37588 2010-04-14 11:21:01 -0700 Make new-run-webkit-tests work for the Qt port 2010-04-21 04:04:41 -0700 1 1 1 Unclassified WebKit New Bugs 528+ (Nightly build) Other OS X 10.5 RESOLVED FIXED P2 Normal --- 37765 1 vestbo webkit-unassigned abarth commit-queue eric ossy oldest_to_newest 212449 0 vestbo 2010-04-14 11:21:01 -0700 Make new-run-webkit-tests work for the Qt port 212453 1 53346 vestbo 2010-04-14 11:23:01 -0700 Created attachment 53346 Patch 212465 2 53346 eric 2010-04-14 11:33:53 -0700 Comment on attachment 53346 Patch Technically we use " instead of ' (or try to in our python). Eventually we need to set up a self._built_bin_path for the linux ports since both DumpRenderTree and ImageDiff will have this trouble. For Qt it's "bin", for Gtk its "Programs", migth as well share the code. But we can come back and do that later. How does this detect the right configuration? I still owe Adam Barth a patch to have WebKit.default_configuration read from $BUILD_DIR/Configuration instead of the current path hack that it does. 212466 3 53346 eric 2010-04-14 11:34:29 -0700 Comment on attachment 53346 Patch I don't want to stand in your way, so r+, but consider the above comments. 212496 4 vestbo 2010-04-14 12:34:39 -0700 (In reply to comment #2) > (From update of attachment 53346 [details]) > Technically we use " instead of ' (or try to in our python). Noted. > Eventually we need to set up a self._built_bin_path for the linux ports since > both DumpRenderTree and ImageDiff will have this trouble. For Qt it's "bin", > for Gtk its "Programs", migth as well share the code. But we can come back and > do that later. You mean something that returns the suffix ("bin", "Programs"), or the full path, so that both the Qt and Gtk port implementations would concat _build_path() and the suffix? Happy to do either as part of this patch just to make it cleaner. > How does this detect the right configuration? I still owe Adam Barth a patch > to have WebKit.default_configuration read from $BUILD_DIR/Configuration instead > of the current path hack that it does. It's not picking up custom configurations, but at least on Linux webkit-build-directory actually works for Debug vs Release, so hard coding to "Release" seems wrong in hindsight. 212528 5 eric 2010-04-14 13:21:36 -0700 (In reply to comment #4) > > Eventually we need to set up a self._built_bin_path for the linux ports since > > both DumpRenderTree and ImageDiff will have this trouble. For Qt it's "bin", > > for Gtk its "Programs", migth as well share the code. But we can come back and > > do that later. > > You mean something that returns the suffix ("bin", "Programs"), or the full > path, so that both the Qt and Gtk port implementations would concat > _build_path() and the suffix? I had figured a _build_path replacement for the places that have more specific needs. like _built_bin_path("DumpRenderTree") would be the call for all ports, and on linux ports that would make the right _build_path("bin", *components) call, but on normal ports it would just call _build_path(*components). Don't know. Hadn't written it yet. It's not essential for this change and is somethign we can clean up later when we have more data about our uses. > It's not picking up custom configurations, but at least on Linux > webkit-build-directory actually works for Debug vs Release, so hard coding to > "Release" seems wrong in hindsight. Sounds good to me. 212531 6 53346 eric 2010-04-14 13:22:06 -0700 Comment on attachment 53346 Patch Thanks for the added explanation. 