Closed
Bug 1094805
Opened 10 years ago
Closed 9 years ago
running talos on a release build locally doesn't give us sourcestamp/repository
Categories
(Testing :: Talos, defect)
Testing
Talos
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: jmaher, Assigned: vidya.vnv, Mentored)
References
Details
(Whiteboard: [good first bug][lang=python])
Attachments
(2 files, 1 obsolete file)
(deleted),
patch
|
wlach
:
review-
|
Details | Diff | Splinter Review |
(deleted),
patch
|
jmaher
:
review+
|
Details | Diff | Splinter Review |
I ran into errors with the new mozversion while testing locally.
Reporter | ||
Comment 1•10 years ago
|
||
a pretty simple bug
Comment 2•10 years ago
|
||
Comment on attachment 8518146 [details] [diff] [review] try/except around missing data (1.0) Review of attachment 8518146 [details] [diff] [review]: ----------------------------------------------------------------- This patch is really dangerous as it could allow us to submit invalid results to the graphserver (remember the NULL issue on MacOS X?). I can't reproduce this problem using the latest released firefox and this command line: talos -e ~/tmp/firefox/firefox --tppagecycles 1 --tpcycles 1 --cycles 1 -a tart --develop It seems like every released version of Firefox includes the necessary information so I'm not sure what problem you ran into?
Attachment #8518146 -
Flags: review?(wlachance) → review-
Reporter | ||
Comment 3•10 years ago
|
||
hmm, this was an official distribution inside my ubuntu package, found with `which firefox`. I can see how this could be dangerous, maybe allow this if we are in --develop mode?
Comment 4•10 years ago
|
||
(In reply to Joel Maher (:jmaher) from comment #3) > hmm, this was an official distribution inside my ubuntu package, found with > `which firefox`. I can see how this could be dangerous, maybe allow this if > we are in --develop mode? Ah, that would make sense. Instead of hacking up Talos, I think it might be better to special case binary handling in mozversion to look for the .ini files in special locations if the path matches a default installation path. On Ubuntu the .ini files we need are in /usr/lib/firefox/. I'll bet something similar is the case on Win32 and MacOS X.
Reporter | ||
Comment 5•10 years ago
|
||
to work on this bug you will need to be setup to develop in mozilla-central: https://developer.mozilla.org/en-US/docs/mozilla-central the mozversion code lives here in the tree: http://dxr.mozilla.org/mozilla-central/source/testing/mozbase/mozversion a first step of this bug should be to support /usr/lib/firefox for application.ini if we cannot find it by default.
Mentor: jmaher, wlachance
Component: Talos → Mozbase
QA Contact: hskupin
Whiteboard: [good first bug][lang=python]
Comment 6•10 years ago
|
||
(In reply to Joel Maher (:jmaher) from comment #5) > a first step of this bug should be to support /usr/lib/firefox for > application.ini if we cannot find it by default. I think we should only do this if the binary path is specified as /usr/bin/firefox. If specified as something like /home/wlach/tmp/firefox/firefox, we should probably error out as now (rather than using a system copy of firefox's ini version information which does not correspond).
Comment 7•10 years ago
|
||
Joel, what type of build is /usr/bin/firefox? Are you on Ubuntu? In such a case it will be a symlink, pointing to /usr/lib/firefox/firefox. Please check the application.ini file in that folder. I assume it does not contain the respository and sourcestamp information. So in the example code as attached, you should use version_info.get('application_repository') instead. This is a recent change, which was part of my patch on bug 1088060. In the cases above you cannot expect that the value exists.
Flags: needinfo?(jmaher)
Reporter | ||
Comment 8•10 years ago
|
||
jmaher@jmaher-ThinkPad-X230:~$ echo `which firefox` /usr/bin/firefox jmaher@jmaher-ThinkPad-X230:~$ ls -la /usr/bin/firefox lrwxrwxrwx 1 root root 25 Oct 13 17:03 /usr/bin/firefox -> ../lib/firefox/firefox.sh jmaher@jmaher-ThinkPad-X230:~$ cat /usr/lib/firefox/application.ini ; This file is not used. If you modify it and want the application to use ; your modifications, move it under the browser/ subdirectory and start with ; the "-app /path/to/browser/application.ini" argument. [App] Vendor=Mozilla Name=Firefox Version=33.0 BuildID=20141013200257 ID={ec8030f7-c20a-464f-9b0e-13a3a9e97384} [Gecko] MinVersion=33.0 MaxVersion=33.0 [XRE] EnableProfileMigrator=1 EnableExtensionManager=1 [Crash Reporter] Enabled=1 ServerURL=https://crash-reports.mozilla.com/submit?id={ec8030f7-c20a-464f-9b0e-13a3a9e97384}&version=33.0&buildid=20141013200257 jmaher@jmaher-ThinkPad-X230:~$
Flags: needinfo?(jmaher)
Comment 9•10 years ago
|
||
As you see there is not such information available. So change the code in talos if you want to use it with an official Firefox package on Ubuntu. This is not a bug in mozversion.
