Magic-Mirror/gecko-dev
1
0
Fork 0
mirror of https://github.com/mozilla/gecko-dev.git synced 2024-10-31 22:25:30 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
322,091 Commits 615 Branches 98 Tags 6 GiB
dc4eed8ded
Commit Graph

67 Commits

Author SHA1 Message Date
Garrett Robinson
e499a38300 Bug 875126 - Make cgdb an option for mach --debugger. r=ted 2013-06-12 16:17:11 -04:00
Nathan Froyd
54f19801f4 Bug 875332 - add cgdb to the known debuggers list; r=jmaher 2013-05-23 10:56:18 -04:00
Nathan Froyd
0e21827b57 Bug 873493 - add lldb to the list of known debuggers; r=jmaher 2013-05-17 11:22:00 -04:00
Joel Maher
acbe7e05ea Bug 827446 - update mochitest, reftest, xpcshell to use mozcrash. r=jhammel 2013-03-21 09:19:34 -04:00
Ed Morley
f6617f5189 Bug 850681 - Part 11: Move processString to the message part of log lines & remove unhelpful process IDs; r=ted 
Our standard logging format is:
TEST-{INFO,UNEXPECTED-FAIL} | testName | message

For leak logs from processes other than the main process (eg plugin processes),
currently processSingleLeakFile outputs:
TEST-{INFO,UNEXPECTED-FAIL} | foo process 123 | testName | message

This patch turns these into:
TEST-{INFO,UNEXPECTED-FAIL} | testName | foo process: message

Note: The process ID is contained in the raw leak log (which gets printed to the
log anyway), so we don't need to repeat it in the main summary log lines. In
addition, the variable process ID often breaks TBPL bug suggestions, so should
be avoided where possible.
 2013-03-18 17:18:08 +00:00
Ed Morley
12b244922e Bug 850681 - Part 10: Tweak comments; r=ted 2013-03-18 17:18:08 +00:00
Ed Morley
3c3dd579bf Bug 850681 - Part 9: Return early if possible during leak summary generation; r=ted 
Now that we close the log file before summary generation (rather than at the
end of processSingleLeakFile()), we can return early in some cases during
leak summary generation.
 2013-03-18 17:18:07 +00:00
Ed Morley
c1a0c09ea1 Bug 850681 - Part 8: Only parse the leak log once; r=ted 
The existing code parses the leak log twice - first time to output the
raw log lines & and the second to output the analysis of them.

As much fun as it is to loop all the things (\o/), let's not.
 2013-03-18 17:18:07 +00:00
Ed Morley
749c5bccd8 Bug 850681 - Part 7: Use the cleaner |with open() as foo| pattern & and don't close/reopen the log unnecessarily; r=ted 2013-03-18 17:18:06 +00:00
Ed Morley
aab3340fe2 Bug 850681 - Part 6: Clean up log parsing conditionals & move the leak threshold reminder to the calling function; r=ted 
The same threshold is used for all leak logs processed from the test run, so
unnecessary to repeat for each.
 2013-03-18 17:18:06 +00:00
Ed Morley
6d7dab1c5b Bug 850681 - Part 5: Clean up logging; r=ted 2013-03-18 17:18:05 +00:00
Ed Morley
f2d6987c21 Bug 850681 - Part 4: Add a partial list of leaked objects to the leak summary, to allow TBPL to make bug suggestions; r=ted 
Current summary line:
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 216 bytes during test execution

With patch parts 1-4:
TEST-UNEXPECTED-FAIL | leakcheck | 180 bytes leaked (Mutex, ReentrantMonitor, nsRunnable, nsTArray_base, nsThread)

