual' operations (Solve and Restart). The midend
takes care of suppressing animation and completion flashes during
transitions between a `special' state and its predecessor, relieving
the game backends of having to do it individually.

(This probably means I could remove some complexity in the
flash_time() functions in most backends, but I haven't done that in
this checkin.)

[originally from svn r5791]

19 years agocomment c'n'p error
Jacob Nevins [Tue, 17 May 2005 11:00:15 +0000 (11:00 +0000)]
comment c'n'p error

[originally from svn r5790]

19 years agoCube's dup_game() function was missing a field. Oops.
Simon Tatham [Tue, 17 May 2005 10:59:47 +0000 (10:59 +0000)]
Cube's dup_game() function was missing a field. Oops.

[originally from svn r5789]

19 years agoThe game IDs for Net (and Netslide) have always been random seeds
Simon Tatham [Mon, 16 May 2005 18:57:09 +0000 (18:57 +0000)]
The game IDs for Net (and Netslide) have always been random seeds
rather than literal grid descriptions, which has always faintly
annoyed me because it makes it impossible to type in a grid from
another source. However, Gareth pointed out that short random-seed
game descriptions are useful, because you can read one out to
someone else without having to master the technology of cross-
machine cut and paste, or you can have two people enter the same
random seed simultaneously in order to race against each other to
complete the same puzzle. So both types of game ID seem to have
their uses.

Therefore, here's a reorganisation of the whole game ID concept.
There are now two types of game ID: one has a parameter string then
a hash then a piece of arbitrary random seed text, and the other has
a parameter string then a colon then a literal game description. For
most games, the latter is identical to the game IDs that were
previously valid; for Net and Netslide, old game IDs must be
translated into new ones by turning the colon into a hash, and
there's a new descriptive game ID format.

Random seed IDs are not guaranteed to be portable between software
versions (this is a major reason why I added version reporting
yesterday). Descriptive game IDs have a longer lifespan.

As an added bonus, I've removed the sections of documentation
dealing with game parameter encodings not shown in the game ID
(Rectangles expansion factor, Solo symmetry and difficulty settings
etc), because _all_ parameters must be specified in a random seed ID
and therefore users can easily find out the appropriate parameter
string for any settings they have configured.

[originally from svn r5788]

19 years agoMissed a vital semicolon off the Cygwin version.c makefile fragment.
Simon Tatham [Sun, 15 May 2005 11:12:05 +0000 (11:12 +0000)]
Missed a vital semicolon off the Cygwin version.c makefile fragment.

[originally from svn r5784]

19 years agoEr, except that I already had a makedist.sh. Transferred new version
Simon Tatham [Sun, 15 May 2005 10:58:09 +0000 (10:58 +0000)]
Er, except that I already had a makedist.sh. Transferred new version
code into that!

[originally from svn r5783]

19 years agoAdded automatic determination of the SVN revision number if a build
Simon Tatham [Sun, 15 May 2005 10:52:23 +0000 (10:52 +0000)]
Added automatic determination of the SVN revision number if a build
takes place in a checked-out copy. Also added mkunxarc.sh which does
the same version determination before building its archive.

[originally from svn r5782]

19 years agoIntroduce a versioning mechanism, and an `About' box in all front
Simon Tatham [Sun, 15 May 2005 10:31:11 +0000 (10:31 +0000)]
Introduce a versioning mechanism, and an `About' box in all front
ends. Versioning will be done solely by Subversion revision number,
since development on these puzzles is very incremental and gradual
and there don't tend to be obvious points to place numbered
releases.

