Send the user's Windows profile type up in crash reports.

Send the user's Windows profile type up in crash reports.

I suspect that some of the version shear crashes happen to users with non-local profiles.  In this case, it's easy to imagine that the version info in the registry is out of sync with that on the filesystem.

BUG=78585
TEST=none

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=145892

[grt (UTC plus 2), 2012-07-09 16:11:38]

Hi Siggi,

I figure it might be useful to see if there's a correlation between non-local
user profiles and the version shear crashes.  What do you think about including
this info in crash dumps?

[Sigurður Ásgeirsson, 2012-07-09 16:39:48]

lgtm

https://chromiumcodereview.appspot.com/10740003/diff/1/chrome/app/breakpad_wi...
File chrome/app/breakpad_win.cc (right):

https://chromiumcodereview.appspot.com/10740003/diff/1/chrome/app/breakpad_wi...
chrome/app/breakpad_win.cc:229: if (::GetProfileType(&profile_bits)) {
nit: is it useful to return a special string for failure in this API? It seems
it's accessing HKLM, so things would be seriously messed up if ever it fails.

[grt (UTC plus 2), 2012-07-09 17:03:10]

https://chromiumcodereview.appspot.com/10740003/diff/1/chrome/app/breakpad_wi...
File chrome/app/breakpad_win.cc (right):

https://chromiumcodereview.appspot.com/10740003/diff/1/chrome/app/breakpad_wi...
chrome/app/breakpad_win.cc:229: if (::GetProfileType(&profile_bits)) {
On 2012/07/09 16:39:48, Sigurður Ásgeirsson wrote:
> nit: is it useful to return a special string for failure in this API? It seems
> it's accessing HKLM, so things would be seriously messed up if ever it fails.

Good idea.  PTAL.

[Sigurður Ásgeirsson, 2012-07-10 10:11:14]

lgtm with an ubernit.

https://chromiumcodereview.appspot.com/10740003/diff/11001/chrome/app/breakpa...
File chrome/app/breakpad_win.cc (right):

https://chromiumcodereview.appspot.com/10740003/diff/11001/chrome/app/breakpa...
chrome/app/breakpad_win.cc:249: profile_type.assign(L"error ");
ubernit: I'd grab GLE to a temp at the very top of the file, just in case.
Also, if StringPrintf is already used in setup.exe, that'd make for marginally
less code.
If not, let's not bloat it :).

[grt (UTC plus 2), 2012-07-10 12:13:48]

https://chromiumcodereview.appspot.com/10740003/diff/11001/chrome/app/breakpa...
File chrome/app/breakpad_win.cc (right):

https://chromiumcodereview.appspot.com/10740003/diff/11001/chrome/app/breakpa...
chrome/app/breakpad_win.cc:249: profile_type.assign(L"error ");
On 2012/07/10 10:11:14, Sigurður Ásgeirsson wrote:
> ubernit: I'd grab GLE to a temp at the very top of the file, just in case.

top of the file?  do you mean the top of this block, so that we're sure to grab
the value set by GetProfileType?

> Also, if StringPrintf is already used in setup.exe, that'd make for marginally
> less code.
> If not, let's not bloat it :).

i'll check...

[Sigurður Ásgeirsson, 2012-07-10 12:33:05]

On Tue, Jul 10, 2012 at 12:13 PM, <grt@chromium.org> wrote:

>
> https://chromiumcodereview.**appspot.com/10740003/diff/**
>
11001/chrome/app/breakpad_win.**cc<https://chromiumcodereview.appspot.com/10740003/diff/11001/chrome/app/breakpad_win.cc>
> File chrome/app/breakpad_win.cc (right):
>
> https://chromiumcodereview.**appspot.com/10740003/diff/**
>
11001/chrome/app/breakpad_win.**cc#newcode249<https://chromiumcodereview.appspot.com/10740003/diff/11001/chrome/app/breakpad_win.cc#newcode249>
> chrome/app/breakpad_win.cc:**249: profile_type.assign(L"error ");
> On 2012/07/10 10:11:14, Sigurður Ásgeirsson wrote:
>
>> ubernit: I'd grab GLE to a temp at the very top of the file, just in
>>
> case.
>
> top of the file?  do you mean the top of this block, so that we're sure
> to grab the value set by GetProfileType?

Sorry, yes.

>
>
>  Also, if StringPrintf is already used in setup.exe, that'd make for
>>
> marginally
>
>> less code.
>> If not, let's not bloat it :).
>>
>
> i'll check...
>
>
https://chromiumcodereview.**appspot.com/10740003/<https://chromiumcodereview...
>

[grt (UTC plus 2), 2012-07-10 15:15:44]

PTAL

[Sigurður Ásgeirsson, 2012-07-10 15:23:33]

lgtm

lgtm

[mmenke, 2012-07-10 21:38:24]

This CL seems to have broken the win build bots.  See
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win&number....
 Breakage was apparently between 145904 and 145913, and the error line sounds
right for this CL:

breakpad_win.obj : error LNK2019: unresolved external symbol
__imp__GetProfileType@4 referenced in function "class
std::basic_string<wchar_t,struct std::char_traits<wchar_t>,class
std::allocator<wchar_t> > __cdecl `anonymous namespace'::GetProfileType(void)"
(?GetProfileType@?A0x0aeceb31@@YA?AV?$basic_string@_WU?$char_traits@_W@std@@V?$allocator@_W@2@@std@@XZ)
..\build\Debug\chrome_app_unittests.exe : fatal error LNK1120: 1 unresolved
externals

[grt (UTC plus 2), 2012-07-10 21:53:26]

r145961 should fix it for good.  could you sync up to that rev and try?
 sorry for the hassle.



On Tue, Jul 10, 2012 at 5:38 PM, <mmenke@chromium.org> wrote:

> This CL seems to have broken the win build bots.  See
> http://build.chromium.org/p/**tryserver.chromium/**
>
buildstatus?builder=win&**number=18436<http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win&number=18436>
> .
>  Breakage was apparently between 145904 and 145913, and the error line
> sounds
> right for this CL:
>
> breakpad_win.obj : error LNK2019: unresolved external symbol
> __imp__GetProfileType@4 referenced in function "class
> std::basic_string<wchar_t,**struct std::char_traits<wchar_t>,**class
> std::allocator<wchar_t> > __cdecl `anonymous namespace'::GetProfileType(**
> void)"
> (?GetProfileType@?A0x0aeceb31@**@YA?AV?$basic_string@_WU?$**char_traits@
> _W@std@@V?$**allocator@_W@2@@std@@XZ)
> ..\build\Debug\chrome_app_**unittests.exe : fatal error LNK1120: 1
> unresolved
> externals
>
>
http://codereview.chromium.**org/10740003/<http://codereview.chromium.org/107...
>