00:18:05  * unixpicklequit (Quit: My Mac has gone to sleep. ZZZzzz…)
00:39:12  * jgiquit (Quit: jgi)
01:06:43  * plutoniixjoined
01:26:15  * unixpicklejoined
02:04:35  * Garbeequit (Quit: Connection closed for inactivity)
02:06:45  * unixpicklequit (Quit: My Mac has gone to sleep. ZZZzzz…)
04:23:04  * NewNewbiequit
04:23:16  * NewNewbiejoined
06:51:39  * thefourtheyejoined
07:49:25  <jochen__>you can always just code-search for it
07:49:42  <jochen__>cs.chromium.org/parser-xml-close-comment
11:08:08  * plutoniixquit (Quit: จรลี จรลา)
11:27:35  * Garbeejoined
13:09:49  * bobmcwjoined
13:10:20  * unixpicklejoined
13:49:42  * plutoniixjoined
14:27:35  * unixpicklequit (Quit: My Mac has gone to sleep. ZZZzzz…)
14:29:11  * hferreirojoined
14:37:07  * jugglinmikejoined
14:56:23  * unixpicklejoined
15:03:13  <caitp>jwolfe: that's a WebKit test under test/webkit
15:03:47  <caitp>v8 inherited the old incorrect HTML comment behaviour from WebKit, the test is just wrong
15:04:15  <caitp>feel free to correct it
15:23:57  * ookqwefpoqkwfpjoined
15:24:02  * ookqwefpoqkwfpquit (Client Quit)
15:24:17  * qwefopqwefokqwopjoined
15:24:27  * qwefopqwefokqwopquit (Client Quit)
16:13:25  <jwolfe>caitp, oh. it's a v8 test, not a webkit test. it's not in chromium/src; it's in v8. that's what confused me. and run-tests.py doesn't run it by default.
16:13:42  <jwolfe>jochen__, that's a cool resource. thanks for sharing.
16:13:58  <caitp>yes, you have to run the webkit suite explicitly
16:14:43  <caitp>it is in fact a webkit test, in that the tests were copied from WebKit (which actually still has this incorrect behaviour)
16:15:05  <jwolfe>apparently `run-tests.py webkit` will do it, even though it's not listed under TESTS: in `run-tests.py --help`.
16:15:14  <jwolfe>caitp, oh you mean the real WebKit? not Blink?
16:15:18  <caitp>yeah
16:15:23  <jwolfe>oh ok
16:15:36  <jwolfe>so how do i run ALL the tests?
16:15:58  <caitp>I don't think there's an easy way to do it :x
16:16:26  <caitp>though you should be able to say `run-tests.py mjsunit webkit cctest unittests test262` to get most of them
16:16:33  <caitp>not sure if that actually works though
16:20:19  <jwolfe>a different question is how do i find out what all checks the precommit bots do?
16:21:33  * RT|Chatzillaquit (Quit: ChatZilla 0.9.86.1 [Firefox 2.0.0.22pre/2010030309])
16:25:41  * seventhjoined
16:31:14  * hferreiroquit (Ping timeout: 272 seconds)
16:34:35  <caitp>jwolfe: if we can get you trybot access, you can run all of those checks yourself without committing
16:34:39  <caitp>cq dryrun also does that
16:36:26  <jwolfe>last time i tried to do a dryrun i wasn't allowed to. i'm not sure how to use that feature effectively
16:37:06  * hferreirojoined
16:37:40  <caitp>running them locally is trickier, jwolfe: actually, it looks like you can run `run-tests.py <arch config> bot_default`, which probably gets you pretty close
16:37:45  <caitp>that runs a much bigger set of tests
16:38:27  <jwolfe>oh cool. ... why wasn't that documented in --help?
16:38:30  <caitp>does not include test262 though
16:38:47  <caitp>dunno
16:38:51  <jwolfe>a bunch of test262 things are failing for me currently. not sure why that is...
16:39:02  <caitp>it's probably prone to changing, so hard to document
16:39:26  <caitp>have you run `gclient sync` recently?
16:40:13  <caitp>if they're date/time-related failures, they may not necessarily mean much
16:40:15  <caitp>I believe they pass on the bots
16:42:34  <caitp>`run-tests.py --exhaustive-variants bot_default test262` looks like it should give you the most exhaustive local test coverage
16:43:51  <jwolfe>is `gclient sync` a thing in the v8 source?
16:43:59  <jwolfe>i don't remember reading that in the wiki
16:44:14  <caitp>gclient is part of depot_tools
16:44:41  <caitp>it will update dependencies in the DEPS file, and run some hooks
16:44:49  <caitp>maybe some other stuff
16:45:11  <jwolfe>oh. i think i've never done that for v8.
16:45:26  <jwolfe>how do i run `gclient sync`? it's giving me an error that it's not configured. i don't remember having that problem in chromium/src.
16:45:51  <caitp>hmm, how did you checkout v8?
16:45:58  <caitp>you had issues using fetch v8 right?
16:46:00  <jwolfe>depot_tools fetch v8
16:46:03  <caitp>ah
16:46:12  <caitp>huh, weird
16:46:24  <caitp>well, I __think__ `git cl config` with all default options should set you up right
16:46:35  <caitp>not positive though, littledan__ might know more
16:46:39  <caitp>or jochen__
16:46:59  <jwolfe>oh, i remember that depot_tools fetch barfed some nonsense files in my current directory that screwed up my ability to do another `depot_tools fetch` or possibly a `gclient sync` in another directory. i remember deleting gclient configuration.
