Prevent DM crash in the PDF sink when src width/height are zero.

Prevent DM crash in the PDF sink when src width/height are zero.

Committed: https://skia.googlesource.com/skia/+/2ab9000b3fa1a61ec9213408af9c800f70eb0657

[djsollen, 2015-04-03 13:31:12]

djsollen@google.com changed reviewers:
+ halcanary@google.com, mtklein@google.com, scroggo@google.com

[hal.canary, 2015-04-03 13:32:08]

lgtm

[hal.canary, 2015-04-03 13:32:12]

The CQ bit was checked by halcanary@google.com

[commit-bot: I haz the power, 2015-04-03 13:32:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1061493002/1

[djsollen, 2015-04-03 13:35:01]

https://codereview.chromium.org/1061493002/diff/1/dm/DMSrcSink.cpp
File dm/DMSrcSink.cpp (right):

https://codereview.chromium.org/1061493002/diff/1/dm/DMSrcSink.cpp#newcode365
dm/DMSrcSink.cpp:365: return "SkDocument::beginPage(w,h) returned NULL";
I think it may be better if we could just produce an empty pdf which would map
to what other backends do.

[mtklein, 2015-04-03 13:37:12]

On 2015/04/03 13:35:01, djsollen wrote:
> https://codereview.chromium.org/1061493002/diff/1/dm/DMSrcSink.cpp
> File dm/DMSrcSink.cpp (right):
> 
> https://codereview.chromium.org/1061493002/diff/1/dm/DMSrcSink.cpp#newcode365
> dm/DMSrcSink.cpp:365: return "SkDocument::beginPage(w,h) returned NULL";
> I think it may be better if we could just produce an empty pdf which would map
> to what other backends do.

We're not going to want to look at that empty pdf, but we don't want to count
this as an error either.  So, generate a non-fatal error?

[commit-bot: I haz the power, 2015-04-03 13:38:35]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/skia/+/2ab9000b3fa1a61ec9213408af9c800f70eb0657

[scroggo, 2015-04-03 14:20:53]

On 2015/04/03 13:37:12, mtklein wrote:
> On 2015/04/03 13:35:01, djsollen wrote:
> > https://codereview.chromium.org/1061493002/diff/1/dm/DMSrcSink.cpp
> > File dm/DMSrcSink.cpp (right):
> > 
> >
https://codereview.chromium.org/1061493002/diff/1/dm/DMSrcSink.cpp#newcode365
> > dm/DMSrcSink.cpp:365: return "SkDocument::beginPage(w,h) returned NULL";
> > I think it may be better if we could just produce an empty pdf which would
map
> > to what other backends do.
> 
> We're not going to want to look at that empty pdf, but we don't want to count
> this as an error either.  So, generate a non-fatal error?

Isn't it an error though? It's not a PDF error, but it seems like an error if we
failed to decode. (Or if we decoded a 0, 0 image and thought it was a success?)
Perhaps we should catch the error before here?

I think this is at least a good place to start though, since it keeps us from
crashing...