Closed
Bug 456466
Opened 16 years ago
Closed 16 years ago
compiling SeaMonkey Linux debug symlink build fails in Jarmaker/Gnomestripe
Categories
(Firefox Build System :: General, defect)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: mnyromyr, Assigned: Pike)
References
Details
Attachments
(1 file, 1 obsolete file)
391 bytes,
patch
|
ted
:
review+
mnyromyr
:
review+
|
Details | Diff | Splinter Review |
Compiling SeaMonkey trunk debug (with --enable-chrome-format=symlink) fails as of today on Linux when JarMaker.py tries to package gnomestripe: processing /home/kd/projekte/mozilla/mozilla.org/src/trunk/mozilla/toolkit/themes/gnomestripe/global/jar.mn Traceback (most recent call last): File "/home/kd/projekte/mozilla/mozilla.org/src/trunk/mozilla/config/JarMaker.py", line 445, in <module> main() File "/home/kd/projekte/mozilla/mozilla.org/src/trunk/mozilla/config/JarMaker.py", line 442, in main jardir=options.j) File "/home/kd/projekte/mozilla/mozilla.org/src/trunk/mozilla/config/JarMaker.py", line 228, in makeJar localedirs) File "/home/kd/projekte/mozilla/mozilla.org/src/trunk/mozilla/config/JarMaker.py", line 299, in processJarSection outHelper, jf) File "/home/kd/projekte/mozilla/mozilla.org/src/trunk/mozilla/config/JarMaker.py", line 348, in _processEntryLine outHelper.symlink(realsrc, out) File "/home/kd/projekte/mozilla/mozilla.org/src/trunk/mozilla/config/JarMaker.py", line 397, in symlink os.symlink(src, out) OSError: [Errno 17] File exists make[8]: *** [libs] Error 1 make[8]: Leaving directory `/home/kd/projekte/mozilla/mozilla.org/obj/sr/mozilla/toolkit/themes/gnomestripe/global' make[7]: *** [libs] Error 2 make[7]: Leaving directory `/home/kd/projekte/mozilla/mozilla.org/obj/sr/mozilla/toolkit/themes/gnomestripe' make[6]: *** [libs] Error 2 make[6]: Leaving directory `/home/kd/projekte/mozilla/mozilla.org/obj/sr/mozilla/toolkit/themes' make[5]: *** [libs] Error 2 make[5]: Leaving directory `/home/kd/projekte/mozilla/mozilla.org/obj/sr/mozilla/toolkit' make[4]: *** [libs_tier_toolkit] Error 2 make[4]: Leaving directory `/home/kd/projekte/mozilla/mozilla.org/obj/sr/mozilla' make[3]: *** [tier_toolkit] Error 2 make[3]: Leaving directory `/home/kd/projekte/mozilla/mozilla.org/obj/sr/mozilla' make[2]: *** [default] Error 2 make[2]: Leaving directory `/home/kd/projekte/mozilla/mozilla.org/obj/sr/mozilla' make[1]: *** [default] Error 2 make[1]: Leaving directory `/home/kd/projekte/mozilla/mozilla.org/obj/sr' make: *** [build] Error 2
Assignee | ||
Updated•16 years ago
|
Severity: blocker → normal
Comment 1•16 years ago
|
||
Dupe of bug 456483?
Reporter | ||
Comment 2•16 years ago
|
||
(In reply to comment #1) > Dupe of bug 456483? Most likely. (Or the other way round, but Sylvain's has a patch.) But I wonder since when debug build breakage is considered "normal"?
Assignee | ||
Comment 4•16 years ago
|
||
This is normal because symlink output isn't stop-world. We were *that* close to unsupporting it completely, even. To copy my review statement from bug 456483, that version required python 2.4, and that's ungood, I just had to fix something else for 2.3. The perl version unconditionally removes the target, let's just do that, too. It's probably cheaper than checking for all the possible error conditions, too.
Attachment #339940 -
Flags: review?(ted.mielczarek)
Attachment #339940 -
Flags: review?(mnyromyr)
Assignee | ||
Updated•16 years ago
|
Assignee: nobody → l10n
Status: NEW → ASSIGNED
Comment 5•16 years ago
|
||
I tried using os.path.exists in a previous version of the patch in bug 456483 but it didn't work. For some reasons I had broken symlinks in my objdir (the buildinfo.html file or something similar), so os.path.exists returned False, the target was not removed and the symlink call raised an exception. Here's how bazaar implement lexists for Python < 2.4: http://www.google.com/codesearch?hl=en&q=lexists+show:F74BlVeB2yE:_0ashqPj9x0:dQkYe0kRM40&sa=N&cd=2&ct=rc&cs_p=http://bazaar-vcs.org/releases/src/bzr-0.14.tar.gz&cs_f=bzr-0.14/bzrlib/osutils.py#l151
Assignee | ||
Comment 6•16 years ago
|
||
... which is why I'm not using os.path.exists in the first place.
Comment 7•16 years ago
|
||
Except your patch is using os.path.exists? + # remove previous link or file + if os.path.exists(out): + os.remove(out)
Assignee | ||
Updated•16 years ago
|
Attachment #339940 -
Attachment is obsolete: true
Attachment #339940 -
Flags: review?(ted.mielczarek)
Attachment #339940 -
Flags: review?(mnyromyr)
Assignee | ||
Comment 8•16 years ago
|
||
Comment on attachment 339940 [details] [diff] [review] remove target for symlinks Old patch, sorry. Duh.
Assignee | ||
Comment 9•16 years ago
|
||
This time for real, try to remove and except fine if that fails.
Attachment #339963 -
Flags: review?(ted.mielczarek)
Attachment #339963 -
Flags: review?(mnyromyr)
Updated•16 years ago
|
Attachment #339963 -
Flags: review?(ted.mielczarek) → review+
Reporter | ||
Comment 10•16 years ago
|
||
Comment on attachment 339963 [details] [diff] [review] remove target for symlinks This works for me with Python 2.5, but I couldn't test with 2.4.
Attachment #339963 -
Flags: review?(mnyromyr) → review+
Assignee | ||
Comment 11•16 years ago
|
||
http://hg.mozilla.org/mozilla-central/rev/4d57bfc569f5, FIXED.
Status: ASSIGNED → RESOLVED
Closed: 16 years ago
Resolution: --- → FIXED
Comment 12•16 years ago
|
||
(In reply to comment #4 by l10n@mozilla.com) > This is normal because symlink output isn't stop-world. We were *that* close to > unsupporting it completely, even. Which is ironic as symlink builds speed up l10n development the most.
Updated•6 years ago
|
Product: Core → Firefox Build System
You need to log in
before you can comment on or make changes to this bug.
Description
•