gm: write all messages to stdout/stderr with "GM:" preamble to distinguish from various debug messa…

gm: write all messages to stdout/stderr with "GM:" preamble to distinguish from various debug messages coming from elsewhere
Committed: https://code.google.com/p/skia/source/detail?r=8126

[epoger, 2013-03-12 20:15:55]

https://codereview.chromium.org/12691009/diff/3002/gm/tests/outputs/compared-...
File
gm/tests/outputs/compared-against-different-pixels-images/output-expected/stderr
(right):

https://codereview.chromium.org/12691009/diff/3002/gm/tests/outputs/compared-...
gm/tests/outputs/compared-against-different-pixels-images/output-expected/stderr:1:
GM: ---- 8888/selftest1: 60000 (of 60000) differing pixels, max per-channel
mismatch R=0 G=111 B=103 A=0
The point of this CL is to make sure gm self-test covers any changes to lines
like these in the output.

[borenet, 2013-03-12 20:19:06]

On 2013/03/12 20:15:55, epoger wrote:
>
https://codereview.chromium.org/12691009/diff/3002/gm/tests/outputs/compared-...
> File
>
gm/tests/outputs/compared-against-different-pixels-images/output-expected/stderr
> (right):
> 
>
https://codereview.chromium.org/12691009/diff/3002/gm/tests/outputs/compared-...
>
gm/tests/outputs/compared-against-different-pixels-images/output-expected/stderr:1:
> GM: ---- 8888/selftest1: 60000 (of 60000) differing pixels, max per-channel
> mismatch R=0 G=111 B=103 A=0
> The point of this CL is to make sure gm self-test covers any changes to lines
> like these in the output.

The idea LGTM but I wish we could wrap those calls into a GmDebug() or some
similar function.  I foresee the preamble being forgotten by other devs, or
future selves.

[epoger, 2013-03-12 20:26:12]

On 2013/03/12 20:19:06, borenet wrote:
> On 2013/03/12 20:15:55, epoger wrote:
> >
>
https://codereview.chromium.org/12691009/diff/3002/gm/tests/outputs/compared-...
> > File
> >
>
gm/tests/outputs/compared-against-different-pixels-images/output-expected/stderr
> > (right):
> > 
> >
>
https://codereview.chromium.org/12691009/diff/3002/gm/tests/outputs/compared-...
> >
>
gm/tests/outputs/compared-against-different-pixels-images/output-expected/stderr:1:
> > GM: ---- 8888/selftest1: 60000 (of 60000) differing pixels, max per-channel
> > mismatch R=0 G=111 B=103 A=0
> > The point of this CL is to make sure gm self-test covers any changes to
lines
> > like these in the output.
> 
> The idea LGTM but I wish we could wrap those calls into a GmDebug() or some
> similar function.  I foresee the preamble being forgotten by other devs, or
> future selves.

Agreed.  I attempted that first, but I couldn't get it to work with the
variable-length arguments passed into fprintf()... any ideas?

[borenet, 2013-03-12 20:41:02]

On 2013/03/12 20:26:12, epoger wrote:
> On 2013/03/12 20:19:06, borenet wrote:
> > On 2013/03/12 20:15:55, epoger wrote:
> > >
> >
>
https://codereview.chromium.org/12691009/diff/3002/gm/tests/outputs/compared-...
> > > File
> > >
> >
>
gm/tests/outputs/compared-against-different-pixels-images/output-expected/stderr
> > > (right):
> > > 
> > >
> >
>
https://codereview.chromium.org/12691009/diff/3002/gm/tests/outputs/compared-...
> > >
> >
>
gm/tests/outputs/compared-against-different-pixels-images/output-expected/stderr:1:
> > > GM: ---- 8888/selftest1: 60000 (of 60000) differing pixels, max
per-channel
> > > mismatch R=0 G=111 B=103 A=0
> > > The point of this CL is to make sure gm self-test covers any changes to
> lines
> > > like these in the output.
> > 
> > The idea LGTM but I wish we could wrap those calls into a GmDebug() or some
> > similar function.  I foresee the preamble being forgotten by other devs, or
> > future selves.
> 
> Agreed.  I attempted that first, but I couldn't get it to work with the
> variable-length arguments passed into fprintf()... any ideas?

I had to deal with this to create SkDebugf for NaCl.  Take a look at va_list et
al: http://skia.googlecode.com/svn/trunk/src/ports/SkDebug_nacl.cpp

[epoger, 2013-03-13 01:21:34]

On 2013/03/12 20:41:02, borenet wrote:
> > Agreed.  I attempted that first, but I couldn't get it to work with the
> > variable-length arguments passed into fprintf()... any ideas?
> 
> I had to deal with this to create SkDebugf for NaCl.  Take a look at va_list
et
> al: http://skia.googlecode.com/svn/trunk/src/ports/SkDebug_nacl.cpp

Thanks for the tip!  That was very helpful... incorporated as patchset 6.

[epoger, 2013-03-13 01:21:40]

https://codereview.chromium.org/12691009/diff/27001/gm/gm_expectations.h
File gm/gm_expectations.h (right):

https://codereview.chromium.org/12691009/diff/27001/gm/gm_expectations.h#newc...
gm/gm_expectations.h:57: static void gm_fprintf(FILE *stream, const char
format[], ...) {
I originally tried to declare this in gm.h, but then I got warnings about the
function being declared but not used in lots of .cpp files.

[borenet, 2013-03-13 11:58:41]

LGTM

https://codereview.chromium.org/12691009/diff/27001/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/12691009/diff/27001/gm/gmmain.cpp#newcode274
gm/gmmain.cpp:274: gm_fprintf(stderr, "\t\t%s pixel_error\n",
fFailedTests[i].fName.c_str());
Thanks for directing these to stderr.

[epoger, 2013-03-13 14:18:48]

Committed patchset #6 manually as r8126 (presubmit successful).