<lilyp>Zelphir: you could version string-utils, but then it'll complain that the one loaded has the wrong version <lilyp>alternatively, use GUILE_LOAD_PATH to your advantage, where you make the string-utils you want show up earlier <Zelphir>lilyp: So if the one I am using directly in my project is loaded before the other one, would it load the string-utils of the package later and thus have all required functions in it? <Zelphir>And how would it work, if the functions in those libraries were disjunct? <apteryx>is there anything I can do to determine what kind of data is held by a Guile process memory? I have a Guile program holding 10 GiB of resident memory, I'd like to know why <spk121>apteryx: there is the 'gc-stats' procedure, but, that doesn't have much detail <apteryx>can I call this on a running process (it's PID 1 ^^') <spk121>apteryx: well you can add a repl to any program and then use that repl to call gc-stats <spk121>That adds a repl to a program that is allocating GC'able blocks. You can log into that repl with 'nc localhost 8080' and then call gc-stats there <spk121>There's probably some way of using environment variables with BDW GC, but, I don't know offhand. <lilyp>apteryx: "sudo herd eval root '(gc-stats)'" ***NullPointerErro1 is now known as NullPointerError
***taiju` is now known as taiju
<rekado>Hi Guilers, I just learned that I cannot tell Guile’s chown to operate on the symlink instead of its target. <rekado>is there a way to accomplish this some other way? <rekado>I only want to chown a symlink and keep the ownership of the target unaltered <unmatched-paren>i have no idea whether guile has an equivalent, but it's worth a try ***Furor is now known as Colere
<rekado>I’m trying to do that in a tiny Guile script; defining FFI stuff is not worth the effort in this case. <rekado>uid_t and gid_t are not defined; I guess I could just use unsigned ints there <rekado>I’d love to see lchown (and Guix’s mkdir-p) in Guile proper. ***daviwil_ is now known as daviwil
<mwette>^that's just the first few lines provided to nyacc's ffi-helper <unmatched-paren>rekado: Hmm, I just realized that lchmod is a little inefficient as it re-binds `proc` every time it's called <antipode>rekado: I sent a patch for various things like lchown (not sure if it actually includes lchown in particular) some months ago, there has been a little progress in getting it merged.