586183 - e10s merge caused bustage for --disable-ipc in embedding/

Status: RESOLVED FIXED

(Core :: IPC, defect)

Description

[Justin Wood (:Callek)]

Attachment: Fix

../webBrowser/libnsWebBrowser_s.a(nsWebBrowser.o): In function `nsWebBrowser::Create()':
/builds/slave/comm-central-trunk-linux64/build/mozilla/embedding/browser/webBrowser/nsWebBrowser.cpp:1236: undefined reference to `XRE_GetProcessType'
collect2: ld returned 1 exit status

Patch by mfinkle. Doug asked to assign to him, this way if no-one pushes by tomorrow morning, he will.

[Whoever pushes, feel free to change commit message to something more meaningful, e.g. including this bug #]

Comment 1

[Mark Finkle (:mfinkle) (use needinfo?)]

the patch should be:

#ifdef MOZ_IPC
   if (XRE_GetProcessType() == GeckoProcessType_Default)
#endif
   {
       // Hook up global history. Do not fail if we can't - just warn.                                                                                                              
       rv = EnableGlobalHistory(mShouldEnableHistory);
       NS_WARN_IF_FALSE(NS_SUCCEEDED(rv), "EnableGlobalHistory() failed");
   }

Comment 2

[Justin Wood (:Callek)]

Attachment: real fix

Lets try attaching again, this time creating my MQ against the right tree.

Comment 4

[Oleg Romashin (:romaxa)]

Isn't so that we can call XRE_GetProcessType in non IPC builds?
it should just return GeckoProcessType_Default always

http://mxr.mozilla.org/mozilla-central/source/toolkit/xre/nsAppRunner.cpp#3837

Comment 5

[[e-mail]]

I'm surprised it isn't #defined away in non-ipc builds.

Comment 6

[Mark Banner (:standard8)]

(In reply to comment #4)
> Isn't so that we can call XRE_GetProcessType in non IPC builds?
> it should just return GeckoProcessType_Default always
> 
> http://mxr.mozilla.org/mozilla-central/source/toolkit/xre/nsAppRunner.cpp#3837

The problem in shared builds is that nsAppRunner is built after embedding/ iirc

Comment 7

[Robert Kaiser]

Pushed the attached patch as
http://hg.mozilla.org/mozilla-central/rev/9b4466b55eaf

Oleg, if that's true, please let go into a followup, as this was something we
needed to fix heavy bustage on Thunderbird and SeaMonkey trees.