Commit Graph

55 Commits

Author SHA1 Message Date
Ted Mielczarek
18e08ab904 bug 460515 - Remove assumption that xpcshell etc in same directory as app executable. r=jwalden 2009-02-09 13:57:27 -05:00
bjarne@runitsoft.com
3a0930b339 Bug 464191 - ssltunnel process persists after Mochitest run. r=jwalden, a=testingonlychange 2009-01-12 11:23:28 -08:00
Shawn Wilsher
5de65e03a2 Bug 473845 - Break a potential cycle with XPConnect and mozStroageService
This changeset removes a cycle that gets created between XPConnect and the
storage service.  This accomplishes this by releasing the cached reference to
XPConnect during xpcom-shutdown.

This was the last remaining leak of the browser-chrome test harness, so this
also sets the leak threshold to zero for that test.

r=bent
2009-01-21 16:52:16 -05:00
Dave Townsend
3d41d01fa0 Bug 435743: Increase expected browser-chrome leaks to compensate for bug 473802. r=ted 2009-01-21 13:18:33 +00:00
Jeff Walden
e875242af2 Sigh, last time this broke wasn't just a passing anomaly. Backed out changeset 82a1d612dde8 2009-01-18 20:12:32 -08:00
bjarne@runitsoft.com
a706f0544a Bug 464191 - ssltunnel process persists after Mochitest run. r=jwalden 2009-01-12 11:23:28 -08:00
Jeff Walden
99efd28560 Backed out changeset a2018012b3ee, may have broken tree (in an inscrutable way, unlike last time :-( ) 2009-01-15 17:16:35 -08:00
bjarne@runitsoft.com
289ea62e3d Bug 464191 - ssltunnel process persists after Mochitest run. r=jwalden 2009-01-12 11:23:28 -08:00
Honza Bambas
48b4e62a7e Bug 456001 - Need automated testing for SSL client auth, r=jwalden 2009-01-15 20:19:15 +01:00
Robert Sayre
b32137338d bug 472706 - add better profiling input for spidermonkey in PGO builds. r=ted
--HG--
rename : modules/libpr0n/test/unit/image3.ico => build/pgo/favicon.ico
2009-01-13 07:34:19 -05:00
Jeff Walden
e0507cdd3f Backed out changeset a12bffb28598; it's in the way of other backouts that need to happen. 2009-01-12 12:36:00 -08:00
bjarne@runitsoft.com
f81c16ba91 Bug 464191 - ssltunnel process persists after Mochitest run. r=jwalden 2009-01-12 11:23:28 -08:00
Jeff Walden
3fe3bfbf72 Bug 469509 - runtests.py.in : don't print leak lines for classes that aren't leaked. r=ted 2009-01-08 16:52:07 -08:00
Serge Gautherie
45fb9c8423 Bug 460548 - Port |Bug 450983 - use leakThreshold for SeaMonkey testers| to Firefox; (Bv1) Reduce |browserChrome| to 72500 from 75000; blanket-r=ted.mielczarek 2009-01-06 17:07:05 +01:00
arno renevier
ad51a390d9 bug 470415 - call exanduser on mochitest logfile argument. r=ted 2008-12-19 15:01:37 +01:00
Justin Dolske
17d229a4bc Revert part of bug 460548, increasing the leak to 75000 again. 2008-12-17 14:21:36 -08:00
Serge Gautherie
5200089bae Bug 469444 - steps.py : MozillaA11YTest misses __init__() and evaluateCommand(); (Av1) Explicitly set default leakThreshold to 0 in runtests.py.in; r=ted.mielczarek 2008-12-13 22:02:18 +01:00
Serge Gautherie
2ea48bae9b Bug 460548 - Port |Bug 450983 - use leakThreshold for SeaMonkey testers| to Firefox; (Bv1) Reduce |browserChrome| to 74000 from 75000; r=ted.mielczarek
(Relanding after unrelated backout.)
2008-12-11 16:09:47 +01:00
Robert Sayre
50809f236c backout bug 460548, checked in on persistent orange. 2008-12-11 02:41:46 -08:00
Serge Gautherie
b77d5d6780 Bug 460548 - Port |Bug 450983 - use leakThreshold for SeaMonkey testers| to Firefox; (Bv1) Reduce |browserChrome| to 74000 from 75000; r=ted.mielczarek 2008-12-11 02:29:25 +01:00
Serge Gautherie
007d700b48 Bug 460548 - Port |Bug 450983 - use leakThreshold for SeaMonkey testers| to Firefox; (Av1) Reduce |browserChrome| to 75000 from Infinity; r=ted.mielczarek 2008-11-20 19:09:33 +01:00
Serge Gautherie
cbfd8ca0ff Bug 456275 - In (Brief/Full) log of (test) builds, remind leak threshold on known leak too; Av2a; r=ted.mielczarek 2008-10-22 01:16:39 +02:00
Serge Gautherie
0ec802ff8d Bug 456275 - In (Brief/Full) log of (test) builds, remind leak threshold on known leak too
(Av2) sync and extend
2008-10-21 02:47:51 +02:00
Serge Gautherie
f4ba40e39a Bug 460382 - Linux Firefox test boxes don't error on mochitest leaks; (Av1) No leaks on Linux too; r=ted.mielczarek 2008-10-18 04:36:50 +02:00
Justin Dolske
a2131b5ec1 Bug 456894 - runtests.py can't handle negative leaks. r=dbaron, r=ted 2008-10-10 14:44:35 -07:00
Ted Mielczarek
b12044f77e bug 417516 - add top level Makefile targets to run test suites. add targets for mochitest. r=bsmedberg 2008-07-27 13:06:58 -04:00
Nick Thomas
dfc871a499 Bug 443090 - Unify unittest failures (runtests leaks part) p=ted.mielczarek, r=sayrer 2008-07-16 17:49:32 +12:00
Ted Mielczarek
