mozilla :: #mozreview

14 Mar 2017
04:33jeremychenzalun: ping
04:33zalunjeremychen: pong
04:33jeremychenzalun: hi, I wonder if reviewboard could export the commit rev for the patch?
04:34zalunwhat do you mean?
04:35jeremychenzalun: I mean, for now, we could click the "view raw" button below the patch on MozReview to see the HG changeset.
04:36jeremychenzalun: I think, the issue would be how to get the correct rev number
04:37zalunthe mozreview revision number?
04:38jeremychensomething like[rev number] could lead us to the raw diff, right?
04:40jeremychenIf we could get the correct revision number, we could try apply this instead using the "Download Diff" link
04:40zaluntechnically - sure, but I think we do use the HG bits as these are tool independent
04:42jeremychenzalun: Well, I may not that familiar with how MozReview works, so could you enlighten me a bit how can I get the raw diff from HG? So we could let bz apply to use the correct diff
04:42zalun For now - I don't know
04:43zalunI think smacleod might have an idea. I will check that and comment on the bug
04:43zalunhow bz is doing it at the moment? I mean what arguments does it use to get the raw diff?
04:44globzalun: links to that
04:44firebotBug 1346321 NEW, MozReview "virtual" commit-message file should not be returned by the "Download Diff" url.
04:45zalunalso - I see the `view raw` has no commit-message, so it might be it uses HG to generate that
04:45globzalun: fixing our diff is the best way forward here. i suggest adding a hook to our fork that allows the list to be filtered
04:47zalunjeremychen: let's stay on fixing mozreview first
04:48zalunI think change in bz would require an API on our side to provide you the right link to HG (but I don;t know all the hidden stuff in mozreview)
04:50jeremychenzalun: ok, what I was proposing (but maybe not the very right way!?) is to export the HG rev number, which is already shown on MozReview UI (see the textarea with "hg pull -r [rev_umber]"), and let "git bz apply" to create patch from HG.
04:50jeremychenI think this is an alternative for not changing the begavior of "Download Diff", since looks like changing this is very difficult at present?
04:50globwe could have a url that redirects to the right hg url, but that seems pointless if we instead fix 'download diff'
04:51jeremychenglob: right
04:51globjeremychen: no, fixing our 'download diff' isn't difficult. the patch on the bug works, but needs some slight changes to move the logic into our extension
04:52globlet's make that happen as fast as we can, and think about doing it "a much better way" later
04:52jeremychenglob: agree. thanks for the explanation and help. :)
05:07pulsebotCheck-in: - Mike Hommey - Bug 1342491 - Handle any kind of exception from git-cinnabar. r=glob
05:26glandiumglob: ping
05:26globglandium: pong
05:27glandiumglob: hey, so, many people have been complaining about git mozreview push failing, but were stuck because of bug 1342491 hiding the actual error
05:27firebot FIXED, git mozreview errors hidden
05:27* glob nods
05:28glandiumglob: it turns out I just got a git mozreview push error myself, and I have bug 1342491 fixed, obviously
05:28globglandium: a ha
05:28glandiumand all I can say is that I got a urllib2.HTTPError: HTTP Error 500: Internal Server Error
05:29globglandium: was it slow?
05:29glandiumI can say it's not the reviewboard submission that failed
05:29glandiumthat is, it's an http error on mercurial
05:30glandiumthe weird thing is that my changeset was there on the server before I re-mozreview pushed
05:30globglandium: do you have exact timestamps for the error?
05:31glandiumunfortunately no, but I can give you my ip
05:31glandium(if that helps)
09:28zalunI've got bug 1346321 hooked, but not in extension.hooks, the problem was with recursive imports
09:28firebot NEW, MozReview "virtual" commit-message file should not be returned by the "Download Diff" url.
09:32zalunI'm not so sure it's sane, but it does work
14:50mcoteglob: should we take reviewboard-hg2 out for now, if it's 500ing all the time?
14:51globmcote: it isn't 500ing "all the time", just sometimes. and there's no hg1 :(
14:51mcoteoh ha
14:51globit would at least make the errors consistent
14:51mcoteokay I thought it was just that half the servers were acting up
18:49mcotezalun: do you need me (or someone) to land
18:49firebotBug 1345740 NEW, diff view shows file diff as commit message
18:57zalunmcote: yes, please - I mentioned it at the standup
18:58mcoteaho kay
18:59mcotewill do that now
19:03pulsebotCheck-in: - Piotr Zalewa - mozreview: check if commit message FileDiff is present by parsing page.model (bug 1345740) r=davidwalsh,mcote
19:10zalunmcote: thanks
15 Mar 2017
No messages
Last message: 11 days and 16 hours ago