[originally from svn r5781]

19 years agoI've had two complaints that Solo ought to recognise the numeric
Simon Tatham [Thu, 12 May 2005 18:25:57 +0000 (18:25 +0000)]
I've had two complaints that Solo ought to recognise the numeric
keypad. The reason it doesn't is because front ends were carefully
translating the numeric keypad into 8-way directional keys for the
benefit of Cube. Therefore, a policy change:
 - front ends process the numeric keypad by sending MOD_NUM_KEYPAD |
   '3' and similar
 - front ends running on a platform with Num Lock SHOULD do this
   _irrespective_ of the state of Num Lock
 - back ends do whatever they see fit with numeric keypad keys.
Result: the numeric keypad now works in Solo, and also works in OS X
Cube (which it previously didn't because I forgot to implement that
bit of the front end!).

[originally from svn r5774]

19 years agoThe two Rubik-like puzzles, Sixteen and Twiddle, now support an
Simon Tatham [Wed, 11 May 2005 20:38:10 +0000 (20:38 +0000)]
The two Rubik-like puzzles, Sixteen and Twiddle, now support an
additional configuration parameter, which is the number of shuffle
moves. By default the grid will be fully shuffled so that you need a
general solution algorithm to untangle it, but if you prefer you can
request a grid which has had (say) precisely four moves made on it,
and then attempt to exactly reverse those four moves.

Currently this feature is only available from the Custom box, and
not in any presets.

[originally from svn r5769]

19 years agoBah. Try the r5766 fix again, this time without the typo.
Simon Tatham [Wed, 11 May 2005 18:49:42 +0000 (18:49 +0000)]
Bah. Try the r5766 fix again, this time without the typo.

[originally from svn r5767]
[r5766 == 701cd045b36f9be1b0b8cfb74d9c191cb5813e98]

19 years agoCompletion flashes were occasionally failing to be cleaned up if a
Simon Tatham [Wed, 11 May 2005 13:03:17 +0000 (13:03 +0000)]
Completion flashes were occasionally failing to be cleaned up if a
subsequent move animation began during them. Tracked this to
overenthusiastic use of clip() and fixed it.

[originally from svn r5766]

19 years agosolve_game() is passed the _initial_ game state, not the most recent
Simon Tatham [Sat, 7 May 2005 16:07:26 +0000 (16:07 +0000)]
solve_game() is passed the _initial_ game state, not the most recent
one; so we can't just set `ret->completed = ret->movecount' and hope
it's been set to something other than zero. Instead, we set both
move counts to 1, which is entirely arbitrary but works.

This fixes a subtle bug with the Solve feature: if you pressed
Solve, then disturbed the grid, then brought it back to the solved
state by making more forward moves (rather than using Undo), then
the first time you did this the `Moves since auto-solve' status line
would reset to zero.

[originally from svn r5759]

19 years agoAdd 3x3 Trivial to the presets list, and make it the default.
Simon Tatham [Sat, 7 May 2005 12:35:04 +0000 (12:35 +0000)]
Add 3x3 Trivial to the presets list, and make it the default.

[originally from svn r5757]

19 years agoAha, here's a nice easy way to generate really hard puzzles. Added
Simon Tatham [Sat, 7 May 2005 12:30:29 +0000 (12:30 +0000)]
Aha, here's a nice easy way to generate really hard puzzles. Added
the missing fifth difficulty level to Solo: `Unreasonable', in which
even set-based reasoning is insufficient and there's no alternative
but to guess a number and backtrack if it didn't work. (Solutions
are still guaranteed unique, however.) In fact it now seems to take
less time to generate a puzzle of this grade than `Advanced'!

[originally from svn r5756]

19 years agoFix outdated comment
Jacob Nevins [Wed, 4 May 2005 13:17:45 +0000 (13:17 +0000)]
Fix outdated comment

[originally from svn r5744]

19 years agoAllow for trailing '\0' in game_text_format() in various games.
Jacob Nevins [Wed, 4 May 2005 12:56:04 +0000 (12:56 +0000)]
Allow for trailing '\0' in game_text_format() in various games.