...or with ellipses if more than 5 objects:
TEST-UNEXPECTED-FAIL | leakcheck | 3184 bytes leaked (BackstagePass, CondVar, Mutex, ReentrantMonitor, SharedScriptableHelperForJSIID, ...)
 2013-03-18 17:18:05 +00:00
Ed Morley
a5009bbcf3 Bug 850681 - Part 3: Output all of the leaked object lines as TEST-INFO rather than a combination of TEST-INFO and TEST-UNEXPECTED-FAIL; r=ted 
Since the main leak summary will soon include an overview of the leaked objects,
we no longer need/want a subset of the individual leak lines showing up in
TBPL's annotated summary (TBPL's parser matches on 'TEST-UNEXPECTED-FAIL').
 2013-03-18 17:18:04 +00:00
Ed Morley
481990d90a Bug 850681 - Part 2: Wait until the full leak log has been processed before outputting the leak total; r=ted 
Will allow the addition of an overview of the leaked objects, in part 4.
 2013-03-18 17:18:04 +00:00
Ed Morley
1ea3d10926 Bug 850681 - Part 1: Use the test name 'leakcheck' rather than 'automationutils.processLeakLog()'; r=ted 
Reduces the chance of line wrapping in the TBPL annotated summary, and makes
my eyes bleed less every time I see leak failures.
 2013-03-18 17:18:03 +00:00
Ryan VanderMeulen
35ddb95cc4 Backed out changeset 1c19d2a03d90 (bug 827446) for packaging bustage on a CLOSED TREE. 2013-03-13 09:22:00 -04:00
Joel Maher
705f3765df Bug 827446 - update mochitest, reftest, xpcshell to use mozcrash. r=jhammel 2013-03-13 07:38:00 -04:00
Ed Morley
eea19fd01b Bug 813577 - Use try/finally in checkForCrashes() to ensure minidump temp directories do not linger; r=gps 2013-02-08 16:32:07 +00:00
Ed Morley
32277d1f42 Bug 782633 - Followup to fix exception; r=me 2013-01-12 18:36:54 +00:00
Ed Morley
14111b91f9 Bug 782633 - Make the "missing output line for total leaks!" failure mode a warning rather than TEST-UNEXPECTED-FAIL; r=ted 2013-01-11 22:20:21 +00:00
Ed Morley
a9d50d2c9b Bug 817349 - Ensure topFrame is always defined; r=ted 2012-12-03 14:30:35 +00:00
Ed Morley
b1ba5d090d Bug 813650 - Part 2: checkForCrashes() should output the top frame of a crash in a TBPL-parsable format; r=ted 2012-12-01 17:02:29 +00:00
Ed Morley
400c9ed4d1 Bug 813650 - Part 1: Don't print minidump stackwalk output incrementally, so we can implement part 2 without changing log order; r=ted 2012-12-01 16:58:40 +00:00
Ed Morley
b138aec7f5 Bug 808410 - checkForCrashes should return true for cases where we hit an exception during minidump stackwalk; r=jmaher 2012-11-20 15:24:28 +00:00
Jesse Ruderman
27df8178bf Bug 789275 - Better error message when trying to process a crash report without symbols. r=wlach. DONTBUILD 2012-09-15 11:33:02 -07:00
Tim Taubert
35294c0950 Bug 728294 - Part 1 - Remove old debug log parser; r=ted 2012-08-03 12:36:59 +02:00
Ed Morley
20d737e7b9 Bug 778433 - Raise ShutdownLeakLogger's MAX_LEAK_COUNT to something less annoying, until bug 728294 lands; r=me DONTBUILD on a CLOSED TREE 2012-07-28 16:22:47 +01:00
Tim Taubert
6ebcbf7468 Bug 728426 - Opening and then closing bookmarks sidebar keeps the bookmarksPanel.xul and/or history-panel.xul document alive; r=dao 2012-07-21 19:37:56 +02:00
Dão Gottwald
4bda531d79 Bug 767896 - disable browser_pageInfo.js 
