<rekado>lechner: do you mean “--base=auto”? <lechner>rekado: Hi, yes! Although i ended up using --base=master. In view of your email about -mentors I wonder if you know anyone who might help with this proposed patch, please http://ix.io/3Z9E <lechner>Hi, that one was for Linux PAM. For anyone wishing to help, why wait for -mentors? Here is another upgrade, for nyacc http://ix.io/3Z9K <lechner>Hi, I created 'handle' via 'scm_from-pointer()' in C. How can I use that value in Guile with a foreign-library-function that expects a POINTER_P, please? I also tried (scm->pointer ...) http://ix.io/3Z9X <lechner>Hi, I used the seemingly somewhat internal 'scm_mkfromstrs' to create an SCM list of SCM strings from the standard C combination of (int argc, char **argc). Is there a function that can do the reverse? <rekado>lechner: have you sent these to guix-patches yet? ***Sauvin is now known as Furor
<lechner>rekado: thanks! i did now. they are #55762 and #55763, respectively <lechner>Hi, is this a way to construct and argv-type array of null-terminated strings for consumption by a foreign library function? (argv (scm->pointer (list->array 1 (map string->pointer options)))) <lloda>idk what argv is there you probably want (scm->pointer (array-ref (list->array 1 ...) 0)) <lloda>ie pointer to the first element of the array <lloda>(scm->pointer array) will give you a pointer to the array struct which is a bunch of stuff you don't want <lloda>also it'd be easier to use a vector here ***chris is now known as Guest5860
<lechner>Hi, why does this give ice-9/eval.scm:159:9: Wrong type to apply: "bingo", please? (map string->pointer ("bingo" "bongo")) <singpolyma>("bingo" "bongo") means to call the string "bingo" as a function with argument "bongo" <singpolyma>You probably want (list "bingo" "bongo") or the similar '("bingo" "bongo") <lechner>singpolyma: thanks so much! i am still struggling with the basics <lechner>Hi, is "literal" in the Guile docs a synonym for "quoted"? <lechner>Hi, this segfaults for argc > 0, presumably because argv points to an unacceptable place (or to unacceptable pointers). What am I doing wrong, please? http://ix.io/3Zc0 ***Guest5860 is now known as chrislck
<lloda>maybe your vector is being collected, not sure <mwette>lechner: wrt http://ix.io/3Zc0, I think you need to collect the string->pointers into a bytevector (based on the size of your machine) and then pass from bytevector->pointer <lechner>mwette: also, thanks again for the quick fix for linux-pam! now compile-ffi warns for pamc (in the compile phase) ;;; <unknown-location>: warning: possibly unbound variable `pamc_handle_t*' the typedef in C is somewhat unusual in that it includes the asterisk. is that something to worry about? <lechner>Hi, does Guile provide a helper functionality for build systems to determine the location of the local site directory? <daviid>lechner: mwette is right, you should look at what guile-gi and g-golf do for those situation - also, you need to keep a reference to the bytevector pointer for the lifetime of the calle <lechner>Hi, how may i use the "interface" returned by resolve-interface, please? I get Unbound variable: service:identify (with a prefix) <lechner>daviid: my code does not fork. what is the lifetime of a dlopen'ed, shared object, please? I do not think it stores any state <daviid>lechner: i don't know, my advice is, first, make it work, if it needs to survive thelifetime of the caller, store the pointer in a global, a hash table for example ... out of curioisity, what is the lib you are binding? <lechner>daviid: PAM modules. i'll circulate a draft once i have basic functionality <lechner>daviid: thanks for that definition of %gi-pointer-size ! <daviid>lechner: why would service-interface be #f, and in the if, it should be (if service-interface ...), not (if (service-interface <daviid>*not (if (service-interface) ...) <daviid>i have little time to help, sorry <cwebber>ok, this isn't quite a guile question, it's an autotools question, but it's for a guile project and I'm not sure where's better to ask <cwebber>I'd like to add something to my .scm file that's like <cwebber>if this field is true according to autoconf, expand this section of code in the .scm.in <daviid>lechner: you need to use emacs and indent your code better, use geiser will also grseatly help you developing, finally, please use a paste service that renders scheme code and has line numbers <cwebber>and I'm not quite sure where to figure out how to do this, even though I know it's probablyb very basic m4 stuff <cwebber>specifically, the configure.ac already has: <cwebber>AC_CHECK_FUNC([inotify_init], [AM_CONDITIONAL(HAVE_INOTIFY, true)], [AM_CONDITIONAL(HAVE_INOTIFY, false)]) <cwebber>so i'd like to see if I can expand conditionally in the .scm.in depending on what the value of HAVE_INOTIFY is <lechner>cwebber: make sure the .in is in AC_CONFIG_FILES <cwebber>lechner: thanks, I'll give it a shot... <lechner>daviid: i do use emacs, just in my own way <daviid>lechner: i don't have time further help now, hope others will shimm in ... good luck <daviid>mwette: I wonder how difficult it would be, maybe you didthink about that, writing a GObject Introspection 'nyacc/ffi-gi-helper' tool/app that would use your tremendous C skills, nyacc, the ffi-helper ... and write GObject Introspectable typelib, from a C lib/API? they are not here anymore, but str1ng did a minimal typelib for Qt, that they could use using g-golf ... just a thought <daviid>cwebber: i would (let ((have-inotify @HAVE_INOTIFY@)) ... in your scm.in file, but if the user removes the lib ... you know i guess <cwebber>daviid: the problem is that it's either importing or not importing things depending on @HAVE_INOTIFY@ <daviid>cwebber: but at autotool expansion, have-notify will be bound to the value 'true' or 'false', which youcn use to condition your runtime scheme code, did i miss something maybe? <daviid>what it right to write 'will be bound'? i wasn't sure... <daviid>*was it right to write... damn, my kb-fu is fuzzy today :) <apteryx>is it possible to rename an autoloaded symbol? <apteryx>also, is there a way to mask autoload warnings (when the library is not found?) <mwette>daviid: haha, been thinking of cleaning up ffi=helper to generated other back ends. That would need to happen first. (then the FH could generate C api code also. <daviid>mwette: not sure how difficult it would be, to acheive good practical/working results i mean, but surely not an easy thing, the GI team itself never wrote a complete GI typelib for cairo for example, not sure why ...