#general
- m
most-cricket-96651
10/08/2015, 12:33 AMYeah, now that I think of it, how about you let me write up my thoughts for build tools specifically I think that would guide it even more. - m
most-cricket-96651
10/08/2015, 12:33 AMSo let me get back to you early next week with something more concrete. (I have a travel day tomorrow and then a half day on Friday) - b
bored-art-40741
10/08/2015, 2:40 AM@witty-crayon-22786: re: https://rbcommons.com/s/twitter/r/2940/ - b
bored-art-40741
10/08/2015, 2:41 AMOur standard procedure is to clean before all CI runs, but to let the CIs both populate and consume from a single remote artifact cache - b
bored-art-40741
10/08/2015, 2:41 AMDoes this change affect that behavior? - b
bored-art-40741
10/08/2015, 2:44 AMIn particular I'm curious about what happens if we get a cache hit in the initial full graph check. At that point, is the compile of the target considered "incremental"? Or I guess we never get to this code path if we've already got a cache hit for the target - b
bored-art-40741
10/08/2015, 2:45 AMYeah never mind, this seems fine. Lots of conditions and ways to not hit this code path to keep in my head - w
witty-crayon-22786
10/08/2015, 4:46 AM@bored-art-40741: if the CI builds multiple targets, there are chances for incremental builds - w
witty-crayon-22786
10/08/2015, 4:46 AMeven with the same sources - w
witty-crayon-22786
10/08/2015, 4:47 AMin our case it happens when sources are built with a different classpath because their ivy resolve has changed - w
witty-crayon-22786
10/08/2015, 4:48 AMwhich is why i mentioned that we will turn this on in CI... we want to still write those, because the primary thing we're trying to avoid is incremental compilation after source changes - w
witty-crayon-22786
10/08/2015, 4:49 AMan incremental compilation due to a classpath change is significantly less risky - b
bored-art-40741
10/08/2015, 2:32 PMIt's impossible for us to build the same target with a different classpath in a given CI run since our classpaths are fixed we always bump the cache key when they change - b
bored-art-40741
10/08/2015, 3:18 PMSo I'd imagine we're safe from that case - e
enough-analyst-54434
10/08/2015, 7:08 PMSounds good - thanks @most-cricket-96651 - d
dry-analyst-73584
10/08/2015, 7:55 PMAnyone have any idea why this might be happening on master?Copy code$ ./pants goals Bootstrapping pants_deps with requirements: /Users/tansy/code/pants/3rdparty/python/requirements.txt /Users/tansy/code/pants/3rdparty/python/twitter/commons/requirements.txt /Users/tansy/code/pants/pants-plugins/3rdparty/python/requirements.txt Failed to create venv. - b
billowy-alarm-80641
10/08/2015, 7:56 PMDid you try
?git clean -fdx - d
dry-analyst-73584
10/08/2015, 7:57 PMjust now, still no luck though - b
billowy-alarm-80641
10/08/2015, 7:58 PMHmm, I think I have seen this before, but I forget what the fix was - b
billowy-alarm-80641
10/08/2015, 7:59 PMMaybe try editing
, and putting abuild-support/pants_venv
in there?set +x - d
dry-analyst-73584
10/08/2015, 8:02 PMset +x? - b
billowy-alarm-80641
10/08/2015, 8:02 PMsorry, -x - b
billowy-alarm-80641
10/08/2015, 8:03 PM - b
billowy-alarm-80641
10/08/2015, 8:03 PMI always get the +/- mixed up there. 😐 - b
billowy-alarm-80641
10/08/2015, 8:03 PM
in a shell script will tell bash to print every command executedset -x - b
billowy-alarm-80641
10/08/2015, 8:03 PMshould help to narrow down the problem - d
dry-analyst-73584
10/08/2015, 8:03 PMawesome, thanks! - d
dry-analyst-73584
10/08/2015, 8:06 PMI think I found it:Copy codeGCC_VERSION=' Agreeing to the Xcode/iOS license requires admin privileges, please re-run as root via sudo.’ - b
billowy-alarm-80641
10/08/2015, 8:06 PMoh yeah - e
enough-analyst-54434
10/08/2015, 8:06 PMEl Cap?