Re: Update on OT-SVG in FreeType

[Moazin K]

Hi Werner,

Thanks a lot for reviewing the MR.

https://gitlab.freedesktop.org/freetype/freetype/-/merge_requests/68

As you can see in the TODO list in the MR, one of the items is having official hooks and documentation on how to create them. I'm trying to write them in the repo below. Maybe ultimately we can host it inside the FreeType group of repositories.

https://gitlab.freedesktop.org/moazin/freetype-librsvg-based-ot-svg-hooks

There are one or two problems I discovered that affect the rendering of some glyphs in NotoColorEmojiSVG and I've opened up an issue in librsvg's repository. There is some more testing to be done, I suspect another problem with librsvg and Cairo. Anyways, those problems will take some time to fix and some more time to be solved in releases, but that shouldn't affect us as we don't have a default, yay! :D

Now I want to make it work with FreeType demo tools, I think the only one that makes sense for SVG glyphs would be ftview? Do we wanna keep the hooks in the demo tools repository directly embedded or maybe do a git submodule? The hooks will change from time to time as I discover these problems and attempt to solve them so maybe submodule makes more sense. What about compilation? Any flags/options to enable/disable OT-SVG glyphs or just enable it all the time in ftview?

Moazin

On Sun, Aug 1, 2021 at 1:55 PM Werner LEMBERG <wl@gnu.org> wrote:

> I've just submitted a draft MR that brings OT-SVG functionality to
> FreeType.  The code added allows client applications to grab the SVG
> document given a glyph ID as well as allows rendering it to a bitmap
> if renderer hooks are plugged-in externally.

Thanks a lot, it looks good to me.

> Also, does anyone have the time to update the internal zlib,
> pleeeeeease?

Hmm, you started that; see branch `origin/internal-zlib-update`.
What's the state of this code?

> Werner, do you think we should have two configuration flags, first one to
> enable the OT-SVG document loading and a second one for the rendering part?
> Or do you think one flag for everything suffices?

A single flag is enough, I believe.


    Werner