This touched a lot more code than initially expected, for two reasons.
First, this broke a big RGBASM assumption: that sections are always being
written to at their end. This plus other problems required touching
basically the entirety of `section.c`.
Second, I tried different solutions to solve the above problem, and along
the way I cleaned up many things around. (I believe that keeping this to
"cleanup" commits yields subpar results, and since it's boring they get
postponed anyways.)
RGBLINK support still needs to be added, but this will come next.
This should significantly improve performance: on pokecrystal builds, perf
reported as much CPU time spent on `yyparse` as on `sym_UseNewMacroArgs`
Measurements show ~6 seconds of improvement on that codebase.
This also fixes#321, as a bonus, due to saner management!
This was utterly stupid. The check right above ensured that `sym` was NULL,
ergo that the argument to `yyerror` *would* segfault.
The only two call sites cannot pass a non-NULL pointer anyways, which I'm
betting is why this went unnoticed.
I did what an optimizing compiler would do anyways: remove the dead code.
This causes it to auto-update whenever the current section's attributes are
updated, simplifying the code and eliminating redundancy.
This should also overall reduce overhead (one extra function call on each
PC evaluation, but less bookkeeping for each byte output)
Windows does not honor `%F` nor `%T` in `strftime`. These are worked around
by writing the full format they serve as a short for.
However, Windows also treats `%z` and `%Z` identically, where SUS instead
requires `%z` to output a ±XXXX offset.
Since the current information is broken (no information), this isn't *breaking*
anything, but at least provides something a human will probably understand.
`__ISO_8601_UTC__` is unaffected because it hardcodes the timezone character,
only `__ISO_8601_LOCAL__` suffers from this.
Stop using that bitfield for everything, including what can be determined otherwise
It also makes it easier to have a sane state, since some bits were (supposedly)
mutually exclusive
The old error stack was fairly obtuse and hard to use for debugging.
This improves it notably by ensuring all line numbers are relative
to the file, and not, say, the macro definition.
This is a breaking change if you were parsing the old stack, but
the change should be painless, and the new stack only brings more info.
The syntax is unchanged for files, macros see their name prefixed
with the file they're defined in and a pair of colors, REPT blocks
simply append a '::REPT~n' to the context they're in, where 'n' is
the number of iterations the REPT has done.
This is especially helpful in macro-heavy code such as rgbds-structs.
REPT blocks nested in macros (and possibly other cases) leaked
memory on every call. Unlike most other memory leaks, which would
be freed at the end of program execution if they were done properly,
those piled up the more compilation went on.
I believe memory usage could have started being fairly high on
large projects following the "one master file INCLUDEs all the rest"
so this may have actually been worth it.
sym_SetMacroArgID used a `sprintf` that could write no \0.
In practice this was benign because %u cannot print 256 chars,
but better future-proof this.
This adds two new directives: newcharmap and setcharmap.
newcharmap creates a new charmap and switches to it.
setcharmap switches to an existing charmap.
