[00:00:43] AaronSchulz: I think so -- https://phabricator.wikimedia.org/T147098#2681229 [01:44:04] anomie: Somehow I ended up reading the ApiRsd file...should the 'OAuth' setting be set to true if the OAuth extension is installed? [09:46:27] legoktm: You'd think... Ask brion? [09:47:39] * (to aid in OAuth client apps switching to support in the future). [09:52:00] https://gerrit.wikimedia.org/r/313652 [17:22:57] TimStarling hi, im wondering do we still need to do https://github.com/wikimedia/mediawiki/commit/7e233425829ee83fc568bec6dc8b00b1ceb39b7f ? [17:23:10] Since were using php 5.5+ now, im wondering does the bug still eixst? [17:24:05] paladox: The autoloader... Has been changed numerous times since? [17:24:20] Oh, but it is still referenced here [17:24:27] https://github.com/wikimedia/mediawiki/blob/master/includes/AutoLoader.php [17:25:02] Reedy https://github.com/wikimedia/mediawiki/blob/master/includes/AutoLoader.php#L26 [17:25:26] Im just wondering do we still need to workaround the bug [17:25:28] ? [17:26:23] I'd suggest filing a task, and tagging it for the perf team [17:26:58] Ok [17:30:33] Reedy https://phabricator.wikimedia.org/T147119 [17:37:31] Reedy im wondering if this is the line https://github.com/facebook/hhvm/blob/a1a2d704a287fa5485d4e76b74f8677369907c97/hphp/runtime/base/heap-scan.h#L74 were looking for [17:37:36] in supporting apcu in hhvm [17:37:36] ? [20:44:21] well, like I said there, it was never fully isolated, so it's hard to know if it was fixed [20:44:44] it was probably fixed, APC quality has improved a lot [20:45:33] ok [22:22:31] legoktm or AaronSchulz hi, could you review https://gerrit.wikimedia.org/r/#/c/299498/ please? [22:22:39] It adds support for apcu [23:36:20] AaronSchulz: T147122 is related to the MWExceptionRenderer changes. I left some notes on the bug about the general problem. I think we might have to rethink how MWExceptionRenderer gets wired into MWExceptionHandler. [23:36:20] T147122: My user talk don't working - https://phabricator.wikimedia.org/T147122 [23:36:55] Or we have to find and fix all MWException subclasses across extensions [23:37:38] probably both really, partial revert, find and fix, and then put it back [23:37:44] Looks to be about 88 implementations of MWException [23:38:44] I think the partial revert would be to jsut put MWException::report back like it was [23:41:01] * bd808 squints at the new code