16:47:14  <caitp>right, yeah
16:47:18  <caitp>that would be an issue
16:47:29  <caitp>I can send you a copy of the v8 solution file
16:47:30  <jwolfe>how am i supposed to have both a chromium/src and a v8 checkout on the same machine?
16:47:43  <jwolfe>solution file? what's that?
16:47:44  <caitp>I have them in different directories
16:47:53  <caitp>~/v8 and ~/chromium
16:48:14  <jwolfe>i have ~/v8 and ~/chromium too. specifically ~/v8/.git and ~/chromium/src/.git
16:49:15  <jwolfe>doing `depot_tools fetch v8` in ~ created ~/.gclient or something, which caused errors for `(cd ~/chromium && depot_tools fetch chromium)`. i might be getting that backwards; it's been a while since i've done this.
16:49:48  <jwolfe>maybe if you start with chromium, then fetching v8 works fine?
16:49:57  <caitp>unfortunately no
16:51:24  <caitp>basically, I have ~/chromium/.gclient (for chromium) and ~/.gclient (for v8)
16:51:36  <caitp>if I wanted to contribute to skia or something else, it would be more complicated
16:52:20  <jwolfe>so really i should have ~/v8container/v8/.git so that ~/v8container/.gclient and ~/chromium/.gclient don't interfere with each other.
16:52:34  <caitp>yes, that would be a way to do it
16:53:35  <caitp>~/chromium.gclient should never be an issue for your v8 checkout though, since ~/chromium is just a wrapper for ~/chromium/src
16:54:41  <caitp>note you could also work on the v8 checkout directly inside the chromium repo, just by updating it to ToT and building it from chromium/src/v9
16:55:21  <caitp>although it would probably still give you problems with gclient sync
16:57:31  <caitp>basically gclient could use some love, but I guess it works well enough for chromium that it isn't really getting it
17:10:11  <jwolfe>i'm too scared to work in ~/chromium/src/v8. i'm afraid that gclient would delete my branches.
17:10:39  <jwolfe>anyway, i got it working by interrupting a fresh `fetch v8` from a tmp dir, and grabbing the .gclient from there.
17:16:49  <caitp>sounds good
17:40:51  * thefourtheyequit (Quit: Connection closed for inactivity)
18:14:37  * scottmgquit (Ping timeout: 250 seconds)
18:15:57  * scottmgjoined
18:21:46  <littledan__>jwolfe, what's the interference you're going through? I didn't have this problem
18:22:07  <littledan__>I tend to git cl upload all the work that I care about, even before it's out for review, and then I don't have to be scared about losing work through messing up the repository state
18:22:41  <jwolfe>littledan__, you mean why am i worried to work in chromium/src/v8?
18:24:57  * unixpicklequit (Quit: My Mac has gone to sleep. ZZZzzz…)
18:26:18  * unixpicklejoined
18:26:20  * unixpicklequit (Client Quit)
18:30:15  * jgijoined
18:36:00  * hferreiroquit (Ping timeout: 272 seconds)
18:40:05  <jwolfe>littledan__, ok, i did more experimenting. the problem is that if you try to do `fetch chromium` in a subdirectory while `fetch v8` is running in the parent directory, then you get an error.
18:40:24  <jwolfe>i suppose the disclaimer about not dealing with partial checkouts was supposed to be my clue for that issue.
18:44:30  <littledan__>don't do it in a child directory; do it in a separate unrelated one
18:48:11  * jgiquit (Quit: jgi)
18:50:20  * jgijoined
18:50:44  * jgiquit (Client Quit)
18:58:56  <jwolfe>littledan__, so you mean like ~/v8/v8/.git and ~/chromium/src/.git?
18:59:45  <jwolfe>if you do `fetch v8` in ~, then doing `fetch chromium` in ~/chromium has a child-parent relationship with the other `fetch`.
19:01:52  <jwolfe>the chromium instructions have a "mkdir" step: https://www.chromium.org/developers/how-tos/get-the-code , but the v8 instructions do not: https://github.com/v8/v8/wiki/Using%20Git#how-to-start . should i have done a `mkdir` for v8 too?
20:12:04  <littledan__>jwolfe, I made separate directories like that, yes
20:12:40  <littledan__>I guess the V8 instructions don't refer to how to develop Chromium. Why don't you propose an edit to the wiki page?
20:52:23  <jwolfe>littledan__, what's the process for proposing an edit to the wiki?
20:52:31  <littledan__>isn't there an edit link?
20:53:06  * unixpicklejoined
20:53:55  <jwolfe>um. i don't think so. Ctrl+F for "edit" doesn't find it.
20:55:16  * unixpicklequit (Read error: Connection reset by peer)
21:00:20  <jwolfe>and as a minor question, what do you think the container directory should be? should it be ~/v8, to result in ~/v8/v8/.git?
21:58:09  * seventhquit (Ping timeout: 246 seconds)
22:11:45  * bobmcwquit (Remote host closed the connection)
22:20:13  * plutoniixquit (Quit: จรลี จรลา)
22:30:46  * RT|Chatzillajoined