[EMAIL PROTECTED] (Ludovic Courtès) writes: > > of `scm_getc ()' translates in one `scm_fill_input ()' call, which in > turn translates in one `select ()' and one `read ()'
Coming back to this bit: that select really does seem pretty doubtful. The more I think the more it sounds like a hangover from 1.6 where select was the only proper place to block if other threads were to keep running. But maybe now the right thing would be to have the scm_leave_guile or scm_without_guile or whatever around each read/write. Avoiding initializing an fd_set and whatnot for select would have to be a good thing. If so then there'd be other places a similar leave might be wanted, `copy-file' and `gethost' for instance (I might have mentioned them before). Oh well, I guess there's plenty of work for anyone who wants to investigate ... One thing I'm pretty sure about though is that the __MINGW32__ conditional on fport_wait_for_input can't be right. I think we had bug reports that mingw has pthreads now, so surely anything wanted for pthreads will be wanted on mingw (or new enough ones or whatever). Maybe the condition was meant to be "if threads" instead of "if not mingw". If configured with null-threads then there'd be no need to worry about where any blocking happens. _______________________________________________ Guile-devel mailing list Guile-devel@gnu.org http://lists.gnu.org/mailman/listinfo/guile-devel
