it's not hard to have a shell script catch things like control-c I pressed control-c in my gcc-config 2 operation... well, it said it completed successfully, but.. yaknow, it didn't complete successfully. Reproducible: Always Steps to Reproduce: 1. 2. 3.
What should it do on control-c ? It might be that the user wanted to terminate it ... ?
not report that it switched the gcc... >What should it do on control-c ? It might be that the user wanted to terminate it ... ? >>well, it said it completed successfully, but.. yaknow, it didn't complete successfully.
i'd suggest we just do the samething as we do with /sbin/rc: trap ":" INT QUIT TSTP the user could possibly ctrl+c in the middle of the switch
Yes, but I cannot see the use, as what takes the longest (and will 99% of the time catch the ctrl-c) is env-update, and as that is python, or trap is overridden ... (which means ctrl-c will anyhow abort the whole thing)
The problem was, that it said the conversion completed... and it wasn't completed. gcc --version read the previous version, while gcc-config said it was the later version... (or.. actually they may have both said the previous version... but... gcc-config said that it has switched me to my new choice of compiler version... and it was in error... Catching a control-c isn't hard... in a shell script.. (well.. what I read anyway) Maybe gcc-config needs to check the output of the program that relinks? stuff... that caught the control-c during it's 99 percent of the time... And, 1 percent failure rate.. is pretty bad actually... (sorta paraphrasing what my dad said... I got him to switch one of his windows NT machines to gentoo...he said: Lots of minor problems) Catching things like control-c should be policy... The point is that people are going to run into these things, and not report them, they'll just reinstall their whole system again with the correct compiler, or... reinstall windowsXP!
1.3.7 does this