Collin Funk wrote:
> That would avoid linking to winpthreads and define nanosleep to
> rpl_nanosleep, I think.

Looking at modules/nanosleep and lib/time.in.h, yes, that is the outcome
that we want in this situation: set REPLACE_NANOSLEEP to 1.

> So would the correct fix to be to change the
> cross guess for "mingw*" to something other than "guessing no"?

This would not be sufficient. The new mingw with the 'static inline'
definition of nanosleep (that Kirill described) will be available
  - for compilation on Windows, via Cygwin and MSYS2,
  - for cross-compilation on Debian.
The changes to m4/nanosleep.m4 must cover both cases.

I think one needs to add a test whether <time.h> already defines
nanosleep as a 'static' function. This can be done through an
AC_COMPILE_IFELSE test; an AC_LINK_IFELSE test is not needed.

Bruno





Reply via email to