mailing-list for TeXmacs Users

[Lionel Elie Mamane <address@hidden>]

On Mon, Aug 07, 2006 at 12:20:58PM +0200, Henri Lesourd wrote:
> Lionel Elie Mamane wrote:

>>#16786 does not block development, but makes some documents using my
>>       plug-in look a bit ugly.

> I just had a look at your example document, when you say "TeXmacs indents
> this for no good reason" in the context of Lemma 1, I don't understand
> where is the superfluous indenting


> Where is the misunderstanding ? (you can have a look in the output
> of my own TeXmacs that I see myself, cf. pdf document attached to
> this email).

Your PDF shows the correct rendering.

TeXmacs 1.0.6 and an old CVS HEAD checkout identifying itself as
1.0.6.2 indents the whole fold that contains Lemma 1. Meaning that the
fold containing Lemma 1 is more to the right than the fold not
containing a lemma.

I now see that CVS HEAD as of a few days ago (that identifies itself
as 1.0.6.5) does not. CVS HEAD also shows the interactivity points
(where you double-click to fold/unfold) in black rather than blue.

I attach a PDF with the rendering of TeXmacs 1.0.6, so that you can
see what the problem was. I'm closing the bug, as it is has been fixed
since I filed it.

>>#16086: I can work around it, by taking care of always writing the
>>        whole contents of the scheme block before opening a new
>>        block, but this makes the other side (the process that talks
>>        to TeXmacs) more complex.

> If the function calls you send to your plugin are always
> synchronous, you could completely avoid relying on the TeXmacs
> socket interface by means of directly writing Scheme s-expressions
> to a Scheme socket.

I'm listening. Where is that documented? grepping for "socket"
case-insensitively in a fresh CVS HEAD checkout of the documentation
finds only a changelog entry:

 <item>Support for connections by sockets (1.0.2.1).

but no further documentation on how this works.

> This way, your code would probably become much simpler, and you
> would avoid depending on the possibly buggy (or possibly unfit for
> your purpose) TeXmacs escaping of the data you send.

Well, actually, being able to use "verbatim" data blocks was
convenient at the prototype stage of development; I'll have to stop
using it for other technical reasons anyway, though, so a clean way to
send s-expressions (only) interests me.

-- 
Lionel