During an interactive drag of an arrow, it's possible to put the mouse
pointer precisely on the pixel the arrow is pointing at. When that
happens, draw_arrow computes the zero-length vector from the pointer
to the target pixel, and tries to normalise it to unit length by
dividing by its length. Of course, this leads to computing 0/0 = NaN.
Depending on the platform, this can cause different effects.
Conversion of a floating-point NaN to an integer is not specified by
IEEE 754; on some platforms (e.g. Arm) it comes out as 0, and on
others (e.g. x86), INT_MIN.
If the conversion delivers INT_MIN, one possible effect is that the
arrow is drawn as an immensely long diagonal line, pointing upwards
and leftwards from the target point. To add to the confusion, that
line would not immediately appear on the display in full, because of
the draw_update system. But further dragging-around of arrows will
gradually reveal it as draw_update rectangles intersect the corrupted
display area.
However, that diagonal line need not show up at all, because once
draw_arrow has accidentally computed a lot of values in the region of
INT_MIN, it then adds them together, causing signed integer overflow
(i.e. undefined behaviour) to confuse matters further! In fact, this
diagonal-line drawing artefact has only been observed on the
WebAssembly front end. The x86 desktop platforms might very plausibly
have done it too, but in fact they didn't (I'm guessing because of
this UB issue, or else some kind of clipping inside the graphics
library), which is how we haven't found this bug until now.
Having found it, however, the fix is simple. If asked to draw an arrow
from a point to itself, take an early return from draw_arrow before
dividing by zero, and don't draw anything at all.
Trivially, no edge of this kind can be part of any legal solution, so
it's a minor UI affordance that doesn't spoil any of the puzzly
thinking to prevent the user accidentally placing them in the first
place.
Suggestion from Larry Hastings, although this is not his patch.
I don't know how I've never thought of this before! Pretty much every
game in this collection has to have a mechanism for noticing when
game_redraw is called for the first time on a new drawstate, and if
so, start by covering the whole window with a filled rectangle of the
background colour. This is a pain for implementers, and also awkward
because the drawstate often has to _work out_ its own pixel size (or
else remember it from when its size method was called).
The backends all do that so that the frontends don't have to guarantee
anything about the initial window contents. But that's a silly
tradeoff to begin with (there are way more backends than frontends, so
this _adds_ work rather than saving it), and also, in this code base
there's a standard way to handle things you don't want to have to do
in every backend _or_ every frontend: do them just once in the midend!
So now that rectangle-drawing operation happens in midend_redraw, and
I've been able to remove it from almost every puzzle. (A couple of
puzzles have other approaches: Slant didn't have a rectangle-draw
because it handles even the game borders using its per-tile redraw
function, and Untangle clears the whole window on every redraw
_anyway_ because it would just be too confusing not to.)
In some cases I've also been able to remove the 'started' flag from
the drawstate. But in many cases that has to stay because it also
triggers drawing of static display furniture other than the
background.
Adding an association with an out-of-bounds square (i.e. by pressing Return
with a dot selected, and then moving the cursor so the `opposite' arrow was
off the screen) would cause space_opposite_dot() to return NULL, in turn
causing ok_to_add_assoc_with_opposite_internal() to return false, failing
the assertion.
This assertion appears to have been introduced in 68363231.
The Rockbox frontend allows games to be displayed in a "zoomed-in"
state targets with small displays. Currently we use a modal interface
-- a "viewing" mode in which the cursor keys are used to pan around
the rendered bitmap; and an "interaction" mode that actually sends
keys to the game.
This commit adds a midend_get_cursor_location() function to allow the
frontend to retrieve the backend's cursor location or other "region of
interest" -- such as the player location in Cube or Inertia.
With this information, the Rockbox frontend can now intelligently
follow the cursor around in the zoomed-in state, eliminating the need
for a modal interface.
If you drag an arrow on to a square which is already filled in as part
of a completed region, or whose counterpart is filled in, or whose
counterpart is actually a dot, then the game can't actually place a
double arrow. Previously, it didn't find that out until execute_move
time, at which point it was too late to prevent a no-op action from
being placed on the undo chain.
Now we do those checks in interpret_move, before generating the move
string that tries to place the double arrow in the first place. So
execute_move can now enforce by assertion that arrow-placement moves
it gets are valid.
Another thing I spotted while trawling the whole source base was that
a couple of games had omitted 'static' on a lot of their internal
functions. Checking with nm, there turned out to be quite a few more
than I'd spotted by eye, so this should fix them all.
Also added one missing 'const', on the lookup table nbits[] in Tracks.
This is the main bulk of this boolification work, but although it's
making the largest actual change, it should also be the least
disruptive to anyone interacting with this code base downstream of me,
because it doesn't modify any interface between modules: all the
inter-module APIs were updated one by one in the previous commits.
This just cleans up the code within each individual source file to use
bool in place of int where I think that makes things clearer.
This commit removes the old #defines of TRUE and FALSE from puzzles.h,
and does a mechanical search-and-replace throughout the code to
replace them with the C99 standard lowercase spellings.
encode_params, validate_params and new_desc now take a bool parameter;
fetch_preset, can_format_as_text_now and timing_state all return bool;
and the data fields is_timed, wants_statusbar and can_* are all bool.
All of those were previously typed as int, but semantically boolean.
This commit changes the API declarations in puzzles.h, updates all the
games to match (including the unfinisheds), and updates the developer
docs as well.