ac96956394 make runtests.py --log-file handle relative pathnames. r=sayrer (no bug) 2008-07-01 15:16:05 -04:00
Jeff Walden
c8154e48f9 Bug 435376 - Start making leaks fatal when running tests in setups where we don't leak. r=sayrer 2008-05-22 15:08:41 -04:00
jwalden@mit.edu
112e96d0fe Bug 431900 - Add a commandline flag to make assertions fatal, for easy use in tinderbox scripts and the like, especially if we eventually set up a run of all the tests that don't trigger assertions and then run them with this argument specified. r=sayrer, a=testonlychange 2008-05-06 10:40:41 -07:00
jwalden@mit.edu
798c6c4d6f Bug 428758 - Clean up runtests.py help messages: move log options into the individual log-level arguments, use custom variables instead of autogenerated ones for --foo=FOO, so FOO is more concise for a bunch of different foo. Also remove the restriction that --file-level must be provided if --log-file is provided by giving a default (we only use that default if a log file is specified). r=testingonlychange 2008-04-14 16:21:25 -07:00
dveditz@cruzio.com
6075900255 fixing test bustage from restrictions imposed by bug 292789 (chrome://mochikit needs contentaccessible=yes). r=gavin 2008-04-12 18:31:00 -07:00
jwalden@mit.edu
74a2f73c04 Bug 418009 - Make runtests.py properly handle stdout for both the server and the application processes, rather than just sitting on the server output and hoping it doesn't grow 'too large'. This may fix some of the problems that were plaguing the PGO box that was using runtests.py a few days ago, and it should fix hangs numerous people have seen running Mochitests lately, where the server process decides to spew gobs of output about stuff being leaked. r=robcee 2008-04-07 22:18:45 -07:00
ajschult@verizon.net
810e830597 Bug 424501: disable gnome bug-buddy dialog from runtests, r=waldo 2008-03-22 09:44:57 -07:00
mnyromyr@tprac.de
7137e8fe45 Bug 423035: can't run single chrome/a11y test automatically; r=Waldo 2008-03-18 12:57:22 -07:00
gavin@gavinsharp.com
93ec72cc45 Bug 423382: float('Inf') doesn't work everywhere, use 1.0e3000 instead, r=jwalden 2008-03-17 06:06:16 -07:00
jwalden@mit.edu
02a07c4d0b Bug 423235 - Make leak detection always run but only be fatal if the threshold is exceeded, for better leak visibility when --leak-threshold isn't specified; also removes the need for me to maintain a --leak-threshold=0 local change without committing it. r=sayrer, a=testonlychange 2008-03-15 21:11:44 -07:00
jwalden@mit.edu
d76b4a5e99 Revert a local change that isn't be ready yet -- need tinderboxen doing leak tests before we can actually make this change, sigh. 2008-03-15 17:52:18 -07:00
jwalden@mit.edu
1bde50a721 Bug 423106 - runtests.py doesn't do anything with --browser-arg. r=ted 2008-03-15 15:48:08 -07:00
jwalden@mit.edu
82499b6e91 Make $objdir/_tests/testing/mochitest/runtests.py runnable from any directory, not just $objdir/_tests/testing/mochitest/. r=sayrer, NPOTB and all that anyway 2008-03-02 20:05:22 -08:00
jwalden@mit.edu
eefd40c732 Bug 420154 - Print finer-grained leak info in runtests.py than just pass/fail and a table, and don't fall over when --leak-threshold is used in a build without refcount logging (just print a warning). r=sayrer 2008-02-29 19:40:45 -08:00
kairo@kairo.at
836da1985a bug 419826 - Make browser mochitests run on SeaMonkey, r=gavin 2008-02-28 13:36:30 -08:00
jwalden@mit.edu
96f0392ae0 Bug 419339 - Followup to change an error condition (which should never be hit) to use ERROR FAIL instead of FAIL in the error message, to trigger the right log-scraping behavior. 2008-02-24 18:59:04 -08:00
jwalden@mit.edu
ae17c91e79 Bug 419339 - Add an argument to runtests.py to specify the maximum allowable number of RLk-style bytes to be leaked, defaulting to any leaks allowable for now. Tinderboxen will need to pick up the extra argument and provide sane values until we can eradicate all leaks, at which point we can change the default to 0 to always check. r=sayrer 2008-02-24 16:54:31 -08:00
sayrer@gmail.com
499bc27b1f Bug 418772. PGO scripts and input. r=ted.mielczarek. 2008-02-21 13:08:39 -08:00
surkov.alexander@gmail.com
f354efac8e Bug 416553 - option to run accessibility tests, r=rcampbell 2008-02-14 02:32:21 -08:00
surkov.alexander@gmail.com
e9b16fd1ff Back out bug 416553 2008-02-13 23:33:19 -08:00
surkov.alexander@gmail.com
9b6e8173e6 Bug 416553 - option to run accessibility tests, r=rcampbell 2008-02-13 22:20:01 -08:00
ajschult@verizon.net
c8e3809c06 Bug 416570: runtests should not pull in bookmarks chrome, r=sayrer 2008-02-11 21:03:09 -08:00
ajschult@verizon.net
b7946e60ca Bug 416775: runtests.py prints lots of newlines after app exits, r=Waldo 2008-02-11 20:30:18 -08:00