[originally from svn r5743]

19 years agoThe Twiddle shuffling algorithm was theoretically parity-unbalanced:
Simon Tatham [Wed, 4 May 2005 12:52:51 +0000 (12:52 +0000)]
The Twiddle shuffling algorithm was theoretically parity-unbalanced:
it performed a fixed number of shuffling moves, and on each one it
had a 2/3 chance of flipping the permutation parity and a 1/3 chance
of keeping it the same. Markov analysis shows that over a run of
1500-odd shuffle moves this will end up being an undetectably small
actual bias in the parity of the generated grid, but it offends my
sense of pedantry nonetheless so here's a small change to make the
number of shuffling moves itself have randomly chosen parity. The
parity of generated grids should now be _exactly_ 50:50.

[originally from svn r5742]

19 years agomarkup typo
Jacob Nevins [Wed, 4 May 2005 12:24:16 +0000 (12:24 +0000)]
markup typo

[originally from svn r5739]

19 years agoI've changed my mind. For the benefit of users with slower
Simon Tatham [Mon, 2 May 2005 16:59:50 +0000 (16:59 +0000)]
I've changed my mind. For the benefit of users with slower
computers, let's save the Solo and Pattern grids at generation time
and regurgitate them when asked to solve, rather than doing all the
work over again.

[originally from svn r5737]

19 years agoFix line endings when pasting on Windows.
Simon Tatham [Mon, 2 May 2005 16:37:20 +0000 (16:37 +0000)]
Fix line endings when pasting on Windows.

[originally from svn r5736]

19 years agoForgot to mention that you can undo a Solve operation.
Simon Tatham [Mon, 2 May 2005 13:27:59 +0000 (13:27 +0000)]
Forgot to mention that you can undo a Solve operation.

[originally from svn r5734]

19 years agoAhem. The `Solve' option in orientable Twiddle needs to correct the
Simon Tatham [Mon, 2 May 2005 13:22:25 +0000 (13:22 +0000)]
Ahem. The `Solve' option in orientable Twiddle needs to correct the
orientations as well as the order!

[originally from svn r5733]

19 years agoSilly (but harmless) typo.
Simon Tatham [Mon, 2 May 2005 13:18:24 +0000 (13:18 +0000)]
Silly (but harmless) typo.

[originally from svn r5732]

19 years agoAdded an automatic `Solve' feature to most games. This is useful for
Simon Tatham [Mon, 2 May 2005 13:17:10 +0000 (13:17 +0000)]
Added an automatic `Solve' feature to most games. This is useful for
various things:
 - if you haven't fully understood what a game is about, it gives
   you an immediate example of a puzzle plus its solution so you can
   understand it
 - in some games it's useful to compare your solution with the real
   one and see where you made a mistake
 - in the rearrangement games (Fifteen, Sixteen, Twiddle) it's handy
   to be able to get your hands on a pristine grid quickly so you
   can practise or experiment with manoeuvres on it
 - it provides a good way of debugging the games if you think you've
   encountered an unsolvable grid!

[originally from svn r5731]

19 years agoOops; forgot to check in the copy-to-clipboard option for Windows.
Simon Tatham [Mon, 2 May 2005 10:55:32 +0000 (10:55 +0000)]
Oops; forgot to check in the copy-to-clipboard option for Windows.

[originally from svn r5730]

19 years agoIntroduce the concept of a `game_aux_info' structure. This is
Simon Tatham [Mon, 2 May 2005 10:12:26 +0000 (10:12 +0000)]
Introduce the concept of a `game_aux_info' structure. This is
constructed at the same time as an internally generated game seed,
so that it can preserve any interesting information known by the
program at generation time but not physically contained within the
text of the game seed itself. (Such as, for example, the solution.)
Currently not used for anything yet, but it will be.

[originally from svn r5729]

19 years agoIt's actually vitally important, it turns out, to have all of the
Simon Tatham [Mon, 2 May 2005 09:42:09 +0000 (09:42 +0000)]
It's actually vitally important, it turns out, to have all of the
`Cut', `Copy' and `Paste' items in the Edit menu of an OS X
application - because there's nothing else that enables the keyboard
cut/copy/paste shortcuts in an edit box! OS X Puzzles can now have
game IDs pasted into it, which it previously couldn't.

