Re: AC_FUNC_MMAP test fails on AIX for MAP_FIXED: who's at fault?

On 2024-07-29 06:36, Zack Weinberg wrote:
> However, perhaps we could have a new family of
> AC_FUNC_MMAP_* macros that check for specific subsets of mmap functionality
> that are generally useful

That could be trouble for cross-compiles, no?

For cross-compiles AC_FUNC_MMAP currently guesses that mmap works on 
Linux kernels and fails everywhere else. This is obviously wrong but I 
guess good enough for programs that use AC_FUNC_MMAP. Not sure I'd like 
to maintain guesses for a more-sophisticated set of mmap configuration 
macros.

Anyway, if you want to head in this direction I suggest talking to Bruno 
Haible for advice and looking at Gnulib for prior art, e.g., 
gl_FUNC_MMAP_ANON sounds like it's the sort of macro that Yury was 
asking for. A nice property of gl_FUNC_MMAP_ANON is that it doesn't do 
runtime checks. (Don't know why it's so complicated, though; you'd have 
to ask Bruno about that.)