Re: [Mingw-w64-public] time guard

2015-01-29 Thread Erik van Pienbroek
Martell Malone schreef op do 29-01-2015 om 15:23 [+]: > This change makes sense. Right now Eric is checking this > change on > Fedora. So we should wait for his results. > Yes but I think he should also build on rc1 without the patch to make > sure that if there is a

Re: [Mingw-w64-public] time guard

2015-01-29 Thread Martell Malone
> > This change makes sense. Right now Eric is checking this change on > Fedora. So we should wait for his results. > Yes but I think he should also build on rc1 without the patch to make sure that if there is a breakage this is the cause. Btw Martell, the fopen_s emulation-function would be som

Re: [Mingw-w64-public] time guard

2015-01-29 Thread Kai Tietz
This change makes sense. Right now Eric is checking this change on Fedora. So we should wait for his results. This might be something we would like to back-merge later to the 4.x branch. Btw Martell, the fopen_s emulation-function would be something good IMO. We should consider then emulation a

[Mingw-w64-public] time guard

2015-01-28 Thread Martell Malone
This should fix the issue of windows programs assuming localtime_r and gmtime_r unavailable. if _POSIX_C_SOURCE is defined then we are specifically asking for them. Quote from http://pubs.opengroup.org/onlinepubs/007904975/functions/xsh_chap02_02.html "A POSIX-conforming application should ensur