--HG--
extra : rebase_source : 315e8b6c151c6c2c23fd2b3275addc7178dd36c2
 2012-07-15 17:04:07 +02:00
Ed Morley
8c73d3a4e7 Bug 772388 - Make ShutdownLeakLogger output the list of leaking tests in a format that TBPL can individually star; r=ttaubert 2012-07-10 13:19:48 +01:00
Gervase Markham
82ff7027aa Bug 716478 - update licence to MPL 2. 2012-05-21 12:12:37 +01:00
Dão Gottwald
7a72c93bb4 Bug 734172 - Stop blaming tests that replace the first tab for leaking. r=ttaubert 2012-05-17 16:41:08 +02:00
Tim Taubert
0530b3807b Bug 753815 - Intermittent INFO | runtests.py | Received unexpected exception while running application; r=ted 2012-05-14 17:02:26 +02:00
Dão Gottwald
c43c50a21d Bug 658738 - reduce MAX_LEAK_COUNT to match the status quo 2012-05-11 12:18:50 +02:00
Dão Gottwald
b42f976cee Bug 752467 - disable leaky tests. r=zpao 2012-05-10 20:09:25 +02:00
Phil Ringnalda
ed28fa6fef Merge m-c to m-i 2012-05-04 19:49:34 -07:00
Tim Taubert
956a1c78af Bug 740242 - runtests.py | Received unexpected exception while running application 'NoneType' object has no attribute 'group'; r=ted 2012-05-04 16:21:24 +02:00
Dão Gottwald
600b237c05 Bug 749361 - reduce MAX_LEAK_COUNT to match the status quo. r=ttaubert 2012-04-27 03:45:49 +02:00
Dão Gottwald
cfc8954d0b Bug 751334 - Redundant TabView.init call in restoreWindow leaks the browser window when the window closes before delayedStartup was called. r=ttaubert 2012-05-04 18:17:25 +02:00
Dão Gottwald
5e01e3dceb Bug 734554 - Increase MAX_LEAK_COUNT in order to reduce the random orange volume 2012-03-13 13:51:56 +01:00
Dão Gottwald
b2a028fe6f No bug - reduce MAX_LEAK_COUNT to match the status quo 2012-03-09 11:39:11 +01:00
Dão Gottwald
6ca7b3d5f9 No bug - reduce MAX_LEAK_COUNT to match the status quo 2012-03-07 11:59:58 +01:00
Tim Taubert
41ee51dd0f Bug 730773 - Track shutdown leaks when DOMWindows *or* DocShells leaked (not and); r=ctalbert 2012-03-02 12:20:20 +01:00
Tim Taubert
242b4e116d Bug 683953 - Part 1 - Parse log output and collect statistics about leaked DOMWindows and DocShells; r=ted 2012-02-25 12:48:05 +01:00
Jesse Ruderman
443df5a432 Bug 703768 - Don't confuse Windows paths with URLs. r=ted 2011-11-29 14:13:08 -08:00
William Lachance
8370e105a7 Bug 679759 - Drop MINIDUMP_STACKWALK_CGI support, let harness download symbols as needed; r=ted 2011-11-15 04:33:21 +00:00
Cameron McCormack
9d4035280b Bug 642175 - Part 2: Allow mochitests to clean up plugin and IPC process crash dumps. r=ted 2011-06-21 12:11:50 +12:00
Jesse Ruderman
82fe1929dc Bug 606389 - when minidump_stack fails, show its error messages and/or exit code. r=ted, a=npotb 2010-11-06 13:30:32 -07:00
Ted Mielczarek
68eff2c341 bug 598507 - wrap test harness process execution with 'arch -arch i386' on OS X 10.5. r=catlee, a=blocking 
--HG--
extra : rebase_source : 7212fa8332fe6853f6dedd6bd23dc937f4b9ec86
 2010-09-23 12:19:31 -04:00
Ted Mielczarek
8ebae2dc0a bug 569965 - automationutils.processLeakLog() should learn not to TEST-UNEXPECTED-FAIL every single object line when there are lots. r=philor 
--HG--
extra : rebase_source : 4ce122222020d522f742b7e77bfacafe8f7f90bc
 2010-06-10 08:27:47 -04:00