[originally from svn r5728]

19 years agoThe addition of a `Copy' menu item on OS X was really beginning to
Simon Tatham [Sun, 1 May 2005 14:05:03 +0000 (14:05 +0000)]
The addition of a `Copy' menu item on OS X was really beginning to
strain my unconventional menu organisation, so I've reverted to
having `File' and `Edit' menus like everyone else.

[originally from svn r5727]

19 years agoCopy-to-clipboard for Rectangles.
Simon Tatham [Sun, 1 May 2005 13:51:46 +0000 (13:51 +0000)]
Copy-to-clipboard for Rectangles.

[originally from svn r5726]

19 years agoCopy-to-clipboard facility for Fifteen, Sixteen and Twiddle.
Simon Tatham [Sun, 1 May 2005 13:22:44 +0000 (13:22 +0000)]
Copy-to-clipboard facility for Fifteen, Sixteen and Twiddle.

[originally from svn r5725]

19 years agoIntroduced a new function in every game which formats a game_state
Simon Tatham [Sun, 1 May 2005 12:53:41 +0000 (12:53 +0000)]
Introduced a new function in every game which formats a game_state
as text. This is used by front ends to implement copy-to-clipboard.
Currently the function does nothing (and is disabled) in every game
except Solo, but it's a start.

[originally from svn r5724]

19 years agoI can never remember what that `TRUE' means in the game structure
Simon Tatham [Sun, 1 May 2005 11:07:13 +0000 (11:07 +0000)]
I can never remember what that `TRUE' means in the game structure
definitions, so let's move it so that it's just next to the
functions it relates to. This also opens the way for me to add more
booleans next to other functions without getting confused as to
which is which.

[originally from svn r5723]

19 years agoRemove outdated comment :-)
Simon Tatham [Sun, 1 May 2005 10:57:47 +0000 (10:57 +0000)]
Remove outdated comment :-)

[originally from svn r5722]

19 years agoI think Windows fonts look better in bold as well.
Simon Tatham [Sun, 1 May 2005 10:57:23 +0000 (10:57 +0000)]
I think Windows fonts look better in bold as well.

[originally from svn r5721]

19 years agoFix game IDs, which I broke in the orientability change. Also
Simon Tatham [Sat, 30 Apr 2005 14:50:33 +0000 (14:50 +0000)]
Fix game IDs, which I broke in the orientability change. Also
introduce a sensible game ID notation for orientable games, and
finally (*blush*) turn the orientability triangles back the right
way up.

[originally from svn r5718]

19 years agoAfter brainstorming with Gareth, we've decided that this is a much
Simon Tatham [Sat, 30 Apr 2005 14:38:20 +0000 (14:38 +0000)]
After brainstorming with Gareth, we've decided that this is a much
simpler and better way to indicate tile orientation than those
colour bars.

[originally from svn r5717]

19 years agoBah, and of course there's a TODO comment I forgot to remove.
Simon Tatham [Sat, 30 Apr 2005 14:14:37 +0000 (14:14 +0000)]
Bah, and of course there's a TODO comment I forgot to remove.

[originally from svn r5714]

19 years agoTwiddle now has an additional mode in which tile orientation
Simon Tatham [Sat, 30 Apr 2005 14:14:14 +0000 (14:14 +0000)]
Twiddle now has an additional mode in which tile orientation
matters. This mode is hard enough to scare the wossnames out of me,
but that's no reason not to put it in anyway!

[originally from svn r5713]

19 years agoPango-derived fonts seem to generally look better if I ask for them
Simon Tatham [Sat, 30 Apr 2005 13:11:22 +0000 (13:11 +0000)]
Pango-derived fonts seem to generally look better if I ask for them
in bold.

