60004e881f
Create a new canvas for each test run; also, don't hang if an exception is thrown, instead print an error and exit
2007-08-02 01:53:53 -07:00
de77792dce
Bug 376489 - Serve reftests from a web server so as not to depend on external resources, if the reftest is marked as needing HTTP; also picks up support for HTTP header and status modification in reftest files. r=dbaron
2007-07-18 14:32:50 -07:00
961f06f7e8
Add checkLoadURI calls to reftest harness. b=371075 r=bzbarsky
2007-06-12 11:25:15 -07:00
9e55a6063a
Fixing bug 379823. Allow conditional skipping of reftests (reporting them as failures). r+sr=dbaron@mozilla.com
2007-05-08 03:21:22 -07:00
39d621400a
Use getAttribute('class') instead of className so that reftest can run SVG tests. On SVG elements className is of type SVGAnimatedString instead of DOMString.
2007-03-27 11:06:35 -07:00
sharparrow1@yahoo.com
5d92f811db
Bug 374050: add printing test mode to reftests. See bug/sample test for how to make printing tests.
2007-03-21 06:07:42 +00:00
asqueella@gmail.com
5fa86e826f
fix bug 371026 - reftest: double setTimeout from onload is not late enough for some tests
...
the fix is to provide a way for the testcase to indicate when it's finished loading.
Also move the reftest self-checks into their own manifest.
r=dbaron
2007-03-07 19:50:02 +00:00
asqueella@gmail.com
8d42473681
bug 371059 - change "FAIL" to "KNOWN FAIL" for expected test failures in reftest (to avoid confusion)
...
r=dbaron
2007-03-01 06:26:23 +00:00
bzbarsky@mit.edu
a36f13fa10
Output the spec of the manifest URI, not a stringification of the XPConnect
...
wrapper. No bug, r=dbaron
2007-02-28 04:35:47 +00:00
dbaron@dbaron.org
c37a1bd0bf
Put back the load timeout code that I backed out yesterday, since it now works. b=369319
2007-02-13 00:54:02 +00:00
dbaron@dbaron.org
c9a615fc69
Ignore load events for subframes. More followup to bug 369319.
2007-02-13 00:34:58 +00:00
dbaron@dbaron.org
b899182dea
Do setTimeout twice so I don't break tests that do a setTimeout. Followup to bug 369319.
2007-02-13 00:03:50 +00:00
dbaron@dbaron.org
66a553c02e
Revert previous change, since it doesn't help.
2007-02-12 21:50:54 +00:00
dbaron@dbaron.org
88f010802c
Put back 20ms timeout for now, to try to fix test failures.
2007-02-12 20:57:30 +00:00
dbaron@dbaron.org
5e1125be0a
Back out failure timeout part of previous change.
2007-02-12 03:45:01 +00:00
dbaron@dbaron.org
ab43f8b499
Make reftest use onload rather than nsIWebProgressListener, and make it time out after load failure. b=369319 r=robcee
2007-02-12 02:21:18 +00:00
dbaron@dbaron.org
fb87815e54
Allow reftest to be marked as failing or random on a specific platform. b=369330 r=robcee
2007-02-08 19:24:43 +00:00
dbaron@dbaron.org
2079d56f5e
Fix missed renaming in reftest. b=368158 r=rcampbell
2007-01-26 06:56:58 +00:00
dbaron@dbaron.org
67a774997a
Add code so that reftest quits properly on Mac OS X. Patch by Rob Campbell <rcampbell@mozilla.com>. r=dbaron b=363985
2007-01-08 18:22:48 +00:00
davel@mozilla.com
e657d949e9
b=344591 r=dbaron trim leading and trailing whitespace from manifest lines
2006-10-28 04:17:21 +00:00
davel@mozilla.com
33a864e3c2
b=344591, r=dbaron
...
fix parsing of comment lines
2006-10-27 19:56:49 +00:00
dbaron@dbaron.org
b45366d6e3
Check in XUL framework for acceptance tests with HTML reference renderings. Not part of build. b=344591
2006-09-05 21:50:54 +00:00
