Experimental Physics and
Industrial Control System

Freddie Akeroyd via Core-talk <[email protected]> · Tue, 16 Jul 2019 21:26:45 -0000

Hi Andrew, the machine that was for a while recently repeatably showing the failure was Window 10, previously we have had it appearing randomly on Windows 7 but then it would go away next build or on a manual rerun of msi. I was able to confirm an fclose(stdout) fixed it when it was misbehaving, I suspect that an fflush() is the key part of the fclose() operation so going with fflush(stdout) until we see further problems seems a reasonable approach. I'm not a great fan of freopen() and liked the idea of removing it as Michael had done, but keeping freopen() and using fflush() does make   "msi -o outfile"   and   "msi > outfile"   behave the same internally and so could be better for now.

Thanks for clarifying the opt_D behaviour, that looks the right thing to do. It was the fact the program was removing a file that it had not itself generated that made me cautious, but it is a makefile target and if it can't generate the dependencies I guess that means the reliability of any existing target file is in question and so should be removed as is done. 

-- 
https://code.launchpad.net/~freddie-akeroyd/epics-base/+git/epics-base/+merge/369763
Your team EPICS Core Developers is subscribed to branch epics-base:3.15.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System