[00:24:22] yes, although for MATCH_STRING_IC I think you should use strcasecmp() [00:25:52] for regexes it would be better to convert them to PCRE at compile time rather than using mb_ereg() [00:27:44] I don't know what is going on with input_substr() [00:29:16] for a single character you can use ($s[$i] ?? '') [00:50:00] I don't know whether the unicode stuff in general is needed, we try to do 8-bit clean parsing in PHP because it is faster [00:51:32] for example if you're doing a case-insensitive match against a literal from the PEG, ASCII case insensitivity is fine as long as the literal is ASCII, which it probably is for us [01:00:52] peg_char_class_test() is also a case in point, it should really be inlined, with the ord_unicode() replaced with ord() if the character class allows it [01:07:57] TimStarling, oh .. you've been typing here. i didn't pay attention. [01:08:22] but, if you want to, you are welcome to do the port ;-) I haven't started yet. [01:09:01] otherwise, can you drop your notes onto the phab task once you are done reviewing? [01:09:08] i am about to sign out in a few mins. [01:10:46] and thanks for the reviewing and input. [01:10:47] ttyl. [02:05:18] I'm adding a comment [06:17:41] <_joe_> TimStarling: should I repackage/install excimer in deployment-prep and production now? [09:49:25] _joe_: it should be pretty harmless to install, but it probably still has bugs in it [09:49:43] so as long as you don't mind packaging it again for bugfixes later on, sure [09:50:57] <_joe_> TimStarling: sure, I don't mind at all [17:32:09] anomie: hello! Sorry I had some kid this afternoon, could not follow up on the memcached error :/ [17:32:21] I am gonna deploy your patch https://gerrit.wikimedia.org/r/#/c/mediawiki/core/+/473551/1/includes/libs/objectcache/MemcachedPeclBagOStuff.php :) [17:32:29] hashar: ok, thanks [17:34:12] anomie: tldr: my oldest kid hurted her a foot with some wood piece and i had to rush to the doctor [17:34:17] nothing too serious thanksfully [17:34:31] I'm glad it wasn't anything too serious [18:38:18] anomie: sorry it took a while, I ran a full sync to deploy the fix :( [18:38:25] or how to waste an hour or so bah [18:38:47] hashar: Oh well. I'm pretty sure I have identified exactly what the problem is, checking one thing yet. [18:40:26] \o/ [18:41:39] hashar: https://gerrit.wikimedia.org/r/#/c/mediawiki/core/+/473578 [18:42:32] anomie: I was looking at that same method [18:42:38] and missed the lack of return !!!!!! [18:43:22] I did too at first. Then when I saw it, "oh, duh." [18:43:40] I am always impressed at how fast you catch those nasty bugs [18:43:56] one sure thing, we need more tests for Job* classes :/ [18:44:47] anomie: then about your comment on https://phabricator.wikimedia.org/T209429#4746781 , my kid injured herself right when I was about to create a patch to add the exception in the log context :( [18:45:26] at least, it was not too bad of an error. I will resume the train once your fix is deployed and confirmed [18:47:17] hashar: I felt bad about that comment once you said about your kid earlier. :/ [18:48:11] worse: I had a similar patch but wanted to add a test to exercise MemcachedPeclBagOStuff::checkResult() yielding a log error in that case [18:48:25] and got lost in writing the test ;-D [18:48:42] I am afraid I am too rusty in PHP/phpunit/mediawiki development [19:22:03] one more train blocker aced by anomie \o/ [19:22:48] (: [21:29:02] Upstream have finally removed the old ui from gerrit! (no longer exists)! [21:34:07] https://gerrit-review.googlesource.com/c/gerrit/+/116790