DM: add pdf

DM: add pdf

BUG=skia:2598
Committed: https://skia.googlesource.com/skia/+/30bf3e2ffcb78fc76e5a62b2ca67638e0411cba9

[hal.canary, 2014-06-03 20:21:29]

https://codereview.chromium.org/312873002/diff/70001/gyp/dm.gyp
File gyp/dm.gyp (left):

https://codereview.chromium.org/312873002/diff/70001/gyp/dm.gyp#oldcode67
gyp/dm.gyp:67: '-lcutils',
where did cutils go?

[mtklein, 2014-06-03 20:27:57]

https://codereview.chromium.org/312873002/diff/70001/gyp/dm.gyp
File gyp/dm.gyp (left):

https://codereview.chromium.org/312873002/diff/70001/gyp/dm.gyp#oldcode67
gyp/dm.gyp:67: '-lcutils',
On 2014/06/03 20:21:29, Hal Canary wrote:
> where did cutils go?

Oh, I think we don't actually need that anymore.  Unrelated. I'll remove it in
another CL.

[hal.canary, 2014-06-03 20:34:28]

lgtm

[mtklein, 2014-06-03 20:35:44]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-06-03 20:36:09]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/312873002/80001

[vandebo (ex-Chrome), 2014-06-03 20:39:57]

https://codereview.chromium.org/312873002/diff/80001/dm/DMPDFTask.cpp
File dm/DMPDFTask.cpp (right):

https://codereview.chromium.org/312873002/diff/80001/dm/DMPDFTask.cpp#newcode60
dm/DMPDFTask.cpp:60: //pdf.canvas()->concat(fGM->getInitialTransform());
Because it passes it into the PDF device:
https://code.google.com/p/chromium/codesearch/#chromium/src/third_party/skia/...

[mtklein, 2014-06-03 20:44:31]

On 2014/06/03 20:39:57, vandebo wrote:
> https://codereview.chromium.org/312873002/diff/80001/dm/DMPDFTask.cpp
> File dm/DMPDFTask.cpp (right):
> 
>
https://codereview.chromium.org/312873002/diff/80001/dm/DMPDFTask.cpp#newcode60
> dm/DMPDFTask.cpp:60: //pdf.canvas()->concat(fGM->getInitialTransform());
> Because it passes it into the PDF device:
>
https://code.google.com/p/chromium/codesearch/#chromium/src/third_party/skia/...

Is there a difference between drawing it before the GM starts and passing it
into the device?

[vandebo (ex-Chrome), 2014-06-03 20:55:38]

On 2014/06/03 20:44:31, mtklein wrote:
> On 2014/06/03 20:39:57, vandebo wrote:
> > https://codereview.chromium.org/312873002/diff/80001/dm/DMPDFTask.cpp
> > File dm/DMPDFTask.cpp (right):
> > 
> >
>
https://codereview.chromium.org/312873002/diff/80001/dm/DMPDFTask.cpp#newcode60
> > dm/DMPDFTask.cpp:60: //pdf.canvas()->concat(fGM->getInitialTransform());
> > Because it passes it into the PDF device:
> >
>
https://code.google.com/p/chromium/codesearch/#chromium/src/third_party/skia/...
> 
> Is there a difference between drawing it before the GM starts and passing it
> into the device?

Yes...

I don't remember all the details and I haven't checked recently, but when the
PDF code was first developed WebKit would consult the device.width() and
height() directly and as a result they need to report the size of the content
area and not the size of the virtual page.  In order to support margins, the
initial transform was added; it isn't detectable once set.  Whereas a setting a
transform on the canvas is visible/detectable. There's one gm that uses the
getInitialTransform() mechanism and it's a regression test.

It's not a desirable feature of the PDF code, just a necessary one.  Once
WebKit/Blink gets fixed up, it should be removed.

[commit-bot: I haz the power, 2014-06-03 20:57:17]

Change committed as 30bf3e2ffcb78fc76e5a62b2ca67638e0411cba9