Right now the results for Litmus tests are sent with an user agent string like "litmusReporter.js/0.0.2 ()". That is very useless when trying to vet such an result because you cannot see which version of the browser the user is using. I would like to propose that we use the full UA and add the above string at the end so we get something like: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1pre) Gecko/20090525 Shiretoko/3.5pre (litmusReporter.js/0.0.2) Clint, could we get this somehow into the next version? That would help us a lot.
This would be a good idea, but I think that Litmus can only handle the 10 digit number of the Build ID, and that it cannot handle the full UA string. I think we have to repair litmus first to accept a full UA string. However, I'm intrigued by the idea that QAC is somehow sending up this "litmusreporter" UA string. How is that being captured in litmus? How did you find this?
See single results like https://litmus.mozilla.org/single_result.cgi?id=184372 We have enough room for the whole UA string. We do it already when working with Litmus via the web interface. I think QAC should only add itself to the default user agent string.
Yeah we store the whole UA string in the db though we don't make it particularly obvious in the result display. For some boneheaded reason, litmusReporter.js doesn't bother to send it... Easy fix.
Assignee: nobody → zach
Status: NEW → ASSIGNED
Created attachment 379769 [details] [diff] [review] Send full UA string in litmusReporter.js litmusReporter.js didn't send the UA because it's a port of the Test::Litmus perl module, which has no concept of navigator.userAgent. Of course, all this generation of XML as text without through escaping is particularly evil (it's a 4 year old hack that got a bit carried away), but it generally works and that's a story for another bug.
Attachment #379769 - Flags: review?
Attachment #379769 - Flags: review? → review?(ctalbert)
Comment on attachment 379769 [details] [diff] [review] Send full UA string in litmusReporter.js Hey Zach, thanks for the patch! Do you need me to check this in? r=ctalbert
Attachment #379769 - Flags: review?(ctalbert) → review+
SVN is balking at my attempts to check this in so that would be great Clint.
Oh, please no further checkins to SVN! Everything is located at http://hg.mozilla.org/webtools/litmus/ now. Or do I miss something? Frederic, could you check this change in?
(In reply to comment #7) > Oh, please no further checkins to SVN! Everything is located at > http://hg.mozilla.org/webtools/litmus/ now. Or do I miss something? > > Frederic, could you check this change in? Yes, you are right, Litmus has been managed on hg since bug 481376 and unless I am mistaken, has never been on SVN -- it was on CVS before.
(In reply to comment #8) > (In reply to comment #7) > > Oh, please no further checkins to SVN! Everything is located at > > http://hg.mozilla.org/webtools/litmus/ now. Or do I miss something? > > > > Frederic, could you check this change in? > > Yes, you are right, Litmus has been managed on hg since bug 481376 and unless I > am mistaken, has never been on SVN -- it was on CVS before. But, wait a minute ;) This bug and patch are not for Litmus themselves, are they. I think it applies to the QA Companion add-on, which may indeed be managed in SVN. So go ahead and check this in, Clint.
Yes, you are right. I shouldn't start to work so early! :)
clint-talberts-macbook:qac clint$ svn commit -m "Bug 494801 - patch by Zach Lipton, r=ctalbert Make QAC report full UA string" chrome/content/litmusReporter.js Sending chrome/content/litmusReporter.js Transmitting file data . Committed revision 26387.
Status: ASSIGNED → RESOLVED
Last Resolved: 9 years ago
Resolution: --- → FIXED
Looks fantastic now: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:188.8.131.52) Gecko/20090715 Firefox/3.5.1 (litmusReporter.js/0.0.3) Marking as verified fixed.
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.