[Mac] Enforce proper usage of main app bundle and helper app bundle....

[Mac] Enforce proper usage of main app bundle and helper app bundle.
- Add a check to make sure the helper isn't used without --type
- Add a check to make sure the main app isn't used with --type.
- Make inproc browser tests invoke the helper that is within the app bundle on disk to recreate the real world environment.

BUG=23645, 52858
TEST=green tree, invoking the main app with --type fails, and invoking the helper without it fails.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=57231

[TVL, 2010-08-24 17:51:39]

Look at diff 1 for try bot results (uploaded again to make sure I hadn't made
other changes).

Passes all the tests run on the mac10.? Test * bots, so it seems like it covers
things.

[Mark Mentovai, 2010-08-24 17:58:13]

LG with questions and things.

http://codereview.chromium.org/3171027/diff/4001/5001
File chrome/app/chrome_dll_main.cc (right):

http://codereview.chromium.org/3171027/diff/4001/5001#newcode529
chrome/app/chrome_dll_main.cc:529: std::string process_type =
Do it here?

You wrote the thing about initialization order down below. Does it make a
difference?

http://codereview.chromium.org/3171027/diff/4001/5001#newcode666
chrome/app/chrome_dll_main.cc:666: DCHECK(!process_type.empty())
Instead of checking for empty, can you check that the --type flag was not
provided? I think this will still be empty if you use --type or --type=.

http://codereview.chromium.org/3171027/diff/4001/5001#newcode667
chrome/app/chrome_dll_main.cc:667: << "The helper app bundle should only be used
with the --type switch!";
CHECK, not DCHECK.

http://codereview.chromium.org/3171027/diff/4001/5001#newcode669
chrome/app/chrome_dll_main.cc:669: DCHECK(process_type.empty())
Same.

http://codereview.chromium.org/3171027/diff/4001/5001#newcode926
chrome/app/chrome_dll_main.cc:926: NOTREACHED() << "Unknown process type";
I’d move this to a CHECK(false). NOTREACHED is a DCHECK(false).

[TVL, 2010-08-24 18:17:23]

http://codereview.chromium.org/3171027/diff/4001/5001
File chrome/app/chrome_dll_main.cc (right):

http://codereview.chromium.org/3171027/diff/4001/5001#newcode529
chrome/app/chrome_dll_main.cc:529: std::string process_type =
On 2010/08/24 17:58:13, Mark Mentovai wrote:
> Do it here?
> 
> You wrote the thing about initialization order down below. Does it make a
> difference?

I was going to, but checking for background pulls in mainbundle, so it wasn't
clear just how much cocoa comes in, and I didn't want to risk mucking with
initialization order given how long it's take to track down problems in the past
with first launch, etc.

http://codereview.chromium.org/3171027/diff/4001/5001#newcode666
chrome/app/chrome_dll_main.cc:666: DCHECK(!process_type.empty())
On 2010/08/24 17:58:13, Mark Mentovai wrote:
> Instead of checking for empty, can you check that the --type flag was not
> provided? I think this will still be empty if you use --type or --type=.

Done.

http://codereview.chromium.org/3171027/diff/4001/5001#newcode667
chrome/app/chrome_dll_main.cc:667: << "The helper app bundle should only be used
with the --type switch!";
On 2010/08/24 17:58:13, Mark Mentovai wrote:
> CHECK, not DCHECK.

Done.

http://codereview.chromium.org/3171027/diff/4001/5001#newcode669
chrome/app/chrome_dll_main.cc:669: DCHECK(process_type.empty())
On 2010/08/24 17:58:13, Mark Mentovai wrote:
> Same.

Done.

[Mark Mentovai, 2010-08-24 18:24:40]

LGTM

http://codereview.chromium.org/3171027/diff/14001/1003
File chrome/app/chrome_dll_main.cc (right):

http://codereview.chromium.org/3171027/diff/14001/1003#newcode661
chrome/app/chrome_dll_main.cc:661: // The helper's info.plist marks it as a
background only app (so it won't
Info.plist (capital I).

Don’t use parentheses here.

http://codereview.chromium.org/3171027/diff/14001/1003#newcode665
chrome/app/chrome_dll_main.cc:665: << "The helper app bundle should only be used
with the --type switch!";
<< "Helper application requires --type.";

http://codereview.chromium.org/3171027/diff/14001/1003#newcode668
chrome/app/chrome_dll_main.cc:668: << "The main app bundle should not be used
with the --type switch!"
<< "Main application forbids --type, saw \"" << process_type << "\".";