new function in the drawing API which permits the display of text
from outside basic ASCII. A fallback mechanism is provided so that
puzzles can give a list of strings they'd like to display in order
of preference and the system will return the best one it can manage;
puzzles are required to cope with ASCII-only front ends.
[originally from svn r8793]
any equivalence class is always the element with the smallest index.
This is slower (the previous behaviour, suggested by Jonas Koelker,
was to choose the new root element to maximise performance), but
still more than acceptably fast and more useful.
[originally from svn r8792]
programs having to clone the latin_solver() function and insert
their own extra deduction routines, they can now just _call_
latin_solver with enough parameters to let it fit its own deductions
into their difficulty framework and call a set of provided function
pointers to do user deductions. Modified Unequal to work in the new
world, of course.
[originally from svn r8791]
the < and > clues are replaced by bars separating every pair of
squares whose contents differ by exactly 1. Unlike standard Unequal,
which presents only a subset of the available clues, in Adjacent the
clues are _all_ present, so you can deduce from the absence of a bar
that the two numbers it would separate are _not_ consecutive.
[originally from svn r8790]
Guess, because Guess expected ^H whereas GTK generated ^?. Other
puzzles that use Backspace do it by being prepared to see either,
which seems wasteful. Now the midend normalises both into ^H, so
front ends can generate whichever they like while puzzles can
safely just look for ^H.
[originally from svn r8786]
buttons in several games if STYLUS_BASED is defined: in games where
you can set a puzzle element to 'on', 'off' or 'not yet set', when
it's hard to mimic a second mouse button, it's better to have the
one 'button' cycle between all three states rather than from 'on'
back to 'unset'.
[originally from svn r8784]
compiling gtk.c with STYLUS_BASED defined: most notably, it provides
a row of numeric 'keys' made of GTK button widgets for puzzles that
want a numeric keypad.
[originally from svn r8783]
in r8414 as a side effect of reorganisation in interpret_move().
Reinstate them.
[originally from svn r8692]
[r8414 == 54801998744a69786a49ebe0688942ae7b57525b]
notice that there was an _official_ way to have menu items marked as
selected or unselected, which is the GtkRadioMenuItem class. Replace
my ad-hoc use of bullet characters with that.
[originally from svn r8601]
[r7980 == ae6c7381272f84c0fd398f6d3e496f571ba42ec8]
available, which magically seems to eliminate most if not all of the
update lag when dragging a point in a large (say 250-point)
Untangle. I don't fully understand how this works, but it seems to!
[originally from svn r8600]
on labels in tables. The effect is that if you resize (e.g.) Solo's
configuration panel, the labels to the left of the various input
boxes will remain just big enough to fit their text in, and all the
extra space will be taken up by the useful part, i.e. the input
boxes themselves. Previously the labels and input boxes would expand
in a 1:1 ratio, causing lots of space to the right of the label text
that wasn't used for anything.
[originally from svn r8599]
Loopy mark LINE_NO grid edges with very faint lines, instead of
leaving them totally undrawn. Helps in complex grid types where the
line layout isn't entirely obvious to a player not already familiar
with it. Disableable by setting LOOPY_FAINT_LINES=n in the
environment, just in case anyone turns out to seriously dislike it.
(You could probably disable it via LOOPY_COLOUR_6 too, but you'd
have to know the exact shade of your background to get that right.)
[originally from svn r8597]
merge_some_cages() was written in the assumption that it would
always be able to do something, in that it returned void on success
and if it couldn't find anything to do it would just loop round
forever trying the same things over and over again.
Now it makes a methodical list of the pairs of cages which are merge
candidates, goes through them in a random order until it finds a
viable one, and returns a boolean indicating whether it succeeded or
ran out of candidates.
A test case which previously hung and now does not is "solo
--generate 1 7jxkdt#12345-10".
[originally from svn r8541]
of each digit, and - perhaps more importantly - the display code
wasn't highlighting violations of that rule as an error. Fix both.
[originally from svn r8540]
the 'Killer Sudoku' puzzle type. As a side effect I've had to
increase the default tile size of Solo, so that the extra numbers
drawn in the squares in Killer mode were still legible.
[originally from svn r8455]
ALIGN_VNORMAL is in use: ALIGN_VNORMAL indicates that the supplied
y-coordinate denotes the _baseline_ of the text, not its top, so
adding on 'asc' to convert to the baseline is wrong.
This only affects Tents, at present.
[originally from svn r8452]
structure in every game_state is terribly wasteful. Move all the
constant bits of it (i.e. everything except the per-square 'blue'
flag) into a ref-counted shared structure.
[originally from svn r8444]
