RE: Flipped Arrow repair reason \ CustomShapes

Hi Miklos
Can you provide more information about the 2nd part of your mail ?
Why do you connect the unit-test part to the CustomShapeGeometry?
Any detail would be appreciated...

       Adam Fyne
        Office: +972-77-517-5008
       Twitter | LinkedIn | Facebook | Blog

-----Original Message-----
From: Miklos Vajna [mailto:vmiklos@suse.cz]
Sent: Wednesday, August 7, 2013 1:39 PM
To: Adam Fyne
Cc: libreoffice@lists.freedesktop.org; tbehrens@suse.com
Subject: Re: Flipped Arrow repair reason \ CustomShapes

On Wed, Aug 07, 2013 at 10:58:32AM +0300, Adam Fyne <Adam.Fyne@cloudon.com>
wrote:
> I have asked our developer that made the fix to explain why he did it,
> and he sent me the attached explanation.

Just a suggestion: it's great if developers doing the real work have their
name attached to patches, instead of you being a proxy. One benefit of that
is if you cause a regression, we won't blame you, but the one who in fact
introduced it. ;-)

> I would be happy to convey him any rejects you have about what he said.

So, it seems that the way the mirrored x/y property can be read via
CustomShapeGeomery (as done in the testcase part of the patch) should work
for writing as well: you found that it's not working, but that's a bug. It
would be excellent if that could be fixed, and then that API would be used
from the VML import filter.

In the end that would result in a consistent API which we all want in the
long run.

Does that make sense?

Thanks,

Miklos