Last year I experimented a lot with a Guile 2 port. As soon as you have some time you could give it a look and we could try to use it as a blueprint for a Guile 3 port. Maybe the JIT will make it possible to disable compilation of the files (which is where I currently experience most of the bugs) since eval in Guile 3 is as fast as in Guile 1.8 (as they say). Of course compilation (if we can make it work) will make bootstrap faster, which would be a great plus.
Max > On 26. Jan 2020, at 09:49, TeXmacs <texm...@lix.polytechnique.fr> wrote: > > On Sat, Jan 25, 2020 at 11:45:08AM +0100, Massimiliano Gubinelli wrote: >> It would be good to have an automatic test-suite to evaluate seriously the >> performance issues involved in this choice, also in the perspective to >> understand what would be the margin of advantage to adapt the codebase to >> Guile 3. >> >> Do you have some ideas of which areas involve Guile performances seriously? > > Since we use Guile as an extension language, > we tend to switch to C++ whenever performance is an issue. > In the past, I rewrote some document search routines in C++ for this reason. > > The only serious problem that I had with Guile was the time to boot all files. > This prompted me to write all the lazy initialization stuff... > and that is precisely what is getting into our way if we now > want to move to Guile 2/3... > > Best wishes, --Joris > > _______________________________________________ > Texmacs-dev mailing list > Texmacs-dev@gnu.org > https://lists.gnu.org/mailman/listinfo/texmacs-dev _______________________________________________ Texmacs-dev mailing list Texmacs-dev@gnu.org https://lists.gnu.org/mailman/listinfo/texmacs-dev
