IRC channel logs
2023-01-17.log
back to list of logs
<rekado>count3rmeasure: yes, it was a reply to your “style question”. <count3rmeasure>thank you, the two were related, and I was a bit distracted, hence the follow up rekado <sneek>I've been running for 14 days <sneek>This system has been up 11 weeks, 5 days, 22 hours, 1 minute <old>dsmith: Thank yout. :config and :init are not working .. I will try with :custom <chrislck>old: agree - it doesn't test that error-type matches <mirai>is there some srfi magic I'm missing? <flatwhatson>mirai: that looks fine to me, not aware of any srfi for a cartesian product <old>hey wingo, you take patches for guile-charting? <old>Got a couple of fixes here, but I don't know if the project is still alive <antipode>sneek: later tell count3rmeasure: If you switch from SRFI records to RNRS records, you can accomplish this by setting the 'protocol' field (without %make-thing wrappers) <antipode>sneek: later tell apteryx: You could delay expansion by using 'local-eval', though it will get re-expanded each time, so not great for performance. <apteryx>why is this not valid: source expression failed to match any pattern in form (condition &platform-not-found-error), where &platform-not-found-error is defined via: (define-condition-type &platform-not-found-error &error platform-not-found-error?) <mirai>what are quasipatterns in ice-9 match? is there any example of it in use? <civodul>apteryx: the syntax is like: (condition (&message (message "foo"))) <civodul>so (condition (&platform-not-found-error)) i guess <civodul>you might want to have a field with the missing platform too <apteryx>ah, so my exception type is a constructor too <lloda>civodul: i pushed the bytevector r7rs doc patch <apteryx>the platform is the object of the query, so it is unknown :-) <apteryx>civodul: but I can add a target-or-system field to the condition type <lloda>i'd like to rescue the array-mutable? patch from last year. Having xxx-immutable-copy needs to be done for each of the array types however, some sort of already have it in C, like string, none has it Scheme side. It'd be weird if you needed to use the array interface to make an immutable root <lloda>an immutable string/vector/bytevector/bitvector i mean <lloda>the main motivation for this is being able to pass an output array to an ffi. You need to be able to check that the array is mutable or it will bomb in the external library <civodul>one can't "officially" create read-only things though <civodul>put differently, it's not part of the API <lloda>tbh i'm not sure creating immutable things is that useful. The example given by Maxime in guile-devel was of testing array-mutable?, because you can have literals that are mutable, with eval <dokma>How long would it take to go through guile source without going into much detail? <civodul>ACTION pushed the linker/assembler changes <civodul>i think i could tag & upload 3.0.9rc1 maybe tomorrow so we get some more testing <count3rmeasure>I'm interested civodul, trying to get involved in dev work and testing releases is an easy way to do that <sneek>Welcome back count3rmeasure, you have 1 message! <sneek>count3rmeasure, antipode says: If you switch from SRFI records to RNRS records, you can accomplish this by setting the 'protocol' field (without %make-thing wrappers) <mfiano>sneek: Later tell civodul: That would be great. I'm also trying to get involved; I am subscribed to the ML and will report any issues I might find with rc1 when you tag it. <daviid>sneek: later tell civodul i've just checked g-golf against the very latest main branch (last commit being the NEWS update - 9a004606e067e5af68d1656ad6553c7ae353e81d) and verything works perfectly, the test-suite and all examples all 'pass' ... <lloda>not sure where the & is coming from <mwette>lloda: It it possible to "make CFLAGS+=-E libguile/posix.o" and look the posix.o to see the CPP expansion ? <apteryx>what is the most convenient way to create write a binary sequence made of integers and strings to a file? <lilyp>apteryx: converting everything to bytevectors and then writing those? <apteryx>OK; is it possilbe to concat the vectors together before writing, or is doing multiple writes easier? <apteryx>mwette: using core guile i think I can work from uint8 lists and then u8-list->bytevector at the end <old>mwette: what's that guile-contrib you have? <mwette>pack/unpack like in python struct module <mwette>I need to doc more but there is a test file w/ examples <old>Is it just scratch stuff or you plan to make it available on guix? <mwette>I think just scratch atm. for guix, means guix.scm file? I'm not a guixer atm. <civodul>apteryx: (guix build syscalls) internally has define-c-struct, which could be useful or an inspiration <sneek>Welcome back civodul, you have 2 messages! <sneek>civodul, mfiano says: That would be great. I'm also trying to get involved; I am subscribed to the ML and will report any issues I might find with rc1 when you tag it. <sneek>civodul, daviid says: i've just checked g-golf against the very latest main branch (last commit being the NEWS update - 9a004606e067e5af68d1656ad6553c7ae353e81d) and verything works perfectly, the test-suite and all examples all 'pass' ... <civodul>mfiano, daviid: that was fast, thank you!!