Bug 133814 - JSC build failure when running build-symbol-table-index.py
Summary: JSC build failure when running build-symbol-table-index.py
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Nobody
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2014-06-12 12:38 PDT by Matthew Mirman
Modified: 2014-06-17 15:17 PDT (History)
8 users (show)

See Also:


Attachments
Fixes build failure on build-symbol-table-index.py (4.33 KB, patch)
2014-06-12 12:42 PDT, Matthew Mirman
no flags Details | Formatted Diff | Diff
Fixes build failure on build-symbol-table-index.py (4.33 KB, patch)
2014-06-12 12:49 PDT, Matthew Mirman
ggaren: review+
commit-queue: commit-queue-
Details | Formatted Diff | Diff
Same attachment, fixed ChangeLog diff (4.43 KB, patch)
2014-06-12 13:27 PDT, Matthew Mirman
no flags Details | Formatted Diff | Diff
Archive of layout-test-results from webkit-ews-15 for mac-mountainlion-wk2 (519.50 KB, application/zip)
2014-06-12 15:37 PDT, Build Bot
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Matthew Mirman 2014-06-12 12:38:13 PDT
Patch forthcoming.
Comment 1 Matthew Mirman 2014-06-12 12:42:57 PDT
Created attachment 232974 [details]
Fixes build failure on build-symbol-table-index.py

Also adds license to scripts.
Comment 2 Matthew Mirman 2014-06-12 12:49:15 PDT
Created attachment 232976 [details]
Fixes build failure on build-symbol-table-index.py

Also adds license.  Hopefully fixes changelog merge error.
Comment 3 Mark Rowe (bdash) 2014-06-12 13:08:59 PDT
Can you clarify the sequence of events that results in getting in to the state where there are no object files to match? Is that a state that it should be possible to get in to, or is it indicative of an earlier problem too?
Comment 4 Geoffrey Garen 2014-06-12 13:10:57 PDT
Comment on attachment 232976 [details]
Fixes build failure on build-symbol-table-index.py

r=me
Comment 5 WebKit Commit Bot 2014-06-12 13:12:26 PDT
Comment on attachment 232976 [details]
Fixes build failure on build-symbol-table-index.py

Rejecting attachment 232976 [details] from commit-queue.

Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.appspot.com', '--bot-id=webkit-cq-02', 'validate-changelog', '--check-oops', '--non-interactive', 232976, '--port=mac']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit

Last 500 characters of output:
kitpy/common/checkout/changelog.py", line 348, in parse_entries_from_file
    yield ChangeLogEntry(''.join(entry_lines[:-1]), revision=most_probable_revision)
  File "/Volumes/Data/EWS/WebKit/Tools/Scripts/webkitpy/common/checkout/changelog.py", line 112, in __init__
    self._parse_entry()
  File "/Volumes/Data/EWS/WebKit/Tools/Scripts/webkitpy/common/checkout/changelog.py", line 207, in _parse_entry
    self._date_line = match.group()
AttributeError: 'NoneType' object has no attribute 'group'

Full output: http://webkit-queues.appspot.com/results/6608226195865600
Comment 6 Matthew Mirman 2014-06-12 13:27:55 PDT
Created attachment 232978 [details]
Same attachment, fixed ChangeLog diff
Comment 7 Build Bot 2014-06-12 15:37:20 PDT
Comment on attachment 232978 [details]
Same attachment, fixed ChangeLog diff

Attachment 232978 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/4999710074994688

New failing tests:
media/W3C/video/networkState/networkState_during_loadstart.html
Comment 8 Build Bot 2014-06-12 15:37:23 PDT
Created attachment 233001 [details]
Archive of layout-test-results from webkit-ews-15 for mac-mountainlion-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: webkit-ews-15  Port: mac-mountainlion-wk2  Platform: Mac OS X 10.8.5
Comment 9 Matthew Mirman 2014-06-12 16:25:22 PDT
(In reply to comment #3)
> Can you clarify the sequence of events that results in getting in to the state where there are no object files to match? Is that a state that it should be possible to get in to, or is it indicative of an earlier problem too?

It might be a problem that .o files aren't being generated, but ultimately if the script fails to produce a set of llvm files all that will happen is that the inliner won't run.    I can't see what exactly is causing this, but one possibility is that the Objects directory exists, but was created earlier and has been emptied and the .o files are being dumped into some other directory.  I'd need a list of environment variables during the failure to see if this was the case.  Another possibility is that the build system is for whatever reason deleting these files before it gets to the script phase, but this seems unlikely.
Comment 10 Matthew Mirman 2014-06-13 15:17:57 PDT
<rdar://problem/17271872>
Comment 11 Matthew Mirman 2014-06-17 15:17:50 PDT
Comment on attachment 232978 [details]
Same attachment, fixed ChangeLog diff

Clearing flags on attachment: 232978

Committed r170076: <http://trac.webkit.org/changeset/170076>
Comment 12 Matthew Mirman 2014-06-17 15:17:56 PDT
All reviewed patches have been landed.  Closing bug.