Spawn a new process to register and unregister DLLs.

Spawn a new process to register and unregister DLLs.

In component=shared_library builds, it isn't safe to load npchrome_frame.dll into a process that has its own AtExit manager, singletons, logging, etc. So now spin off a new run of the given test executable to do the registration.

BUG=110492
TEST=none (covered by cf_win trybot)

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=119049

[grt (UTC plus 2), 2012-01-24 21:44:32]

I just updated some error handling.  Please refresh if you've started your
review.  Thanks.

[robertshield, 2012-01-24 22:40:01]

lg, nits

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
File chrome_frame/test_utils.cc (right):

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:129: << "). Are you running as Admin?";
"Are you" -> "Make sure"

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:171: int argc;
 = 0

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:173: if (argc != 4 || ::lstrcmp(argv[1],
kCallRegistrationEntrypointSwitch) != 0)
I wonder if this should printf() a usage string and exit if the flag is passed
but argc != 4?

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:186: &entrypoint[0]);
Maybe a brief comment explaining that this does the character conversion, I had
to double check that was the case.

[grt (UTC plus 2), 2012-01-25 02:46:36]

Thanks for the suggestions.  PTAL.

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
File chrome_frame/test_utils.cc (right):

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:129: << "). Are you running as Admin?";
On 2012/01/24 22:40:01, robertshield wrote:
> "Are you" -> "Make sure"

Done.

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:171: int argc;
On 2012/01/24 22:40:01, robertshield wrote:
>  = 0

Done.

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:173: if (argc != 4 || ::lstrcmp(argv[1],
kCallRegistrationEntrypointSwitch) != 0)
On 2012/01/24 22:40:01, robertshield wrote:
> I wonder if this should printf() a usage string and exit if the flag is passed
> but argc != 4?

Done.

https://chromiumcodereview.appspot.com/9146054/diff/2001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:186: &entrypoint[0]);
On 2012/01/24 22:40:01, robertshield wrote:
> Maybe a brief comment explaining that this does the character conversion, I
had
> to double check that was the case.

Done.

[robertshield, 2012-01-25 03:07:40]

lgtm

https://chromiumcodereview.appspot.com/9146054/diff/6001/chrome_frame/test_ut...
File chrome_frame/test_utils.cc (right):

https://chromiumcodereview.appspot.com/9146054/diff/6001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:176: printf("Usage: %S %S <path to dll>
<entrypoint>\n", argv[0], argv[1]);
nit: argv[1] -> kCallRegistrationEntrypointSwitch.

[grt (UTC plus 2), 2012-01-25 03:44:11]

Thanks.

https://chromiumcodereview.appspot.com/9146054/diff/6001/chrome_frame/test_ut...
File chrome_frame/test_utils.cc (right):

https://chromiumcodereview.appspot.com/9146054/diff/6001/chrome_frame/test_ut...
chrome_frame/test_utils.cc:176: printf("Usage: %S %S <path to dll>
<entrypoint>\n", argv[0], argv[1]);
On 2012/01/25 03:07:41, robertshield wrote:
> nit: argv[1] -> kCallRegistrationEntrypointSwitch.

Okay.

[commit-bot: I haz the power, 2012-01-25 04:44:51]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/grt@chromium.org/9146054/9001

[commit-bot: I haz the power, 2012-01-25 04:44:55]

Presubmit check for 9146054-9001 failed and returned exit status 1.

Running presubmit commit checks ...

** Presubmit Warnings **
New code should not use wstrings.  If you are calling an API that accepts a
wstring, fix the API.
    chrome_frame/test_utils.cc:140

Presubmit checks took 1.1s to calculate.