> On 17. Jan 2022, at 23:28, [email protected] <[email protected]>
> wrote:
>
>> Do I understand you correclty that the patch I
>> implemented break some assumption of your code
>
> In general, yes. I'd have to read through the
> code -- I glanced at it a few months ago, but
> can't remember any details, except that it struck
> me as written for Guile 1.8? and perhaps
> awkward for s7 (something about modules?).
>
Yes, the current code was written for Guile 1.8 and we use some module stuff,
which I simulated via lets. (It is some time I wrote that code too, so now I
forgot the fine details). We have a large module which contains all the user
accessible procedures and we have a system to overload them (tm-define, you can
look it up in the file ./TeXmacs/progs/kernel/texmacs/tm-define.scm). My guess
is that this module is getting populated with a lot of symbols and this slows
down S7 without my patch. But frankly I do not understand well the logic of
lookup in S7. Would you mind to give us a (maybe rapid) overview of how lookup
is implemented? I would really like to have some "correct" solution which make
S7 works with TeXmacs. We are currently evaluating seriously to use it in the
future but we would like to have to iron out these problems (lookup & memory
footprint) which even if small produce large effects in our usecase.
I guess the problem with lookup is localized in this only large module that we
have. What I prefer not do is to change all our implementation because of it.
Ideally I would like to keep the code "implementation agnostic" as it is now,
and provide a compatibility layer to adapt to various scheme implementations.
This is currently the case, we have implementations working with Guile 1.6,
1.8, 3.0 and S7.
Best
Max
_______________________________________________
Cmdist mailing list
[email protected]
https://cm-mail.stanford.edu/mailman/listinfo/cmdist
