Send uptime as milliseconds for Mac and iOS.

Send uptime as milliseconds for Mac and iOS.

 All other platform are sending uptime as milliseconds. Changing the
implementation to do the same on Mac and iOS.
Committed: https://code.google.com/p/google-breakpad/source/detail?r=928

[Mark Mentovai]

LG otherwise.

https://breakpad.appspot.com/355001/diff/1/src/client/ios/Breakpad.h
File src/client/ios/Breakpad.h (right):

https://breakpad.appspot.com/355001/diff/1/src/client/ios/Breakpad.h#newcode140
src/client/ios/Breakpad.h:140: // BREAKPAD_PROCESS_UP_TIME          The total
time in ms the process has been
Spell out “milliseconds.”

https://breakpad.appspot.com/355001/diff/1/src/client/mac/Framework/Breakpad.h
File src/client/mac/Framework/Breakpad.h (right):

https://breakpad.appspot.com/355001/diff/1/src/client/mac/Framework/Breakpad....
src/client/mac/Framework/Breakpad.h:195: // BREAKPAD_PROCESS_UP_TIME         
The total time in ms the process has been
Same.

https://breakpad.appspot.com/355001/diff/1/src/client/mac/crash_generation/Co...
File src/client/mac/crash_generation/ConfigFile.mm (right):

https://breakpad.appspot.com/355001/diff/1/src/client/mac/crash_generation/Co...
src/client/mac/crash_generation/ConfigFile.mm:120: sprintf(processUptimeString,
"%zd", 1000 * processUptime);
Doing this in 32-bit time_t-land opens up the possibility for overflow in as
little as 24 days. That’s not actually unreasonable for a process uptime. You
should force this to the 64-bit domain.

[Mark Mentovai]

LG otherwise.

[qsr]

https://breakpad.appspot.com/355001/diff/1/src/client/ios/Breakpad.h
File src/client/ios/Breakpad.h (right):

https://breakpad.appspot.com/355001/diff/1/src/client/ios/Breakpad.h#newcode140
src/client/ios/Breakpad.h:140: // BREAKPAD_PROCESS_UP_TIME          The total
time in ms the process has been
On 2012/03/01 19:56:54, Mark Mentovai wrote:
> Spell out “milliseconds.”

Done.

https://breakpad.appspot.com/355001/diff/1/src/client/mac/Framework/Breakpad.h
File src/client/mac/Framework/Breakpad.h (right):

https://breakpad.appspot.com/355001/diff/1/src/client/mac/Framework/Breakpad....
src/client/mac/Framework/Breakpad.h:195: // BREAKPAD_PROCESS_UP_TIME         
The total time in ms the process has been
On 2012/03/01 19:56:54, Mark Mentovai wrote:
> Same.

Done.

https://breakpad.appspot.com/355001/diff/1/src/client/mac/crash_generation/Co...
File src/client/mac/crash_generation/ConfigFile.mm (right):

https://breakpad.appspot.com/355001/diff/1/src/client/mac/crash_generation/Co...
src/client/mac/crash_generation/ConfigFile.mm:120: sprintf(processUptimeString,
"%zd", 1000 * processUptime);
On 2012/03/01 19:56:54, Mark Mentovai wrote:
> Doing this in 32-bit time_t-land opens up the possibility for overflow in as
> little as 24 days. That’s not actually unreasonable for a process uptime. You
> should force this to the 64-bit domain.

Done.

[Mark Mentovai]

LGTM, but…

https://breakpad.appspot.com/355001/diff/4002/src/client/mac/crash_generation...
File src/client/mac/crash_generation/ConfigFile.mm (right):

https://breakpad.appspot.com/355001/diff/4002/src/client/mac/crash_generation...
src/client/mac/crash_generation/ConfigFile.mm:121: sprintf(processUptimeString,
"%llu",
snprintf with arraysize(processUptimeString) would be safer, although there’s
currently no way for this to overflow that array.

https://breakpad.appspot.com/355001/diff/4002/src/client/mac/crash_generation...
src/client/mac/crash_generation/ConfigFile.mm:127:
sprintf(processCrashtimeString, "%zd", tv.tv_sec);
Here too.