Bug 64276 – Attachment 99745 failed to build on Windows EWS, but the Windows EWS bot didn't comment in the bug saying so

• Component: Tools / Tests
• Status: RESOLVED CONFIGURATION CHANGED

[Adam Roben (:aroben)]

In bug 63330, attachment 99745 [details] failed to build on the Windows EWS. You can see this by the red "win" box underneath the patch. But you can also see that there's no comment in the bug indicating that the patch failed to build.

I wonder what happened here?

[Adam Barth]

Can you grab the logs from the bot?

[Lucas Forschler]

I checked all of the EWS bots and this log file does not exist.

[Eric Seidel (no email)]

This looks like the log?
http://queues.webkit.org/results/8988457

This is the attachment page:
http://queues.webkit.org/patch/99745

I just clicked on the bubble next to the attachment, I assume you all did the same, so I must be missing something here.

[Eric Seidel (no email)]

29>Linking...
29>gtest.lib(gtest.obj) : warning LNK4075: ignoring '/EDITANDCONTINUE' due to '/SAFESEH' specification
29>LINK : warning LNK4098: defaultlib 'MSVCRTD' conflicts with use of other libs; use /NODEFAULTLIB:library
29>StringOperators.obj : error LNK2019: unresolved external symbol "public: __thiscall WTF::String::String(char const *)" (??0String@WTF@@QAE@PBD@Z) referenced in function "private: virtual void __thiscall TestWebKitAPI::WTF_StringOperators_Test::TestBody(void)" (?TestBody@WTF_StringOperators_Test@TestWebKitAPI@@EAEXXZ)
29>StringOperators.obj : error LNK2019: unresolved external symbol "private: static class WTF::PassRefPtr<class WTF::StringImpl> __cdecl WTF::AtomicString::add(char const *)" (?add@AtomicString@WTF@@CA?AV?$PassRefPtr@VStringImpl@WTF@@@2@PBD@Z) referenced in function "public: __thiscall WTF::AtomicString::AtomicString(char const *)" (??0AtomicString@WTF@@QAE@PBD@Z)
29>C:\cygwin\home\buildbot\Webkit\WebKitBuild\Debug\bin\TestWebKitAPI.exe : fatal error LNK1120: 2 unresolved externals
29>Build log was saved at "file://C:\cygwin\home\buildbot\Webkit\WebKitBuild\Debug\obj\TestWebKitAPI\BuildLog.htm"
29>TestWebKitAPI - 3 error(s), 2 warning(s)

is the error I see?

[Eric Seidel (no email)]

Oh, I see now.  You're wondering why the win-ews didn't comment?  Bugzilla was down for a while this weekend, it's possible it was down briefly when this thing tried to build 5 days ago.  Not sure why we don't see logs of the build event though.

[Eric Seidel (no email)]

I can change the patch status page to include information on where the log message came from if that would help?

http://trac.webkit.org/browser/trunk/Tools/QueueStatusServer/templates/patch.html

It should really just use this include instead of rolling its own:
http://trac.webkit.org/browser/trunk/Tools/QueueStatusServer/templates/includes/singlequeuestatus.html

[Adam Roben (:aroben)]

(In reply to comment #5)
> You're wondering why the win-ews didn't comment?

Yes.

> Bugzilla was down for a while this weekend, it's possible it was down briefly when this thing tried to build 5 days ago.  Not sure why we don't see logs of the build event though.

I uploaded attachment 99745 [details] almost a week ago. I'd assume that the EWS bot processed well before this weekend.

[Eric Seidel (no email)]

Yes, it processed your patch 5 days, 21 hours ago. http://queues.webkit.org/patch/99745

[Ahmad Saleem]

[e-mail] - Is something required further on this? I don't think by reading through comment? Appreciate your input. Thanks!

[Alexey Proskuryakov]

EWS has been reimplemented from scratch since then.