Provide a separate app bundle for subprocesses like the renderer on the Mac....

Provide a separate app bundle for subprocesses like the renderer on the Mac.
Remove LSUIElement and related hacks from the browser's app bundle.

BUG=8044
TEST=Observe one Chromium and one or more Chromium Helper processes in
     Activity Monitor

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=22981

[Mark Mentovai, 2009-08-07 21:43:42]

The app portion of this is complete, but there are still a couple of build tasks
I need to do:

 - Whack the helper's Info.plist for Breakpad and other things
 - Introduce the helper to dump_app_syms

[TVL, 2009-08-07 22:45:42]

having really looked yet, but while I think of it.

base/process_util? and something within app (and maybe the custom code I put
within the perf tests to get memory sizes) may be using process names to find
browsers vs. renderers, you'll need to check how they will deal with this change
also.

http://codereview.chromium.org/164177/diff/16/22
File chrome/app/helper-Info.plist (left):

http://codereview.chromium.org/164177/diff/16/22#oldcode148
Line 148: <string>app.icns</string>
do you want an icon on the help?

[TVL, 2009-08-08 17:04:16]

http://codereview.chromium.org/164177/diff/16/21
File chrome/chrome.gyp (right):

http://codereview.chromium.org/164177/diff/16/21#newcode4506
Line 4506: ['exclude', ''],
would .* as the exclude pattern be clearer?

[Mark Mentovai, 2009-08-10 18:03:33]

thomasvl@chromium.org wrote:
> base/process_util? and something within app (and maybe the custom code I
> put within the perf tests to get memory sizes) may be using process
> names to find browsers vs. renderers, you'll need to check how they will
> deal with this change also.

I'd rather just commit it and see what goes wonky...

I hate process_util so hard.

> http://codereview.chromium.org/164177/diff/16/22#oldcode148
> Line 148: <string>app.icns</string>
> do you want an icon on the help?

I don't think so.  It's not a UI process, its icon won't ever show up
anywhere anyway.  Even in the cracktivity monitor, LSUIElemented
processes get no icon.  We get it now on unrebundled Chrome's browser
process because of the TransformProcess call, but we don't get it on
the renderers.

If there were something that showed the icon, I'd consider giving it
one, but I'd think it should be badged somehow.  Since it doesn't show
up anywhere I'm aware of, we might as well save the 216kB (! - really,
that's what app.icns is now) in our app bundle.

> http://codereview.chromium.org/164177/diff/16/21#newcode4506
> Line 4506: ['exclude', ''],
> would .* as the exclude pattern be clearer?

It's just as clear to me, but I can see how it'd probably be clearer
to others, so I've changed this.

Mark

[Mark Mentovai, 2009-08-10 20:43:00]

I've rolled in the changes to dump_app_syms and tweak_app_infoplist.  Ready for
another look.

[TVL, 2009-08-10 21:03:10]

lgtm

http://codereview.chromium.org/164177/diff/1067/1078
File build/mac/dump_app_syms (right):

http://codereview.chromium.org/164177/diff/1067/1078#newcode48
Line 48: DSYM_TAR_PATH="${BUILT_PRODUCTS_DIR}/${SRC_APP_NAME}.dSYM.tar.bz2"
should this is dSYMs now since it is more then one dsym in the archive?

[ladiesman13_gmx.com, 2009-08-11 16:54:35]

Yes. What that mean??? Lgtm

Sent from my iTouch!!

On Aug 10, 2009, at 2:03 PM, thomasvl@chromium.org wrote:

>
> lgtm
>
>
> http://codereview.chromium.org/164177/diff/1067/1078
> File build/mac/dump_app_syms (right):
>
> http://codereview.chromium.org/164177/diff/1067/1078#newcode48
> Line 48:
> DSYM_TAR_PATH="${BUILT_PRODUCTS_DIR}/${SRC_APP_NAME}.dSYM.tar.bz2"
> should this is dSYMs now since it is more then one dsym in the  
> archive?
>
> http://codereview.chromium.org/164177