IRC channel logs


back to list of logs

<VG234>OK. Back after being on here with a guile 1.8.8 problem with emacs and cygwin. I thought I would update for completeness. Earlier I had what turned out to be a guile configuration issue on cygwin. It appeared my install was incorrect. Cygwin doesn't have the latest guile so I downloaded and compiled natively. After installing some needed libraries which were noted in the readme, I got solid install. It did fail on the activate readline
<VG234>setting in my .guile config, but I'm guessing that is normal. So, bottom line, you can compile and run guile in cygwin64 on windows 7 64bit.
<VG234>Sorry, lost internet connection.
<jao>hi. is there a cond-expand feature differentiating 2.1 from 2.0?
<davexunit>jao: yes, just left a comment on github about it
<davexunit>since 2.1 is the unstable series for 2.2
<jao>davexunit, thanks!
<davexunit>as always, thanks for geiser!
<davexunit>ACTION is using geiser right now to develop SDL2 bindings
<jao>davexunit, yw. btw, is there also a guile-2.0? or is that guile-2?
<davexunit>jao: that I'm not sure of.
<jao>ACTION doesn't have a guile 2.1.x handy
<mark_weaver>jao: there's no guile-2.0, just guile-2
<mark_weaver>see %cond-expand-features in ice-9/boot-9.scm
<jao>and is guile-2 defined in 2.1/2.2?
<davexunit>jao: fwiw, guix provides a 'guile-next' package.
<mark_weaver>jao: yes
<jao>or are tehy mutually exclusive?
<mark_weaver>no, guile-2.2 will include both guile-2 and guile-2.2
<jao>oh, so i cannot cond-expand on "< 2.2"
<jao>well, no big deal
<mark_weaver>jao: cond-expand supports 'and', 'or', and 'not', so you have everything you need
<mark_weaver>see the 'srfi-0' docs in the guile manual or elsewhere
<jao>there's no feature marking guile-2.0 explicitly
<jao>so i don't have everything i would like
<mark_weaver>well, it's obviously too late to do anything about that
<mark_weaver>but out of curiosity, why would you want this?
<jao>(cond-expand (guile-2.0) (else ;; things for any guile > 2.0 ...
<davexunit>(and guile-2 (not guile-2.2)) ?
<mark_weaver>well, I don't know what the plan is for the 'guile-2.2' feature in guile-2.4
<jao>yeah, that's fine... i'll simply use 2.2 and revisit when there's a guile 2.4
<mark_weaver>jao: out of curiosity, what do you need to conditionalize based on 2.0 vs 2.2?
<jao>compilation to 'bytecode or 'objcode and load-thunk-from-memory vs. make-program
<mark_weaver>ah, yes
<mark_weaver>we probably need a better API to paper over those differences.
<davexunit>jao: how does geiser look up those scheme files?
<jao>davexunit, what scheme files?
<davexunit>I'm trying to get geiser to load your patched evaluation.scm file
<davexunit>without having to go through all the trouble of building a new geiser or restarting emacs
<jao>ah, M-x geiser-reload
<jao>well, yes, that'd work
<davexunit>I basicaly want to get it to use a git checkout of mine for that file
<davexunit>the version I use is installed via guix
<jao>simply restarting the repl too if you're using a checkout
<davexunit>I'm not actually using the elisp there
<davexunit>I was hoping that I could hack some elisp function to get geiser to read scheme code out of my git checkout
<jao>davexunit, there are only changes in evaluation.scm, so you need to reinstall that
<jao>no elisp involved
<davexunit>is there a variable that specifies where geiser should look for these files, though?
<jao>how are you loading geiser?
<davexunit>(require 'geiser-install)
<jao>(but i've never tested mixing things like that)
<davexunit>I'll see if it works
<davexunit>hmm didn't seem to work
<davexunit>testing this fix is difficult
<davexunit>jao: if I build geiser in a git checkout, could I just add that directory to the elisp load path?
<jao>what directory?
<davexunit>the source tree
<jao>no, not that easy
<davexunit>I want to start up an emacs instance without any of my normal config and just use this development version of geiser
<jao>why don't you just use the full git checkout?
<davexunit>that's what I'm trying to do
<jao>that's easy
<jao>(load-file "<path-to-geiser-checkout>/elisp/geiser.el")
<jao>that will setup everything, including the load path
<davexunit>okay, thanks
<davexunit>jao: unbound variable: bytcode?
<davexunit>and then after fixing the typo
<davexunit>Unbound variable: bytecode?
<jao>davexunit, just pushed a fix
<nalaginrut>morning guilers~
<davexunit>jao: it works!!!
<jao>davexunit, excellent! i can go to sleep now :)
<davexunit>me too ;)
<davexunit>this is exciting
<jao>thanks a lot for your help
<davexunit>you're welcome
<davexunit>thanks for the fix
<davexunit>now I can see how well my game engine works with guile 2.1
<jao>good luck!
<jao>nite all
<nalaginrut>I encountered terrible lib dependencies hell in my debian, then I want to try Guix, but now I realized that gcc can't be installed because of dependencies issue...
<nalaginrut>ACTION is reinstalling... so sad
<mark_weaver>nalaginrut: I recommend just using the binary install method.
<mark_weaver>if you use guix, you'll end up using our binaries anyway
<nalaginrut>mark_weaver: ok, I'll try
<nalaginrut>ACTION is considering to make sure Artanis support any plugin-able server, and write a strong server for Artanis first with libuv. In case delimited-continuation based server core hasn't optimizing enough for a productive need...
<dsmith-w`>Monday Greetings, Guilers
***dsmith-w` is now known as dsmith-work
<turbopape>Hi ! My PR on geiser was not accepted, as I break it for 2.0, but jaor used cond-expand to derive a solution for it :)
<turbopape>Good news, now geiser properly evaluates guile 2.1.1 :)
<davexunit>turbopape: yup!
<davexunit>very nice
<CustosLimen>so I was thinking, I can force spurious wakeups from pthread_cond_wait with an ld preload that implements it using pthread_cond_timedwait
<CustosLimen>then I can add test for this owner corruption to test suite