Op Mon, 20 Dec 2004 11:32:16 +0100 schreef Corinna Vinschen in <[EMAIL PROTECTED]>: : On Dec 17 21:59, Bas van Gompel wrote: : > Op Fri, 17 Dec 2004 10:43:01 +0100 schreef Corinna Vinschen : > : Hmm, if stderr is not unbuffered in mingw, then that should be fixed : > : in mingw, shouldn't it? : > : > I guess so... : > : > I'll try and look into this, if noone else does.
Danny does not seem inclined to apply my patch... I'll try once more. : > What about the patch? It shouldn't hurt, and the flush of stderr can : > be removed, once this has been fixed in mingw. : : I'd rather see a fix than a workaround, if that's possible. You know : that the typical lifetime of a workaround is a multiple of the lifetime : of the actual problem ;-) True. Here is a version which will work as expected iff the mingw-patch gets accepted. 2004-12-23 Bas van Gompel <[EMAIL PROTECTED]> * cygcheck.cc (eprintf): Flush stdout before output. (display_error): Use eprintf. --- src/winsup/utils/cygcheck.cc 18 Nov 2004 05:20:23 -0000 1.64 +++ src/winsup/utils/cygcheck.cc 22 Dec 2004 06:30:07 -0000 @@ -102,6 +102,9 @@ void eprintf (const char *format, ...) { va_list ap; + + fflush (stdout); + va_start (ap, format); vfprintf (stderr, format, ap); va_end (ap); @@ -114,10 +118,10 @@ static int display_error (const char *name, bool show_error = true, bool print_failed = true) { if (show_error) - fprintf (stderr, "cygcheck: %s%s: %lu\n", name, + eprintf ("cygcheck: %s%s: %lu\n", name, print_failed ? " failed" : "", GetLastError ()); else - fprintf (stderr, "cygcheck: %s%s\n", name, + eprintf ("cygcheck: %s%s\n", name, print_failed ? " failed" : ""); return 1; } L8r, Buzz. -- ) | | ---/ ---/ Yes, this | This message consists of true | I do not -- | | / / really is | and false bits entirely. | mail for ) | | / / a 72 by 4 +-------------------------------+ any1 but -- \--| /--- /--- .sigfile. | |perl -pe "s.u(z)\1.as." | me. 4^re