Chronological Thread 

< Chronological >      < Thread >

---

On Sat, 2006-01-21 at 19:24 +0100, Joris van der Hoeven wrote:
> On Sat, Jan 21, 2006 at 06:06:16PM +0100, Ingolf Schäfer wrote:
> > I think it is definitely possible and should be considerded, but I would
> > suggest to work in a different direction. I think we should not try to
> > port to QT/WX /GTK/... but we should aim to use cairo instead of direct
> > X11 calls. This would make TeXmacs easier to port and would also be an
> > enrichment to TeXmacs.
>
> For the new graphics mode, Henri and I are indeed considering changing
> the underlying libraries. Cairo and OpenGL are both good candidates.
> We might also want to opt for an abstract layer (like the current
> ps_device) with different implementations.

I would be thrilled to see that.

Out of curiosity, did you investigate Pango? I had a very superficial
look at it years ago, and it looked like it had a reasonable font model.
Better integration with of the operating system font services would be
conducive to more correct PDF generation.

I used TeXmacs recently to produce some documentation, and was
disappointed of seeing that the PDF output was still inferior to the PS
output. The fonts were not jagged (anymore), but characters did not
align properly on the baseline.

> Notice however that the graphical routines are probably the easiest part,
> because of the current ps_device abstraction. More complex is the event
> handling, which is also based on X, currently. Moreover, this is getting
> more and more complex. For instance, the next version should come with
> support for Japanese, Chinese and maybe Korean. However, our current
> implementation heavily depends on the X input method.

Nice to hear that this important project is under way.

I hope that would lead to a reorganisation of the keyboard input
handling. Last time I looked at it, it was dysfunctional: it was
confused by the presence of extraneous modifiers, such as Mod4
(associated to Numlock on some systems).

If the existing pattern matching approach is to be retained, it would
probably be necessary to represent keys as structured objects and match
against those, instead of using pure text-based pattern matching.

Regards.
--
-- ddaa

Attachment: signature.asc
Description: This is a digitally signed message part

---

• Cairo, OpenGL, and CJK input methods, David Allouche, 04/18/2006

---

Archive powered by MHonArc 2.6.19.