[00:10:44] bd808: any chance for https://gerrit.wikimedia.org/r/#/c/224520/ after all? :) [00:10:51] oh heh [00:10:53] you just replied [00:10:57] :) [00:15:10] bd808: a bit gross, but it'll have to be glob.glob('*.[pc][hd][pb]') [00:15:15] instead of '*.{php,cdb}' [00:15:19] which isn't supported [00:15:24] really? yuck [00:15:44] yeah, the docs say so and i just did a quick experiment to verify [00:15:56] what shell is that? [00:17:48] ori: would *.??? work? [00:18:01] or even *.* [00:18:32] *.* should work and not pick up the dir [00:19:04] yeah both would work [00:19:41] and no shell, it's fnmatch behind the scenes [00:19:42] https://hg.python.org/cpython/file/2.7/Lib/fnmatch.py#l85 [00:20:24] sudo_check_call is subprocess.Popen with shell=Ture [00:20:29] True even [00:20:48] yeah but i'm referring to the if glob.glob('%s/*.cdb' % out_dir): [00:21:02] which I took out and which I thought you were asking me to re-introduce [00:21:49] *.* would match . and .. I think [00:21:51] hmm.. yeah if there is nothing there cp will return 1 as well won't it [00:22:42] let me test that [00:23:15] *.* does not match . and .. [00:23:32] "If the directory contains files starting with . they won’t be matched by default." https://docs.python.org/2/library/glob.html [00:23:46] yuck. even *.{php,cdb} fails unle there is at least one of each file type [00:24:03] *.[pc][hd][pb] :) [00:25:37] we still need to exclude the $out_dir/upstream dir or the cp will fail too so *.* should work there [00:26:03] so ugly glob and *.* on the cp will work I guess [00:26:25] why not *.[pc][hd][pb] for both? with a comment explaining that this is temporary and apologizing to posterity [00:26:30] (that's what i currently have staged) [00:26:47] let me test *.[pc][hd][pb] in the shell [00:27:32] yeah that seems to work ok [00:28:30] amended [00:42:20] ori: I just updated beta to include your scap patch [00:43:55] bd808: sweet, thanks. I'll test it. [15:48:18] ori: we should setup beta cluster to build and use LCStoreStaticArray if it looks promising. The only thing that doesn't get exercise there is the new branch initial build [15:48:48] bd808: I may have, um, migrated production to LCStoreStaticArray [15:48:59] oh I know [15:49:01] :) [15:49:19] but now you have a chance to do the right thing in beta cluster too [15:49:21] the new branch initial build is the thing I'm most worried about [15:49:36] well twentyafterfour will be doing that today [15:49:42] you should warn him [15:49:54] yeah [15:50:54] I could tell that you were racing to try it in prod. If I was a better person I would have stopped you and made you do it in beta first [15:52:07] it would have not gotten done that way [15:52:37] seriously? and extra 6 hours would doom the project? [15:53:19] I like to see shit get done but I also like to be able to test and debug failures [15:56:07] ok, I'll do beta next