Component: Mozbase → Talos
QA Contact: hskupin
Comment 10•10 years ago
|
||
I wonder if we could solve this problem just by having mozversion follow symbolic links of the binary.
Reporter | ||
Comment 11•10 years ago
|
||
:wlach, back to --develop mode to not require this?
Flags: needinfo?(wlachance)
Comment 12•10 years ago
|
||
(In reply to William Lachance (:wlach) from comment #10) > I wonder if we could solve this problem just by having mozversion follow > symbolic links of the binary. mozversion is already doing this. I implemented this not that long ago in bug 1061809.
Comment 13•10 years ago
|
||
Oh, and if you don't want to run such kind of builds with Talos you might wanna do it as you did in the patch. In our case for Mozmill tests we simply assume the latest release. Not the best, but so far it works.
Reporter | ||
Comment 14•10 years ago
|
||
here is how to get setup with talos: https://wiki.mozilla.org/Buildbot/Talos/Running#Running_locally_-_Source_Code use the patch on this bug and if we get an exception and we are not in --develop mode, then we need to quit the test run with an error.
Reporter | ||
Comment 15•10 years ago
|
||
wlach, any further thoughts on a final direction here? I keep running into this locally. I don't mind hacking this out for --develop.
Assignee: jmaher → nobody
Status: ASSIGNED → NEW
Flags: needinfo?(wlachance)
Comment 16•10 years ago
|
||
If it's not too hacky, yeah, I think we should ignore missing revision information in --develop mode. I'm not sure if this is really a good first bug; I think I'd rather someone have a bit of experience with talos before tackling it. I might make it a "good next bug".
Flags: needinfo?(wlachance)
Assignee | ||
Comment 17•9 years ago
|
||
Attachment #8562192 -
Flags: review?(jmaher)
Reporter | ||
Comment 18•9 years ago
|
||
Comment on attachment 8562192 [details] [diff] [review] Patch to check for sourcestamp/repository on a release build locally. Review of attachment 8562192 [details] [diff] [review]: ----------------------------------------------------------------- just one nit and an idea for making this a little better! ::: talos/run_tests.py @@ +189,5 @@ > + browser_config['repository'] = version_info['application_repository'] > + browser_config['sourcestamp'] = version_info['application_changeset'] > + except: > + if not browser_config['develop']: > + print "unable to find changeset or repository: %s" % version_info nit: please remove the blank space at the end of the line. @@ +190,5 @@ > + browser_config['sourcestamp'] = version_info['application_changeset'] > + except: > + if not browser_config['develop']: > + print "unable to find changeset or repository: %s" % version_info > + sys.exit() I would like to hack these values as we might need them, how about adding: else: browser_config['repository'] = 'develop' browser_config['sourcestamp'] = 'develop'
Attachment #8562192 -
Flags: review?(jmaher) → review-
Assignee | ||
Comment 19•9 years ago
|
||
Attachment #8562192 -
Attachment is obsolete: true
Attachment #8562220 -
Flags: review?
Attachment #8562220 -
Flags: review? → review?(jmaher)
Reporter | ||
Comment 20•9 years ago
|
||
Comment on attachment 8562220 [details] [diff] [review] Added else statements and removed blankspace Review of attachment 8562220 [details] [diff] [review]: ----------------------------------------------------------------- thanks!
Attachment #8562220 -
Flags: review?(jmaher) → review+
Reporter | ||
Comment 21•9 years ago
|
||
https://hg.mozilla.org/build/talos/rev/9e89d4094c48
Reporter | ||
Updated•9 years ago
|
Assignee: nobody → vidya.vnv
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
You need to log in
before you can comment on or make changes to this bug.
Description
•