IRC channel logs
2024-07-08.log
back to list of logs
<vagrantc>janneke: there are very few changes with nyacc 1.09.4 -> 1.09.5 and it seems to build fine with mes 0.27 on amd64 and i386 on debian ... just wondering if there is anything in particular i should watch out for? <snuik>vagrantc, you have 1 message <snuik>vagrantc, janneke says: oops! and thanks for testing; i'm testing a couple of patches on the version-0.26 branch right now <vagrantc>i might just push the newer nyacc ... worst case is some hideous versions for a spell <vagrantc>Googulator: no idea... newer mes 0.27 claims to be compatible with newer versions of nyacc ... just not the very latest version of nyacc <Googulator>"compatible" as in "can use it with the pregenerated parsers", or "can bootstrappably regenerate the parsers"? <Googulator>...without relying on psyntax-pp, on top of that <vagrantc>i am not the person to answer that ... it builds fine on debian :) <janneke>vagrantc: i didn't see 1.09.5, so that may well work <vagrantc>two successful build tests on two architectures ... what could possibly go wrong? :) <janneke>+ * fixed bug 65849: crash on multiple #warning <janneke>so the main feature fixes a bug i reported on .4, yay :) <janneke>[or some gramatically correct version of that comment] <janneke>*grammatically -- phew, time for some rest <janneke>having multiple "#warning ..."s work is actually pretty nice for debugging cpp flags <vagrantc>Googulator: i guess ... check the new upstream version for related commits? <vagrantc>Googulator: ah, now i see the confusion. :) <janneke>iwbn if samplet would upstream that patch (except for the define-module bits which belong in mes i guess) <Googulator>upstreaming a fix into 1.09.x or 1.10.x when mes depends on no newer than 1.00.x is less useful <Googulator>assuming that Matthew Wette isn't as hostile to pregen-free bootstrapping as certain other maintainers (I'm looking at you, Bruno Haible...) <janneke>mwette has been quite accommodating in the past <janneke>ACTION has added the missing mes bits (define F_OK), (system foreign) to mes' wip branch <janneke>ah, and mes should also provide vector-fold and string-every <vagrantc>+ /<<PKGBUILDDIR>>/bin/mes --no-auto-compile -e main -L /usr/share/guile/site/3.0 -C /usr/lib/guile/3.0/site-ccache /<<PKGBUILDDIR>> <vagrantc>/scripts/mescc.scm -- -m 32 -c -D HAVE_CONFIG_H=1 -I ../lib -I ../include -I ../include -I include -v -static -nostdinc -nostdlib -f <vagrantc>/<<PKGBUILDDIR>>/scripts/mescc: line 69: 6902 Segmentation fault ${SCHEME-$MES} --no-auto-compile -e main -L /usr/share/guile/ <vagrantc>site/3.0 -C /usr/lib/guile/3.0/site-ccache $bindir/mescc.scm $sep "$@" <vagrantc>also when i just try to build on mescc-gcc (e.g. not using ./configure --with-bootstrap) <stikonas>Googulator: yeah, but we could try to check if those nyacc patches work with newer nyacc and if yes, then try to upstream them <stikonas>oh yes, both of you already mentioned upstreaming too <stikonas>Googulator: anyway, first I'm testing upgrade to mes 0.27 without nyacc changes <janneke>Googulator: i believe that was already added in v0.27 <stikonas>suprisingly few changed hashes for mes upgrade... <stikonas>but I guess this time libc didn't chagne