cleanup: fix gm's use of ErrorBitfield to be consistent (it's a bitfield, not an enum)

cleanup: fix gm's use of ErrorBitfield to be consistent (it's a bitfield, not an enum)
Committed: https://code.google.com/p/skia/source/detail?r=8028

[epoger, 2013-03-07 19:51:03]

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp#newcode1
gm/gmmain.cpp:1: /*
Toplevel comment: The reason I'm doing this is to prepare for adding a flag to
return success even if there are pixel mismatches, as listed in
https://goto.google.com/ChecksumRemainingSteps .

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp#newcode87
gm/gmmain.cpp:87: const static ErrorBitfield kNoGpuContext_ErrorBitmask         
= 0x01;
While I was at it, changed the constant names to follow C++ style guide (instead
of looking like preprocessor macros)

[borenet, 2013-03-07 20:10:28]

The logic LGTM but I'm wondering about some of the variable names.

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp#newcode260
gm/gmmain.cpp:260: if (errorType == (errorType & kIgnorable_ErrorBitmask)) {
I like this, because it allows us to have multiple error types be ignorable.

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp#newcode614
gm/gmmain.cpp:614: retval = kReadingReferenceImage_ErrorBitmask;
So kReadingReferenceImage_ErrorBitmask occurs for either a missing baseline
image OR a missing expected checksum?  I think it's a little misleading to call
it an error reading a reference image, in that case.  In fact, "error reading
reference image" suggests to me that a reference image exists but could not be
read (maybe due to bad file format or something), but I think what this really
means is "no reference image exists."  Would "NoReferenceImage" or
"NoExpectations" be better?

[epoger, 2013-03-07 20:24:20]

PTAL at patchset 2

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp#newcode614
gm/gmmain.cpp:614: retval = kReadingReferenceImage_ErrorBitmask;
On 2013/03/07 20:10:28, borenet wrote:
> So kReadingReferenceImage_ErrorBitmask occurs for either a missing baseline
> image OR a missing expected checksum?  I think it's a little misleading to
call
> it an error reading a reference image, in that case.  In fact, "error reading
> reference image" suggests to me that a reference image exists but could not be
> read (maybe due to bad file format or something), but I think what this really
> means is "no reference image exists."  Would "NoReferenceImage" or
> "NoExpectations" be better?

Good thoughts, thanks... I changed it to kMissingExpectations_ErrorBitmask. 
Whaddayasay?

[borenet, 2013-03-07 20:27:44]

LGTMv2.0

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/12640004/diff/1/gm/gmmain.cpp#newcode614
gm/gmmain.cpp:614: retval = kReadingReferenceImage_ErrorBitmask;
On 2013/03/07 20:24:20, epoger wrote:
> On 2013/03/07 20:10:28, borenet wrote:
> > So kReadingReferenceImage_ErrorBitmask occurs for either a missing baseline
> > image OR a missing expected checksum?  I think it's a little misleading to
> call
> > it an error reading a reference image, in that case.  In fact, "error
reading
> > reference image" suggests to me that a reference image exists but could not
be
> > read (maybe due to bad file format or something), but I think what this
really
> > means is "no reference image exists."  Would "NoReferenceImage" or
> > "NoExpectations" be better?
> 
> Good thoughts, thanks... I changed it to kMissingExpectations_ErrorBitmask. 
> Whaddayasay?

LGTM!

[epoger, 2013-03-07 20:28:39]

Committed patchset #2 manually as r8028 (presubmit successful).