[originally from svn r5709]

19 years agoNew puzzle: `twiddle', generalised from a random door-unlocking
Simon Tatham [Sat, 30 Apr 2005 12:54:22 +0000 (12:54 +0000)]
New puzzle: `twiddle', generalised from a random door-unlocking
gadget in Metroid Prime 2.

[originally from svn r5708]

19 years agoAdded a `--generate' command-line option in the GTK port of every
Simon Tatham [Fri, 29 Apr 2005 17:07:19 +0000 (17:07 +0000)]
Added a `--generate' command-line option in the GTK port of every
puzzle, to make it construcct puzzle IDs and output them on stdout.
Also checked in print.py, a script which reads puzzle IDs on stdin
and produces PostScript output. With these, you can generate pages
of Pattern, Rectangles and Solo puzzles to take on trains with you.

[originally from svn r5707]

19 years agoCorrect block coordinate display when standalone solver shows its
Simon Tatham [Fri, 29 Apr 2005 08:11:01 +0000 (08:11 +0000)]
Correct block coordinate display when standalone solver shows its
working.

[originally from svn r5706]

19 years agoAha, even better: a Makefile hack that causes auto-detection of GTK
Simon Tatham [Thu, 28 Apr 2005 12:50:56 +0000 (12:50 +0000)]
Aha, even better: a Makefile hack that causes auto-detection of GTK
2 and fallback to GTK 1.2.

[originally from svn r5705]

19 years agoFix inconsistent vertical centring of text in GTK (and update Solo's
Simon Tatham [Wed, 27 Apr 2005 19:44:34 +0000 (19:44 +0000)]
Fix inconsistent vertical centring of text in GTK (and update Solo's
todo list, which complained about this because Solo was the only
puzzle noticeably affected).

[originally from svn r5694]

19 years agoPatches for GTK 2. Puzzles already _built_ under GTK 2, but now it
Simon Tatham [Wed, 27 Apr 2005 19:39:20 +0000 (19:39 +0000)]
Patches for GTK 2. Puzzles already _built_ under GTK 2, but now it
builds better:
 - the GTK makefile now defines $(GTK_CONFIG) which you can
   override, so you can build for GTK 2 with no makefile-editing
   simply by running `make GTK_CONFIG="pkg_config gtk+-2.0"'
 - we use Pango to find appropriate fonts, which means the text in
   the puzzles actually (gasp!) adapts its size to the
   circumstances. Unfortunately, I've been unable to do this
   portably without depending on _either_ a Pango function that
   isn't present in older versions _or_ the underlying window system
   being X11; I'd appreciate someone doing better.

[originally from svn r5693]

19 years agoCommand-line solver was dividing up non-square puzzles the wrong way
Simon Tatham [Wed, 27 Apr 2005 16:59:18 +0000 (16:59 +0000)]
Command-line solver was dividing up non-square puzzles the wrong way
round.

[originally from svn r5691]

19 years agoPattern also contains an internal solver, so here's a command-line
Simon Tatham [Wed, 27 Apr 2005 11:13:33 +0000 (11:13 +0000)]
Pattern also contains an internal solver, so here's a command-line
interface to it just in case it comes in handy.

[originally from svn r5686]

19 years agoThe command-line solver's difficulty gradings now use the same
Simon Tatham [Tue, 26 Apr 2005 17:03:56 +0000 (17:03 +0000)]
The command-line solver's difficulty gradings now use the same
terminology as the puzzle GUI.

[originally from svn r5683]

19 years agoImplement the remaining modes of reasoning in nsolve, and thus
Simon Tatham [Tue, 26 Apr 2005 11:19:00 +0000 (11:19 +0000)]
Implement the remaining modes of reasoning in nsolve, and thus
enable configurable puzzle difficulty. I'm only generating grids up
to Times level (complicated non-recursive analysis but guessing
never required); I wouldn't object to providing a Telegraph
difficulty level (guessing required) but it turns out to be very
hard indeed to generate at random. I might still add it later
(probably under the name `Unreasonable' :-) if I can think of an
efficient way to find them.

