Re: Chart test failures on mac: text size differences

[Tamas Zolnai]

Hi Khaled,

On Sunday, February 26, 2017 23:30 GMT, Khaled Hosny <khaledhosny at eglug.org> wrote: 
 
> On Sat, Feb 25, 2017 at 09:58:58PM +0000, Tamas Zolnai wrote:
> > Hi Khaled,
> > 
> > > > However I think it would be better to debug this problem with a unit
> > > > test in vcl module which tests directly the text rendering. I would
> > > > suggest that if you are working making text rendering consitent. The
> > > > vcldemo does something similar, but it's just doing the rendering and
> > > > does not compare the text size to expected values. That code can be
> > > > reused to write some unit tests about text rendering.
> > >
> > > If someone knows how to write text layout unit tests for VCL (i.e. by
> > > showing some example) I might try adding some, otherwise I don’t think
> > > OutputDevice provides much of useful API for doing layout tests that
> > 
> > > would be helpful at VCL level.
> > 
> > Is vcldemo code not good enough as example? (vcl/workben/vcldemo.cxx)
> > Check the drawText() or drawComplex() method. As I see this method
> > uses only vcl objects (OutputDevice, vcl::Font).
> > OutputDevice::GetTextBoundRect() can be used to get the text width and
> > height. Or did I miss something?
> 
> I don’t think text width or height are that interesting, but they are
> probably a good approximation, so I repurposed one of the existing VCL
> tests to test this[1]. I tried testing also the text bounding rectangle
> but it showed a 1 pixel difference on Mac, this should be fixed in [2].
> 
> Feel free to add more test cases that you think are worth having.
> 
> 1. https://gerrit.libreoffice.org/gitweb?p=core.git;a=commitdiff;h=eb7b03b052ffe8c2c577b2349987653db6c53f76
> 2. https://gerrit.libreoffice.org/gitweb?p=core.git;a=commitdiff;h=af871d02914c9fc6a08079f67b4af71a198e166a
> 
> Regards,
> Khaled
> 
> P.S. The bounding rectangle is the ink extents of the text, i.e. the
> smallest rectangle that fits around the text without leaving any white
> space around it, which is not the same as text width and height. This
> should’t normally have any effect on text layout, as it should depend
> only on logical extents (the width and height). We still use platform
> APIs to determine each glyph bounding rectangle which is a potential
> source of differences here. We probably should try to use HarfBuzz here
> as well.
 
Hmm, I see. It seems the problem causing chart test failures is not on VCL level. This small difference not seems to be the one we looking for and also the log files (what you attached earlier) show that the text height is different not the bounding rectangle. Well, I'm not familiar with this text layouting stuff, so I don't know what other levels it has. Anyway it's good that now we have a test for VCL level.

Thanks,
Tamás