[Windows] Include the about:flags experiments in crash reports.

[Windows] Include the about:flags experiments in crash reports.

  * Increases the number of command line switches included in dumps from 2 to 15.
  * Saves the command line flags for all process types (before it was just for browser).
  * Includes the "fake" command line flags that were added by about:flags experiments.

This change will make it possible to cluster crashes with enabled experiments from our server-side analysis.

BUG=60992
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=101009

[eroman, 2011-09-13 05:31:35]

Changing reviewer from thakis to cpu.

[cpu_(ooo_6.6-7.5), 2011-09-13 22:09:59]

lgtm

http://codereview.chromium.org/7866033/diff/12002/chrome/app/breakpad_win.cc
File chrome/app/breakpad_win.cc (right):

http://codereview.chromium.org/7866033/diff/12002/chrome/app/breakpad_win.cc#...
chrome/app/breakpad_win.cc:115: CommandLine::ForCurrentProcess()->argv();
todo (eroman) : filter useless switches that confuse the TPMs and bloat the
report.

[eroman, 2011-09-13 22:29:56]

Thanks for the review!

http://codereview.chromium.org/7866033/diff/12002/chrome/app/breakpad_win.cc
File chrome/app/breakpad_win.cc (right):

http://codereview.chromium.org/7866033/diff/12002/chrome/app/breakpad_win.cc#...
chrome/app/breakpad_win.cc:115: CommandLine::ForCurrentProcess()->argv();
On 2011/09/13 22:09:59, cpu wrote:
> todo (eroman) : filter useless switches that confuse the TPMs and bloat the
> report.

Done.

[Nico, 2011-09-13 23:04:41]

LGTM!

http://codereview.chromium.org/7866033/diff/6003/chrome/app/breakpad_win.cc
File chrome/app/breakpad_win.cc (right):

http://codereview.chromium.org/7866033/diff/6003/chrome/app/breakpad_win.cc#n...
chrome/app/breakpad_win.cc:122: for (; argv_i < argv.size() && num_added <
kMaxSwitches; ++argv_i) {
Increment num_added in the for loop instead of in the loop body

http://codereview.chromium.org/7866033/diff/6003/chrome/common/child_process_...
File chrome/common/child_process_logging_linux.cc (right):

http://codereview.chromium.org/7866033/diff/6003/chrome/common/child_process_...
chrome/common/child_process_logging_linux.cc:82: // TODO: http://crbug.com/60993
NOTIMPLEMENTED()?

http://codereview.chromium.org/7866033/diff/6003/chrome/common/child_process_...
File chrome/common/child_process_logging_win.cc (right):

http://codereview.chromium.org/7866033/diff/6003/chrome/common/child_process_...
chrome/common/child_process_logging_win.cc:156: if (!exe_module)
DCHECK?

[eroman, 2011-09-13 23:27:32]

FYI: I have made a minor change to the parameters as, shown here:

http://codereview.chromium.org/7866033/diff2/12002:15011/chrome/app/breakpad_...

The problem was I had been testing this using the multi-DLL build. In the
regular build on the other hand, the breakpad code (chrome.exe) has a separate
global CommandLine from that seen by chrome.dll. Passing in the pointer should
avoid this problem.

Once I verify that this works, I will commit.

http://codereview.chromium.org/7866033/diff/6003/chrome/app/breakpad_win.cc
File chrome/app/breakpad_win.cc (right):

http://codereview.chromium.org/7866033/diff/6003/chrome/app/breakpad_win.cc#n...
chrome/app/breakpad_win.cc:122: for (; argv_i < argv.size() && num_added <
kMaxSwitches; ++argv_i) {
On 2011/09/13 23:04:41, Nico wrote:
> Increment num_added in the for loop instead of in the loop body

Done.

The reason I was doing it in the body rather than the for loop was to make it
easy to skip over adding certain flags. When I address the TODO, I will move the
increment back out of the for loop.

http://codereview.chromium.org/7866033/diff/6003/chrome/common/child_process_...
File chrome/common/child_process_logging_linux.cc (right):

http://codereview.chromium.org/7866033/diff/6003/chrome/common/child_process_...
chrome/common/child_process_logging_linux.cc:82: // TODO: http://crbug.com/60993
On 2011/09/13 23:04:41, Nico wrote:
> NOTIMPLEMENTED()?

Done.

http://codereview.chromium.org/7866033/diff/6003/chrome/common/child_process_...
File chrome/common/child_process_logging_win.cc (right):

http://codereview.chromium.org/7866033/diff/6003/chrome/common/child_process_...
chrome/common/child_process_logging_win.cc:156: if (!exe_module)
On 2011/09/13 23:04:41, Nico wrote:
> DCHECK?

I just copy-pasted this from the other functions in this file. Not familiar
enough to know if DCHECK is correct.