[originally from svn r5682]

19 years agoFix trivial UI glitch involving clicking on the border outside the
Simon Tatham [Tue, 26 Apr 2005 08:20:25 +0000 (08:20 +0000)]
Fix trivial UI glitch involving clicking on the border outside the
grid. I'm really starting to get annoyed by the default round-
towards-zero behaviour of C integer division.

[originally from svn r5681]

19 years agoRemove some legacy debugging code which was breaking the MinGW
Simon Tatham [Mon, 25 Apr 2005 14:37:33 +0000 (14:37 +0000)]
Remove some legacy debugging code which was breaking the MinGW
build, and which is redundant given the new standalone solver
functionality.

[originally from svn r5672]

19 years agoOops; _actually_ add the reasoning mode I mentioned in the last
Simon Tatham [Mon, 25 Apr 2005 14:17:14 +0000 (14:17 +0000)]
Oops; _actually_ add the reasoning mode I mentioned in the last
checkin.

[originally from svn r5671]

19 years agoVarious changes prompted by my boss taking an interest:
Simon Tatham [Mon, 25 Apr 2005 14:03:53 +0000 (14:03 +0000)]
Various changes prompted by my boss taking an interest:
 - added a compilation option -DSTANDALONE_SOLVER which makes both
   of Solo's internal solvers accessible from the command line.
 - fix a bug in nsolve turned up by testing in this mode: it failed
   to iterate at all! Oddly, this massive improvement to the
   effectiveness of nsolve hasn't emptied the generated grids by
   very much.
 - add an extra mode of reasoning to my to-do list (which is the
   dual of one already there, so I'm kicking myself).

[originally from svn r5670]

19 years agoOutstandingly cute mathematical transformation which allows me to
Simon Tatham [Sun, 24 Apr 2005 10:06:47 +0000 (10:06 +0000)]
Outstandingly cute mathematical transformation which allows me to
lose a lot of code duplication in nsolve while preserving efficiency.

[originally from svn r5667]

19 years agoIntroduce configurable symmetry type in generated puzzles, and drop
Simon Tatham [Sun, 24 Apr 2005 09:21:57 +0000 (09:21 +0000)]
Introduce configurable symmetry type in generated puzzles, and drop
the default symmetry from order-4 down to order-2, which seems to
mitigate the excessively-full-grid problem by permitting more
freedom to remove stuff.

[originally from svn r5666]

19 years agoDoc tweaks for Solo.
Jacob Nevins [Sat, 23 Apr 2005 18:16:54 +0000 (18:16 +0000)]
Doc tweaks for Solo.

[originally from svn r5665]

19 years agoSimplify clip region handling under Windows, which also makes Solo's
Simon Tatham [Sat, 23 Apr 2005 17:09:19 +0000 (17:09 +0000)]
Simplify clip region handling under Windows, which also makes Solo's
clipping policy work properly. I haven't proved why it didn't work
the previous way, but I have a good guess: I think that clip regions
are handled by reference. So I saved the old clip region out of the
DC, then did an IntersectClipRect, and then selected the old clip
region back in again - but the old clip region had never been
_de_-selected, because IntersectClipRect didn't change which object
was selected but rather it modified-in-place the one that already
was selected. So my attempt to restore the old clip region did
nothing whatsoever, and thus clipping to two different rectangles
during the same draw sequence failed. Now I'm completely destroying
the clip region during unclip(), which seems to work better.

[originally from svn r5662]

19 years agoVisual C points out a couple of typos.
Simon Tatham [Sat, 23 Apr 2005 16:41:35 +0000 (16:41 +0000)]
Visual C points out a couple of typos.

[originally from svn r5661]

19 years agoInitial checkin of `Solo', the number-placing puzzle popularised by
Simon Tatham [Sat, 23 Apr 2005 16:35:28 +0000 (16:35 +0000)]
Initial checkin of `Solo', the number-placing puzzle popularised by
the Times under the name `Sudoku'.

[originally from svn r5660]