This function gives the front end a way to find out what keys the back
end requires; and as such it is mostly useful for ports without a
keyboard. It is based on changes originally found in Chris Boyle's
Android port, though some modifications were needed to make it more
flexible.
This allows me to use different types for the mutable, dynamically
allocated string value in a C_STRING control and the fixed constant
list of option names in a C_CHOICES.
Now midend.c directly tests the returned pointer for equality to this
value, instead of checking whether it's the empty string.
A minor effect of this is that games may now return a dynamically
allocated empty string from interpret_move() and treat it as just
another legal move description. But I don't expect anyone to be
perverse enough to actually do that! The main purpose is that it
avoids returning a string literal from a function whose return type is
a pointer to _non-const_ char, i.e. we are now one step closer to
being able to make this code base clean under -Wwrite-strings.
To do this, I've completely replaced the API between mid-end and front
end, so any downstream front end maintainers will have to do some
rewriting of their own (sorry). I've done the necessary work in all
five of the front ends I keep in-tree here - Windows, GTK, OS X,
Javascript/Emscripten, and Java/NestedVM - and I've done it in various
different styles (as each front end found most convenient), so that
should provide a variety of sample code to show downstreams how, if
they should need it.
I've left in the old puzzle back-end API function to return a flat
list of presets, so for the moment, all the puzzle backends are
unchanged apart from an extra null pointer appearing in their
top-level game structure. In a future commit I'll actually use the new
feature in a puzzle; perhaps in the further future it might make sense
to migrate all the puzzles to the new API and stop providing back ends
with two alternative ways of doing things, but this seemed like enough
upheaval for one day.
At minimum size (3x3) Galaxies can generate a pre-solved single dot
game.
You have to add and remove a line to get the victory flash which is a
bit weird, so just prevent this.
puzzle backend function which ought to have it, and propagate those
consts through to per-puzzle subroutines as needed.
I've recently had to do that to a few specific parameters which were
being misused by particular puzzles (r9657, r9830), which suggests
that it's probably a good idea to do the whole lot pre-emptively
before the next such problem shows up.
[originally from svn r9832]
[r9657 == 3b250baa02a7332510685948bf17576c397b8ceb]
[r9830 == 0b93de904a98f119b1a95d3a53029f1ed4bfb9b3]
new_desc. Oddities in the 'make test' output brought to my attention
that a few puzzles have been modifying their input game_params for
various reasons; they shouldn't do that, because that's the
game_params held permanently by the midend and it will affect
subsequent game generations if they modify it. So now those arguments
are const, and all the games which previously modified their
game_params now take a copy and modify that instead.
[originally from svn r9830]
basically just so that it can divide mouse coordinates by the tile
size, but is definitely not expected to _write_ to it, and it hadn't
previously occurred to me that anyone might try. Therefore,
interpret_move() now gets a pointer to a _const_ game_drawstate
instead of a writable one.
All existing puzzles cope fine with this API change (as long as the
new const qualifier is also added to a couple of subfunctions to which
interpret_move delegates work), except for the just-committed Undead,
which somehow had ds->ascii and ui->ascii the wrong way round but is
otherwise unproblematic.
[originally from svn r9657]
midend_status(), and given it three return codes for win, (permanent)
loss and game-still-in-play. Depending on what the front end wants to
use it for, it may find any or all of these three states worth
distinguishing from each other.
(I suppose a further enhancement might be to add _non_-permanent loss
as a fourth distinct status, to describe situations in which you can't
play further without pressing Undo but doing so is not completely
pointless. That might reasonably include dead-end situations in Same
Game and Pegs, and blown-self-up situations in Mines and Inertia.
However, I haven't done this at present.)
[originally from svn r9179]
state is in a solved position, and a midend function wrapping it.
(Or, at least, a situation in which further play is pointless. The
point is, given that game state, would it be a good idea for a front
end that does that sort of thing to proactively provide the option to
start a fresh game?)
[originally from svn r9140]
_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]
amount of code. James has ripped out the solver's version of
check_complete(), in favour of using the one I wrote for the
game-playing UI. My one checks connectedness, which means that the
solver will now not believe non-solutions to puzzles where
connectedness becomes a difficult issue. Examples of game IDs which
are now solved correctly but were previously not are 5x3:ubb and
7x7:ajfzmfqgtdzgt.
[originally from svn r7362]
binary from the Galaxies source file. The function of the new
`galaxiespicture' is to take a .xbm bitmap on standard input and
convert it into a Galaxies game ID using both black and white dots,
such that when solved the puzzle displays the input bitmap.
In the process of this I've implemented a post-processing pass after
the main game generation, to prevent clusters of adjacent
singletons. James H already solved that problem for unconstrained
game generation, but for some reason it came back when I did this.
However, the post-processing pass is still turned off for normal
usage, on the basis that (a) if it ain't broke don't fix it, and (b)
it's rather slow and best avoided if not necessary.
[originally from svn r7354]
- missing static in filling.c
- better robustness in execute_move() in filling.c
- remove side effects in assert statements
- remove rogue diagnostic in galaxies.c
- remove // comment in map.c
- add more stylus-friendly UI to Pattern
- bias Unequal towards generating inequality clues rather than numeric
[originally from svn r7344]
is mostly done with ifdefs in windows.c; so mkfiles.pl generates a
new makefile (Makefile.wce) and Recipe enables it, but it's hardly
any different from Makefile.vc apart from a few definitions at the
top of the files.
Currently the PocketPC build is not enabled in the build script, but
with any luck I'll be able to do so reasonably soon.
[originally from svn r7337]
