#general
- r
rough-minister-58256
11/03/2015, 12:04 AMic - e
enough-analyst-54434
11/03/2015, 12:04 AMOthers have more commits-per-week than I now already, which is a milestone I've been looking forward to - e
enough-analyst-54434
11/03/2015, 12:06 AMLooks like OSX is going green - the #4 shard at least is green now and has failed > 10x in a row prior - r
rough-minister-58256
11/03/2015, 12:07 AMso then for pantsd, we should ship sans BG-cache for 1.0, then engine.BG, /then/ engine.BG-cache? - e
enough-analyst-54434
11/03/2015, 12:07 AMBG-cache is a risk judgement - r
rough-minister-58256
11/03/2015, 12:08 AMyeah - e
enough-analyst-54434
11/03/2015, 12:08 AMI have not read Sergei's change yet - w
witty-crayon-22786
11/03/2015, 12:08 AMi don't think it is currently possible to cache the buildgraph without also invalidating the options - w
witty-crayon-22786
11/03/2015, 12:08 AMso options caching/invalidation block the buildgraph anyway, no? - e
enough-analyst-54434
11/03/2015, 12:09 AMNot to mention - if real - Dan Bently's anecdote about restart watchman - e
enough-analyst-54434
11/03/2015, 12:09 AMas the new clean-all - r
rough-minister-58256
11/03/2015, 12:09 AMwell, to be fair thats based on our internal git fork’s usage of watchman. of which it is using one-major-version-lower version - e
enough-analyst-54434
11/03/2015, 12:09 AMaha - r
rough-minister-58256
11/03/2015, 12:09 AMthe buck guys have put in a ton of work on watchman - r
rough-minister-58256
11/03/2015, 12:10 AMhave seen numerous bugs fixed - e
enough-analyst-54434
11/03/2015, 12:10 AMI don't doubt it - they would be motivated - r
rough-minister-58256
11/03/2015, 12:10 AMand we reap the benefit - e
enough-analyst-54434
11/03/2015, 12:11 AMAt any rate - for OSS, I don't care about big perf wins for 1.0 branch so much, Square would be a mojor consumer of 1.0 for the next ~1 year and they are happy enough - r
rough-minister-58256
11/03/2015, 12:11 AMso, I think watchman can work as long as folks (incl us going forward) continue to be vigilant about fixing the issues that do crop up - r
rough-minister-58256
11/03/2015, 12:11 AMalso, the reason we have to killall watchman alot - e
enough-analyst-54434
11/03/2015, 12:12 AMYes, I'd want us to be good contributors and stay aggressively up2date - r
rough-minister-58256
11/03/2015, 12:12 AMis apparently not related to watchman at all - but a fallback behavior that our forked git has - r
rough-minister-58256
11/03/2015, 12:12 AMrelated to population of the fs_cache etc. - e
enough-analyst-54434
11/03/2015, 12:12 AMBut this all sounds by definition like 2.0 / post branch stuff to me - e
enough-analyst-54434
11/03/2015, 12:12 AMie: we can be aggresive post-branch - r
rough-minister-58256
11/03/2015, 12:12 AM2.0 == engine? - e
enough-analyst-54434
11/03/2015, 12:13 AM2.0 == all things very new, new engine, task apis, BUILD api, daemon - e
enough-analyst-54434
11/03/2015, 12:14 AM3.0 == we throw away 1/2 our code because sploot is out there and installable on Joe avergae's machine - w
witty-crayon-22786
11/03/2015, 12:14 AMan argument could perhaps be made for daemon in 1.0 if it didn't actually cache anything risky probably? - e
enough-analyst-54434
11/03/2015, 12:14 AMThat's what I was suggesting above