Summary: | garden-o-matic should know how to fetch test results from the (NRWT) bots | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Adam Barth <abarth> | ||||||||
Component: | New Bugs | Assignee: | Adam Barth <abarth> | ||||||||
Status: | RESOLVED FIXED | ||||||||||
Severity: | Normal | CC: | dglazkov, eric, ojan, webkit.review.bot | ||||||||
Priority: | P2 | ||||||||||
Version: | 528+ (Nightly build) | ||||||||||
Hardware: | Unspecified | ||||||||||
OS: | Unspecified | ||||||||||
Attachments: |
|
Description
Adam Barth
2011-07-05 16:13:45 PDT
Created attachment 99758 [details]
Patch
@ojan: Can we make the server understand the normal JSONP "callback" parameter? That would make the client side code cleaner (and more robust). (In reply to comment #2) > @ojan: Can we make the server understand the normal JSONP "callback" parameter? That would make the client side code cleaner (and more robust). That's fine with me. The only issue is the local results.html file needs to load full_results.json from the local filesystem, which as you are well aware we can't do XHR to the local filesystem in Chrome. So, at least for that script, we need the callback built in to the written out JSON. If we're going to have it for that one, may as well have it for all of them? That said, we could have the server replace the hard-coded call with the callback, but that gains us less. Open to better suggestions of course. I agree this isn't ideal. I think it's fine for the server to replace the default callback with one specified in the URL. That seems like the easiest solution. Created attachment 99765 [details]
Patch
Comment on attachment 99765 [details]
Patch
OK.
Comment on attachment 99765 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=99765&action=review > Tools/Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:15 > + function fetchResults(builder_name, success) { nit: I tend to call paramters likes this "onsuccess" or "successHandler" or something that makes it clear it's a function not a boolean. Whoops. Didn't mean to reset the cq+. I'll leave it cq? for now in case you want to address my nit. Feel free to ignore it though. Created attachment 99766 [details]
Patch
> nit: I tend to call paramters likes this "onsuccess" or "successHandler" or something that makes it clear it's a function not a boolean.
Yeah, I was emulating the name jquey uses in its $.get function, but onsuccess does seem much clearer. :)
Comment on attachment 99766 [details] Patch Clearing flags on attachment: 99766 Committed r90424: <http://trac.webkit.org/changeset/90424> All reviewed patches have been landed. Closing bug. |