Some plumbing for the beginning of getting handles into snapshot/minidump

Some plumbing for the beginning of getting handles into snapshot/minidump

Follows https://codereview.chromium.org/1400413002/.

R=mark@chromium.org
BUG=crashpad:21, crashpad:46, crashpad:52
Committed: https://chromium.googlesource.com/crashpad/crashpad/+/4600643a78e8080713bfef6b12b7ab7b6956c6bf

[scottmg, 2015-10-14 22:59:05]

Patchset #8 (id:140001) has been deleted

[scottmg, 2015-10-14 22:59:15]

Patchset #7 (id:120001) has been deleted

[Mark Mentovai, 2015-10-14 23:55:13]

https://codereview.chromium.org/1407643004/diff/160001/compat/non_win/dbghelp.h
File compat/non_win/dbghelp.h (right):

https://codereview.chromium.org/1407643004/diff/160001/compat/non_win/dbghelp...
compat/non_win/dbghelp.h:641: //!     operating system. On Windows, this
corresponds to `OBJ_INHERIT`,
You got this documentation verbatim from Microsoft, but the meaning of
“operating system” to them is entirely different than its meaning in Crashpad,
which runs on non-Microsoft operating systems.

Your “on Windows” addition is then a little weird, because HANDLEs aren’t a
thing on non-Windows at all, and we wouldn’t see a MINIDUMP_HANDLE_DESCRIPTOR. I
think that Microsoft meant what we’d call the OS version.

Same for the remainder of the fields in this structure.

https://codereview.chromium.org/1407643004/diff/160001/snapshot/handle_snapsh...
File snapshot/handle_snapshot.cc (right):

https://codereview.chromium.org/1407643004/diff/160001/snapshot/handle_snapsh...
snapshot/handle_snapshot.cc:21: handle(),
0 for integer fields. Basically the same comments that I made in the place that
this moved from.

https://codereview.chromium.org/1407643004/diff/160001/snapshot/handle_snapsh...
File snapshot/handle_snapshot.h (right):

https://codereview.chromium.org/1407643004/diff/160001/snapshot/handle_snapsh...
snapshot/handle_snapshot.h:38: //! \brief The ACCESS_MASK for the handle in this
process. Ref:
No Ref: or URL in the \brief.

https://codereview.chromium.org/1407643004/diff/160001/snapshot/test/test_pro...
File snapshot/test/test_process_snapshot.cc (right):

https://codereview.chromium.org/1407643004/diff/160001/snapshot/test/test_pro...
snapshot/test/test_process_snapshot.cc:112: CHECK(false);
Would be easy to implement, but if there’s no use for it and you’re not going
to, you should say so in a comment.

[scottmg, 2015-10-15 00:26:05]

Patchset #8 (id:180001) has been deleted

[scottmg, 2015-10-16 20:44:40]

Patchset #8 (id:200001) has been deleted

[scottmg, 2015-10-16 20:50:17]

Patchset #8 (id:220001) has been deleted

[scottmg, 2015-10-16 20:51:02]

https://codereview.chromium.org/1407643004/diff/160001/compat/non_win/dbghelp.h
File compat/non_win/dbghelp.h (right):

https://codereview.chromium.org/1407643004/diff/160001/compat/non_win/dbghelp...
compat/non_win/dbghelp.h:641: //!     operating system. On Windows, this
corresponds to `OBJ_INHERIT`,
On 2015/10/14 23:55:13, Mark Mentovai wrote:
> You got this documentation verbatim from Microsoft, but the meaning of
> “operating system” to them is entirely different than its meaning in Crashpad,
> which runs on non-Microsoft operating systems.
> 
> Your “on Windows” addition is then a little weird, because HANDLEs aren’t a
> thing on non-Windows at all, and we wouldn’t see a MINIDUMP_HANDLE_DESCRIPTOR.
I
> think that Microsoft meant what we’d call the OS version.
> 
> Same for the remainder of the fields in this structure.

Done.

https://codereview.chromium.org/1407643004/diff/160001/snapshot/handle_snapsh...
File snapshot/handle_snapshot.cc (right):

https://codereview.chromium.org/1407643004/diff/160001/snapshot/handle_snapsh...
snapshot/handle_snapshot.cc:21: handle(),
On 2015/10/14 23:55:13, Mark Mentovai wrote:
> 0 for integer fields. Basically the same comments that I made in the place
that
> this moved from.

Done.

https://codereview.chromium.org/1407643004/diff/160001/snapshot/handle_snapsh...
File snapshot/handle_snapshot.h (right):

https://codereview.chromium.org/1407643004/diff/160001/snapshot/handle_snapsh...
snapshot/handle_snapshot.h:38: //! \brief The ACCESS_MASK for the handle in this
process. Ref:
On 2015/10/14 23:55:13, Mark Mentovai wrote:
> No Ref: or URL in the \brief.

Done.

https://codereview.chromium.org/1407643004/diff/160001/snapshot/test/test_pro...
File snapshot/test/test_process_snapshot.cc (right):

https://codereview.chromium.org/1407643004/diff/160001/snapshot/test/test_pro...
snapshot/test/test_process_snapshot.cc:112: CHECK(false);
On 2015/10/14 23:55:13, Mark Mentovai wrote:
> Would be easy to implement, but if there’s no use for it and you’re not going
> to, you should say so in a comment.

Implemented, I was just planning to do it in a follow up once there's actually
test code calling it.

[Mark Mentovai, 2015-10-16 22:39:24]

LGTM

https://codereview.chromium.org/1407643004/diff/240001/util/win/process_info.h
File util/win/process_info.h (right):

https://codereview.chromium.org/1407643004/diff/240001/util/win/process_info....
util/win/process_info.h:172: // Handles() is logically const, but updates this
member on first retrieval.
Reference https://code.google.com/p/crashpad/issues/detail?id=9. This const
problem is getting out of hand, we should fix it.

[scottmg, 2015-10-16 22:52:19]

https://codereview.chromium.org/1407643004/diff/240001/util/win/process_info.h
File util/win/process_info.h (right):

https://codereview.chromium.org/1407643004/diff/240001/util/win/process_info....
util/win/process_info.h:172: // Handles() is logically const, but updates this
member on first retrieval.
On 2015/10/16 22:39:24, Mark Mentovai wrote:
> Reference https://code.google.com/p/crashpad/issues/detail?id=9. This const
> problem is getting out of hand, we should fix it.

Done.

It does seem sort of "right" for Snapshot to be all const, but in practice
laziness always seems to eventually make const interfaces troublesome.

[scottmg, 2015-10-16 22:57:47]

Patchset #10 (id:280001) has been deleted

[scottmg, 2015-10-16 22:58:30]

Patchset #10 (id:300001) has been deleted

[scottmg, 2015-10-16 22:58:46]

Committed patchset #9 (id:260001) manually as
4600643a78e8080713bfef6b12b7ab7b6956c6bf (presubmit successful).