to have some means of automatically spotting this sort of problem on
a desktop platform, but I can't immediately think of one; building a
trick compiler which thinks `int' is 16 bits would be the obvious
option, but it would immediately break the ABIs to all the system
functions.
[originally from svn r6184]
slash in this square is red, so it does indeed imply that some of
the edge markings are also red; but it doesn't mean _all_ the edge
markings must be red. So instead of assuming ERRSLASH implies all
edge error markers, we set the correct set of edge error markers at
the same time as setting ERRSLASH, at which point we know which kind
of slash it is so we know which ones to set.
[originally from svn r6175]
in red if it's impossible to fulfill them (either through too many
neighbours connecting to them, or too many not connecting to them),
and edges are highlighted in red if they form part of a loop.
In order to do this I've had to revamp the redraw function
considerably. Each square is now drawn including its top and left
grid edges, but _not_ its bottom or right ones - which means that I
need to draw an extra strip of empty squares outside the actual grid
in order to draw the few pixels which appear on the grid bottom and
right borders and also to red-highlight border clues.
[originally from svn r6174]
to reverse the effect of the mouse buttons. Gareth has been
complaining about this for days: apparently he finds precisely the
opposite control system intuitive to me.
This is a horrendous hack, and pushes me one step closer to losing
my temper and designing a proper preferences architecture.
[originally from svn r6168]
Also in this checkin (committed by mistake - I meant to do it
separately), a behind-the-scenes change to Slant to colour the two
non-touching classes of diagonals in different colours. Both colours
are set to black by default, but configuration by way of
SLANT_COLOUR_* can distinguish them if you want.
[originally from svn r6164]
the return value rather than in *error. In the old days type-
checking would have caught this, but now of course they're the same
type.
[originally from svn r6161]
or two, a debugging fix, a couple of explicit initialisations of
variables that were previously read uninitialised, and a fix for a
whopping great big memory leak in Slant owing to me having
completely forgotten to write free_game().
[originally from svn r6159]
nikoli.co.jp (which has quite a few puzzles that they don't seem to
have bothered to translate into English).
Minor structural change: the disjoint set forest code used in the
Net solver has come in handy again, so I've moved it out into its
own module dsf.c.
[originally from svn r6155]
