Ekaitz Zarraga 👹
@civodul @janneke @wingo One of the most limiting things I found is the context switches between C and Scheme, that sometimes can trigger some `eval` that makes another context switch.
Also, moving a few things (like arithmetic operators) to opcodes makes everything WAAAAY faster, but that disallows making things like (define (+ . x) whatever) if it's not done carefully. Do you have any trick for this?
Janneke
@ekaitz_zarraga @civodul @wingo
You and Timothy mentioned a very specific semantic aspect of the Guile dialect, IIRC in the context of modules and possibly macro expansion?
Janneke
@ekaitz_zarraga @civodul @wingo
Right. So I figured, either it would be great for Mes and Germ to learn from Guile how to do that efficiently (not very likely), or it would be good to know that the Guile developers are aware of this and could possibly even consider a semantic change to also improve Guile's performance.
No idea what the impact of that change would be though.
Ekaitz Zarraga 👹
@janneke @civodul @wingo Yes, the module system is one of the things that produces most of this Scheme->C->Scheme dance and also guile having a different behavior when it's AOT compiled vs when it's interpreting makes everything very hard: It forces us to expand+compile+evaluate top-level expressions one by one, with all the problems that has.
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.