<amz3>did anyone look at my rambling on the guile-user mailling lsiit? <amz3>I was thinking about something else; I will post something else <amz3>I did 4 posts alreay on the smallest month of the year, nobody will bit me this year :) <amz3>even if there are mmonologues <spauldo>just spent way longer than I'd like to admit trying to debug a function that had (+1 i) in it <janneke>it's not at all strange what you write <lloda`>wingo: libguile/continuations.c:186, is that frame->stack_top ?? <wingo>btw i ran some tests on my 2-socket E5-2620v3 system <wingo>6 cores/socket, hyperthreading off. <wingo>for the fibers web server it gets around 10K reqs/s for a hello-world app on one core <wingo>so still some scaling work to be done <wingo>and also some single-core perf to work on too <wingo>lloda`: heh i meant to remove an assigned-but-not-used variable but it seems i removed the decl but not the assignment :P <civodul>wingo: i think i'll cut 2.0.14 today or tomorrow just so we get those fixes out (repro builds in particular) <wingo>civodul: sgtm; afaiu we are still missing the forward port of that patch, right? <civodul>wingo: yes, it's still on my to-do list <civodul>merging stable-2.0 in master and adjusting things <civodul>is there something particular to pay attention to when merging stable-2.0 in master BTW? <wingo>i think at this point forward cherry-picks are the right thing; afair anyway <wingo>maybe we can still merge, not sure <civodul>well there aren't so many commits anyway <civodul>i'll try merge and cherry-pick if it doesn't work <wingo>i think i've done cherry-picks most recently so there might be conflicts <lloda`>is NEWS the place to keep deprecation notices? I miss some. <lloda`>wingo: I ask b/c e.g. uniform-array-read!/write were deprecated in 2.0 but there's no specific deprecation notice, just 'this interface has been deprecated'. I think a notice like 'x is gone, use y instead' is important to help people port. <lloda`>now they are gone and there's no mention of them anywhere. <lloda`>I can add a post-facto notice in NEWS after '** All deprecated code removed', would that be ok? <wingo>i think that notice is there <lloda`>it doesn't mention those specific functions and their replacements <lloda`>I'll put up a patch with what I have in mind <wingo>ah sorry i didn't understand your proposal. sgtm <ArneBab_>wingo: after paroneayea asked that on FOSDEM, is there a chance of getting wisp included in Guile? <ArneBab_>and if yes, what would I have to do for that? <wingo>ArneBab_: what would that look like? :) <wingo>also why would you want that? in some ways it can be easier to maintain something outside of guile <ArneBab_>wingo: it would mean that people could run wisp code just like they can run elisp-code and ecma-code <wingo>and if there is a wisp feature you need to rely on, are you ok with waiting some time for a guile release? <wingo>or telling your users to use guile from git? <civodul>i'm not sure this is a good idea, though i remember paroneayea had almost convinced during their talk <ArneBab_>wisp is pretty stable these days. I only experimented with one change in the past year, and I think it would create more problems than it solves. <ArneBab_>wingo: I’m ok with waiting for Guile releases. <ArneBab_>all the stuff I’ve been doing is just exploring what can be done with the wisp-syntax defined in SRFI-119 <paroneayea>I think it would be nice to have wisp in guile proper <paroneayea>wisp's primary feature is that it provides a nice, non-parenthetical syntax which can make some things easier to write (game scripts, etc) but *especially* it makes overcoming the fear of parenthesis easier I think <paroneayea>that, and as far as languages bundled with Guile, which presumably we want more of <paroneayea>wingo: ArneBab_: maybe we should make a branch, wip-wisp, on savannah? <paroneayea>ArneBab_: I guess you should request commit access then, if you don't have it? <paroneayea>ArneBab_: I'd be interested in testing and doing some light review of the branch as it comes along, but of course I will not be the *best* reviewer <paroneayea>because I don't know much about guile internals or compilers <paroneayea>ArneBab_: personally I would recommend that even though the wisp repo has the "wisp is bootstrapped from lisp but then is written in lisp" thing, I would just write it in scheme for guile proper <ArneBab_>paroneayea: I’d just add the files parsed to scheme <paroneayea>ArneBab_: sorry, I didn't quite catch what you meant by that? <paroneayea>ArneBab_: btw what do you think of the extension .wsp or .wisp? <paroneayea>ha, .wsp means Windows Sharepoint in some places tho :) <ArneBab_>personally I like .w (for the simplicity), but I’m not attached enough to it that I’d argument about that. <paroneayea>C-u M-x debbugs-gnu ... "Show archived bugs?" y ; wait for eternity :) <jmd>Is there a proceedure like mkstemp! but for directories? <paroneayea>jmd: iirc no... I had to write something like that at some point but I forget where it was... <paroneayea>ArneBab_: rekado: you dealt with delay things on your videos you reviewed right? How did you determine what number to put in for the delay in seconds <paroneayea>I think I got the delay right, but I'm not sure about the sign I gave to it... I set it to negative... :) <didi>How do I find a file relative to a source file? <paroneayea>picked up Art of the Metaobject Protocol. Maybe by reading it I can have a better understanding to help contribute to GOOPS. <unknown_lamer>unless something has changed in the last ten years (am I old now?) the GOOPS MOP is pretty limited and could use some major sprucing up <paroneayea>the thing that worries me most is generic method dispatch being O(n) the number of implementers <unknown_lamer>IIRC, there is a dispatch cache. but i might be thinking of sbcl. if there isn't an inline dispatch cache should be doable <paroneayea>there's a dispatch cache but it appears to be using an alist? <paroneayea>unknown_lamer: I've asked for confirmation but nobody has confirmed / denied on that one ;) <paroneayea>unknown_lamer: anyway there are other places where it feels like things could maybe be nicer, and are nicer for CLOS users. <unknown_lamer>goops does beat clos in that generics don't have to have a fixed arglist <unknown_lamer>but missing things like method combinations which can be useful <unknown_lamer>curse you 9-6 as a sysadmin killing my will to hack / no longer having a job writing common lisp <amz3>a tutorial on goops would be awesome, I am wondering how to model an orm in guile <paroneayea>augh this live repl select bug is driving me nuts <paroneayea>I've patched out the offending commit in most of my guix.scm's for right now <gammarus>hi, need help need to convert a time from an api in this format "2011-03-24T20:30:47Z" into something like 13 Mar 2011. <amz3>gammarus: guile doesn't support dates before 3000 <amz3>string->date how surprising <amz3>the software is probably based on minikanren work <gammarus>I have been through the manual and found that. but have struggled to get it to work. <gammarus>thx. I am not a particulary experienced programmer but am trying to learn guile and ncurses. <amz3>gammarus: I can't get it work either <gammarus>yes those were the type of errors I was getting <amz3>I can't help i don't understand <amz3>gammarus: send a mail on guile-user mailling list if you don't get an anwer <paroneayea>daviid: tried debugging that weird error in guile-lib with guile2.2 <paroneayea>it's a weird heisenbug! The moment I put in print statements, the bug goes away. <paroneayea>I'm afraid it's actually a bug in goops somewhere ***heroux_ is now known as heroux
<ArneBab_>paroneayea: I adjusted the delay about 5-10 times. It only started working when I *stopped* adjusting the start and stop time of the video accordingly. <catonano_>amz3`: Barliman is based on minikkkanren indeed