212790 7 53346 commit-queue 2010-04-15 02:44:09 -0700 Comment on attachment 53346 Patch Rejecting patch 53346 from commit-queue. Unexpected failure when landing patch! Please file a bug against webkit-patch. Failed to run "['WebKitTools/Scripts/webkit-patch', '--status-host=webkit-commit-queue.appspot.com', 'land-attachment', '--force-clean', '--build', '--test', '--non-interactive', '--ignore-builders', '--build-style=both', '--quiet', '53346', '--parent-command=commit-queue', '--no-update']" exit_code: 1 Last 500 characters of output: ', 'commit', '--all', '-F', '-'], input=message) File "/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/webkitpy/common/system/executive.py", line 85, in run_command return Executive().run_command(*args, **kwargs) File "/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/webkitpy/common/system/executive.py", line 186, in run_command string_to_communicate = str(input) UnicodeEncodeError: 'ascii' codec can't encode character u'\xf8' in position 26: ordinal not in range(128) 213339 8 eric 2010-04-16 11:53:05 -0700 Sorry again Tor. I'll fix this today. 214165 9 eric 2010-04-19 12:24:10 -0700 Sorry. If bug 37765 isn't landed today, I'll land this change by hand. 214561 10 ossy 2010-04-20 03:58:02 -0700 (In reply to comment #8) > Sorry again Tor. I'll fix this today. It would be good, because webkit-patch hates me too. :) 214894 11 53346 eric 2010-04-20 12:57:42 -0700 Comment on attachment 53346 Patch Sorry again for the trouble. 215228 12 53346 eric 2010-04-21 03:48:59 -0700 Comment on attachment 53346 Patch The fix for bug 37765 was rolled out. So this will fail to land. Hopefully I'll reland the fix tomorrow at which point I'll cq+ this again. 215234 13 53346 eric 2010-04-21 04:03:31 -0700 Comment on attachment 53346 Patch Clearing flags on attachment: 53346 Committed r57965: <http://trac.webkit.org/changeset/57965> 215235 14 eric 2010-04-21 04:03:42 -0700 All reviewed patches have been landed. Closing bug. 215236 15 eric 2010-04-21 04:04:41 -0700 I just went ahead and landed it by hand (using my locally fixed copy of webkit-patch). Again, apologies for the trouble. 53346 2010-04-14 11:23:01 -0700 2010-04-21 04:03:30 -0700 Patch bug-37588-20100414142259.patch text/plain 1448 vestbo ZGlmZiAtLWdpdCBhL1dlYktpdFRvb2xzL0NoYW5nZUxvZyBiL1dlYktpdFRvb2xzL0NoYW5nZUxv ZwppbmRleCAxZDFiM2FiLi41MmNkMjc4IDEwMDY0NAotLS0gYS9XZWJLaXRUb29scy9DaGFuZ2VM b2cKKysrIGIvV2ViS2l0VG9vbHMvQ2hhbmdlTG9nCkBAIC0yLDYgKzIsMTUgQEAKIAogICAgICAg ICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KIAorICAgICAgICBNYWtlIG5ldy1ydW4td2Vi a2l0LXRlc3RzIHdvcmsgZm9yIHRoZSBRdCBwb3J0CisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJr aXQub3JnL3Nob3dfYnVnLmNnaT9pZD0zNzU4OAorCisgICAgICAgICogU2NyaXB0cy93ZWJraXRw eS9sYXlvdXRfdGVzdHMvcG9ydC9xdC5weToKKworMjAxMC0wNC0xNCAgVG9yIEFybmUgVmVzdGLD uCAgPHRvci5hcm5lLnZlc3Rib0Bub2tpYS5jb20+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9C T0RZIChPT1BTISkuCisKICAgICAgICAgbmV3LXJ1bi13ZWJraXQtdGVzdHM6IHRyeSB0byBkZXRl Y3QgYWx0ZXJuYXRlIGFwYWNoZSBwYXRoCiAgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3Jn L3Nob3dfYnVnLmNnaT9pZD0zNzU4NwogCmRpZmYgLS1naXQgYS9XZWJLaXRUb29scy9TY3JpcHRz L3dlYmtpdHB5L2xheW91dF90ZXN0cy9wb3J0L3F0LnB5IGIvV2ViS2l0VG9vbHMvU2NyaXB0cy93 ZWJraXRweS9sYXlvdXRfdGVzdHMvcG9ydC9xdC5weQppbmRleCA4OWEzNWU4Li42N2NkZWZlIDEw MDY0NAotLS0gYS9XZWJLaXRUb29scy9TY3JpcHRzL3dlYmtpdHB5L2xheW91dF90ZXN0cy9wb3J0 L3F0LnB5CisrKyBiL1dlYktpdFRvb2xzL1NjcmlwdHMvd2Via2l0cHkvbGF5b3V0X3Rlc3RzL3Bv cnQvcXQucHkKQEAgLTMxLDYgKzMxLDcgQEAKIGltcG9ydCBsb2dnaW5nCiBpbXBvcnQgb3MKIGlt cG9ydCBzdWJwcm9jZXNzCitpbXBvcnQgc2lnbmFsCiAKIGZyb20gd2Via2l0cHkubGF5b3V0X3Rl c3RzLnBvcnQud2Via2l0IGltcG9ydCBXZWJLaXRQb3J0CiAKQEAgLTkwLDYgKzkxLDkgQEAgY2xh c3MgUXRQb3J0KFdlYktpdFBvcnQpOgogICAgICAgICAgICAgICAgICMgJ2tpbGxhbGwnIHdlYiBz ZXJ2ZXJzLgogICAgICAgICAgICAgICAgIHNlbGYuX3NodXRfZG93bl9odHRwX3NlcnZlcihOb25l KQogCi0gICAgZGVmIGRlZmF1bHRfY29uZmlndXJhdGlvbihzZWxmKToKLSAgICAgICAgIyBGSVhN RTogRG8gdGhpcyBwcm9wZXJseQotICAgICAgICByZXR1cm4gIlJlbGVhc2UiCisgICAgZGVmIF9i dWlsZF9kcml2ZXIoc2VsZik6CisgICAgICAgICMgVGhlIFF0IHBvcnQgYnVpbGRzIERSVCBhcyBw YXJ0IG9mIHRoZSBtYWluIGJ1aWxkIHN0ZXAKKyAgICAgICAgcmV0dXJuIFRydWUKKworICAgIGRl ZiBfcGF0aF90b19kcml2ZXIoc2VsZik6CisgICAgICAgIHJldHVybiBzZWxmLl9idWlsZF9wYXRo KCdiaW4vRHVtcFJlbmRlclRyZWUnKQo=