add interface for WriteMinidump which allows the caller to supply file handles instead of paths whe…

add interface for WriteMinidump which allows the caller to supply file handles
instead of paths where the minidumps should be written.

BUG=N/A
TEST=N/A
R=ivan.penkov@gmail.com, mark@chromium.org

Committed: https://code.google.com/p/google-breakpad/source/detail?r=1191

[Ivan Penkov]

.

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
File client/windows/crash_generation/minidump_generator.cc (right):

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
client/windows/crash_generation/minidump_generator.cc:347: full_memory_dump ?
full_dump_file : NULL);
Why not just pass in full_dump_file handle?  Besides, for invalid handles you
should use INVALID_HANDLE_VALUE insetad of NULL.

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
client/windows/crash_generation/minidump_generator.cc:375: if (!dump_file ||
dump_file == INVALID_HANDLE_VALUE) {
Why check for NULL?  The check for INVALID_HANDLE_VALUE should be sufficient.

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
client/windows/crash_generation/minidump_generator.cc:378: 
You need to also check the validity of the second handle when (dump_type &
MiniDumpWithFullMemory) != 0

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
File client/windows/crash_generation/minidump_generator.h (right):

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
client/windows/crash_generation/minidump_generator.h:82: bool
WriteMinidump(HANDLE process_handle,
Also add a comment that the handles are not closed upon function exit.

[Cris Neckar]

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
File client/windows/crash_generation/minidump_generator.cc (right):

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
client/windows/crash_generation/minidump_generator.cc:347: full_memory_dump ?
full_dump_file : NULL);
On 2013/06/04 23:21:49, Ivan Penkov wrote:
> Why not just pass in full_dump_file handle?  Besides, for invalid handles you
> should use INVALID_HANDLE_VALUE insetad of NULL.

Done.

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
client/windows/crash_generation/minidump_generator.cc:375: if (!dump_file ||
dump_file == INVALID_HANDLE_VALUE) {
On 2013/06/04 23:21:49, Ivan Penkov wrote:
> Why check for NULL?  The check for INVALID_HANDLE_VALUE should be sufficient.

Done.

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
client/windows/crash_generation/minidump_generator.cc:375: if (!dump_file ||
dump_file == INVALID_HANDLE_VALUE) {
On 2013/06/04 23:21:49, Ivan Penkov wrote:
> Why check for NULL?  The check for INVALID_HANDLE_VALUE should be sufficient.

Done.

https://breakpad.appspot.com/602002/diff/1/client/windows/crash_generation/mi...
client/windows/crash_generation/minidump_generator.cc:378: 
On 2013/06/04 23:21:49, Ivan Penkov wrote:
> You need to also check the validity of the second handle when (dump_type &
> MiniDumpWithFullMemory) != 0

Done.

[Ivan Penkov]

LGTM

[Mark Mentovai]

LGTM

[Cris Neckar]

Committed patchset #2 manually as r1191 (presubmit successful).