My Mac has just upgraded itself to include a version of clang which
warns if you use abs() on a floating-point value, or fabs() on an
integer. Fixed the two occurrences that came up in this build (and
which were actual build failures, because of -Werror), one in each
direction.
I think both were benign. The potentially dangerous one was using abs
in place of fabs in grid_find_incentre(), because that could actually
lose precision, but I think that function had plenty of precision to
spare (grid point separation being of the order of tens of pixels) so
nothing should have gone seriously wrong with the old code.
happening was that at the point of calling grid_make_consistent, the
grid had no faces or vertices, probably because grid_trim_vigorously
had removed them all, causing grid_make_consistent to try to allocate
a negative amount of memory and die in snewn.
Fixed by detecting this case in new_desc_penrose and retrying until
generation is successful. (It wasn't happening 100% of the time, just
_most_ of the time.) The same verification step is also used in
validate_desc_penrose in case a user manages to manually construct a
set of parameters leading to this failure mode.
[originally from svn r9840]
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]
meaning to revisit for a while. The new version generates more nicely
symmetric grids (best at even heights, but still improved even at odd
heights), and avoids any 'ears' on the grid (triangles at the corners
connected to only one other triangle, which tend to be boringly easy
places to start solving).
I've reused the grid-description-string mechanism invented for the
Penrose tilings as a versioning mechanism for the triangular grids, so
that old game descriptions should still be valid, and new triangular-
grid game descriptions begin with an "0_" prefix to indicate that they
are based on the new-style construction.
[originally from svn r9824]
4-vector representation, rather than mucking about with sines and
cosines after grid generation. _Should_ make no difference in the
generated grids (there's a theoretical risk of an unlucky rounding
error just about managing to push some point in or out of bounds, but
I think it's vanishingly small), but simplifies the coordinate-
flattening procedure, and in particular increases its chance of
getting vertical lines actually vertical.
(Prior to this change, the game ID
10x10t12:G2554,-31,108_a3b12h0a212a3d102b2a23a2e3b01b0a2c2a0c0 was
generating a not-quite-vertical edge at top left, in the Java port but
not on Linux; I suspect differences in sin and cos as the cause of the
discrepancy. With the rotation done like this, the points'
x-coordinates are now computed without reference to their
y-coordinates.)
[originally from svn r9168]
additions of missing 'static' and explicit 'void' in parameter lists,
plus one or two other things like explicitly casting chars in variadic
argument lists to int and using DBL_MAX if HUGE_VAL isn't available.
[originally from svn r9166]
to add two kinds of Penrose tiling to the grid types supported by
Loopy.
This has involved a certain amount of infrastructure work, because of
course the whole point of Penrose tilings is that they don't have to
be the same every time: so now grid.c has grown the capacity to
describe its grids as strings, and reconstitute them from those string
descriptions. Hence a Penrose Loopy game description consists of a
string identifying a particular piece of Penrose tiling, followed by
the normal Loopy clue encoding.
All the existing grid types decline to provide a grid description
string, so their Loopy game descriptions have not changed encoding.
[originally from svn r9159]
part that converts from abstract grid coordinates into screen
coordinates. This should speed up window-resizing by eliminating
pointless reiteration of the complicated part of the algorithm: now
when a game_drawstate is renewed, only the conversion into screen
coordinates has to be redone.
[originally from svn r9157]
overly complicated algorithm that uses it to home in on the grid edge
closest to a mouse click. That algorithm is being stressed beyond its
limit by the new grid type, and it's unnecessary anyway given that no
sensibly sized puzzle grid is going to be big enough to make it
prohibitively expensive just to do the trivial approach of iterating
over all edges and finding the closest of the eligible ones.
[originally from svn r9108]
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]
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]
