#general
- b
bored-art-40741
09/21/2015, 6:11 PMYup API section - w
witty-crayon-22786
09/21/2015, 6:11 PMi don't think that's new though. - b
bored-art-40741
09/21/2015, 6:12 PMJust reading through here trying to find the big stuff - w
witty-crayon-22786
09/21/2015, 6:13 PMi've been wrong before about the effects of global -> isolated on analysis - w
witty-crayon-22786
09/21/2015, 6:13 PMif we have analysis for a classpath entry, we treat it as internal - b
bored-art-40741
09/21/2015, 6:13 PMcc @benjy on this convo - b
bored-art-40741
09/21/2015, 6:13 PM@happy-kitchen-89482: - b
bored-art-40741
09/21/2015, 6:14 PMLet's see if I can find an old artifact - b
bored-art-40741
09/21/2015, 6:14 PMTo compare - h
happy-kitchen-89482
09/21/2015, 6:14 PMwhat what? - h
happy-kitchen-89482
09/21/2015, 6:16 PMUnder isolated we're not doing any split/merge any more, right? - b
bored-art-40741
09/21/2015, 6:16 PMCopy code$ du -h --max-depth=1 1008K ./classes 3.0M ./analysis_tmpdir - b
bored-art-40741
09/21/2015, 6:16 PMThat's an old one - b
bored-art-40741
09/21/2015, 6:16 PMHonestly not that much smaller - b
bored-art-40741
09/21/2015, 6:16 PMMaybe something else is at play here - w
witty-crayon-22786
09/21/2015, 6:22 PM@bored-art-40741: simple example: - w
witty-crayon-22786
09/21/2015, 6:22 PMCopy codeArchive: blah.jar Length Method Size Ratio Date Time CRC-32 Name -------- ------ ------- ----- ---- ---- ------ ---- 0 Defl:N 2 0% 09-21-15 11:21 00000000 META-INF/ 68 Defl:N 67 2% 09-21-15 11:21 5fdf217d META-INF/MANIFEST.MF 0 Stored 0 0% 09-21-15 11:20 00000000 compile/ 0 Stored 0 0% 09-21-15 11:20 00000000 compile/jvm/ 0 Stored 0 0% 09-21-15 11:20 00000000 compile/jvm/zinc/ 0 Stored 0 0% 09-21-15 11:20 00000000 compile/jvm/zinc/isolated-analysis/ 19670 Defl:N 4379 78% 09-18-15 18:30 531db062 compile/jvm/zinc/isolated-analysis/wilyns.common.src.main.scala.com.twitter.wilyns.common.common.analysis.portable 0 Stored 0 0% 09-21-15 11:20 00000000 compile/jvm/zinc/isolated-classes/ 2643 Defl:N 2279 14% 09-18-15 18:30 e9604683 compile/jvm/zinc/isolated-classes/wilyns.common.src.main.scala.com.twitter.wilyns.common.common.jar 0 Stored 0 0% 09-21-15 11:20 00000000 compile/jvm/zinc/isolated-logs/ 3858 Defl:N 1021 74% 09-18-15 18:30 87b1c800 compile/jvm/zinc/isolated-logs/wilyns.common.src.main.scala.com.twitter.wilyns.common.common.log -------- ------- --- ------- 26239 7748 71% 11 files - w
witty-crayon-22786
09/21/2015, 6:22 PMso the analysis compresses by 78% - w
witty-crayon-22786
09/21/2015, 6:23 PM(at least in a jar... i know these are usually tar.gzs) - b
bored-art-40741
09/21/2015, 6:23 PMI'm no longer surprised by the ratio or size of the artifacts compared to the old ones, but the perf is bad - b
bored-art-40741
09/21/2015, 6:23 PMI'm wondering if we're doing something in the main thread we didn't used to - b
bored-art-40741
09/21/2015, 6:23 PMThis step should be faster, since as Benjy said, we no longer need to split/merge - w
witty-crayon-22786
09/21/2015, 6:24 PMyea - b
bored-art-40741
09/21/2015, 6:24 PMMaybe we're rebasing outside of the worker? - w
witty-crayon-22786
09/21/2015, 6:24 PMfwiw, the relativize is still happening - w
witty-crayon-22786
09/21/2015, 6:24 PMer, yea... what you said. - w
witty-crayon-22786
09/21/2015, 6:26 PM@bored-art-40741: just to rule out the obvious thing... might the fact that all of foursquare's builds are suddenly cold have something to do with this - b
bored-art-40741
09/21/2015, 6:26 PMVery unlikely - w
witty-crayon-22786
09/21/2015, 6:27 PM...and it might stabilize later? - w
witty-crayon-22786
09/21/2015, 6:27 PMk