Re: Possible Pango 1.4 ideas

[Owen Taylor <otaylor redhat com>]

<dleach qnx com> writes:

>  > Hi,
> > 
> > I noticed the discussions on gtk-devel about GTK+ 2.4 plans, so I
> > wondered if we should try to get some better layout features into Pango
> > for the next release. Quite a few people seem interested in using Pango
> > for DTP or printing now!
> [snip]
> 
> 
> I got another request re: the shapers.  Does anyone think it would
> be a _good_ _thing_ (tm) to remove all the GUI/render specific code
> from the individual shapers by abstracting them out further, through
> maybe a callback mechanism (for glyph fetch, index fetch, mapping request, etc)?
> It would make porting the shapers easier, at least for me ;).

Shapers are, by design, specific to a particular font technology. As
it happens, some of the backends share font technology ... in
particular the Xft and FT2 backends are very similar. In these cases,
we should certainly be looking at ways to reduce duplicated code.

But there isn't much code that could be shared between, say, Xft, 
core X, and the Uniscribe-based Win32 backend.

Frequently there are also some things (such as tables of linguistic
information) that are shared between backends for different font
technology.

But I think a hope of only implement a few language-independent
callbacks and have all the languages "just work" is pretty much doomed
to disappointment. Unless, perhaps, the backend is similar enough to
Xft/fontconfig to fit into that framework.

(That would mean, at a minimum using OpenType fonts for complex
scripts, and either using FreeType or having a compatible version of
the stuff in the pango/opentype directory.)

Regards,
                                        Owen