As suggested by https://github.com/rednex/rgbds/issues/350#issuecomment-498030458
The order `count` then `byte` was decided after some discussion:
- First argument consistent with single-arg syntax
- Intuitive at least to some people other than myself
- Consistent with other assemblers, at least ca65
This avoids redundancy between them (and also having to port fixes and features)
The error messages have been preserved through a string reporting mechanism
The files being diffed (especially for variants) are temp files, so their
names are pretty nondescript. This improve error output, using ANSI escape
sequences to make those lines stand out.
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
This is actually not necessary, because RGBLINK would warn about missing labels.
Besides, through semi-esoteric ways, it is possible to define more labels in this scope,
and there's no reason to prevent that.
POSIX leaves undefined the order of output if stderr is injected into stdout,
and in practice it differs on Windows (Linux buffers both streams separately,
Windows interleaves them as they arrive without buffering).
This should help testing on other platforms
This is especially useful when an EQUS expands to another one, to help
track them.
This is done separately from the file stack as the EQUS stack is separate
(which is itself because EQUS are managed *way* differently).
Macro and rept buffers were not always being terminated with newlines
and/or were vulnerable to the final newline being escaped, allowing
buffer overflows to occur. Now, they are terminated with newlines using
the same mechanism as the file buffer.
Null characters in the middle of strings interact badly with the RGBDS
codebase, which assumes null-terminated strings. There is no reason to
support null characters in input source code, so the simplest way to deal
with null characters is to reject them early.
