IRC channel logs


back to list of logs

<lechner>Hi, how do I interpret this error from 'guild compile-ffi' please?
<drakonis>it seems you didnt pass a pair?
***taiju` is now known as taiju
<lechner>drakonis: i just ran the command on these three files. it failed on the last one
<lechner>maybe this odd inline comment, or the continuation lines confused nyacc?
<lechner>#define PAM_BPC_FOR_CLIENT(/* pamc_bp_t */ prompt)
<mwette>lechner: probably c preprocessor issue, can you file a bug report?
<mwette>much of the challenge in coding FFI helper has been the C preprocessor
<lechner>mwette: wow, having reached the author i'm in the right place. thanks so much for the great FFI module! will do on the bug
<lagash>lechner: it's not Guile (it's CL), but if you like "consfiguring" you should check out Consfigurator :-)
***Guest8848 is now known as roptat
***ecraven- is now known as ecraven
<mwette>lechner: I think I fixed: one liner:
<lampilelo>mwette strikes again with quick fixes
<lechner>mwette: looks great! will test and get back to you.
<lechner>mwette: on a sidenote, why does the order of the header files here matter, please? symbols like PAM_SUCESS from _pam_types.h are not included when that file is listed last
***chris is now known as Guest3456
<mwette>lechner: maybe not the way it's intended to be done (in pam); if pam usually wants you to include one file, but you want to suck in the dep's then it's better to use #:include on the intended ones and use a #:inc-filter to suck in the files under security/. See some examples in examples/ffi for how to use.
<mwette>FYI, nyacc-1.07.0 was released today
<civodul>mwette: congrats!
<stis>mwette: \o/
<mwette>took some time; I've been working microcontrollers and also guile debugging
<mwette>There is a lot there (traps, dwarf support, etc) but getting a full debugger will take some time: need to understand what's there and then adding on top
<lechner>mwette: Hi, did your one-liner make it into 1.07? I only saw it on the dev branch
<lechner>maybe i looked in the wrong branch. did you switch from 'master' to 'main' recently?
<lechner>mwette: thanks for the quick fix, and also for the hint about #:inc-filter !
***yewscion is now known as Guest7325
***yewscion20 is now known as yewscion
<lechner>Hi, do i have to do anything other than (use-modules (ffi pam)) to get access to PAM_SUCCESS, please?
<lechner>I also tried (ffi-pam-symbol-val PAM_SUCCESS)
<mwette>(ffi-pam-symbol-val 'PAM_SUCCESS)
<mwette>the one liner is in 1.07
<mwette> yes in main -- I still have to merge into master
<mwette>merged to master now, commig 88da7e
<lechner>mwette: thanks! sorry, i'm a Guile noobie but with great enthusiasm. why do you have master and main? btw, i now name all my development trunks "history"
<mwette>master is deprecated (e.g., at, etc); the guile repo is using main as well
<mwette>for git, master is the old convention, main is the new convention
<Noisytoot>No, git still uses master as the default branch, it's just GitHub (and some other forges) that use main by default.
<mwette>Oh. Thanks for the clarification.
<lechner>Hi, is there a state monad in Guile?
<mwette>monads in guix:
<lechner>mwette: thanks!
<lechner>They monads look great! Why are they only in Guix?
<lilyp>Because Guix foud them useful :)
<lechner>lilyp: Thanks! Do they bring a specific advantage to Guix?
<lilyp>Well, in some aspects Guix is more purely functional than generic Scheme stuff, but I don't want to start a lengthy debate at this hour.
*lilyp → bed
<lechner>lilyp: Thanks!
<lechner>Hi, is the shorted 'define' syntax for procedures widely accepted, or do some folks still prefer the long form with lambda? In other words, is it pedantic to like the latter? Thanks!
<rekado>a good reason to use the long form is when doing memoization or including private state via closure
<mwette>Yes. Shortened form is widely used. I use the long form for private var's , like rekado says. (define foo (let ((counter 0)) (lambda () (set! counter (+1 counter)) counter)))