IRC channel logs


back to list of logs

***heroux_ is now known as heroux
<nalaginrut>morning guilers~
***_zxq9_ is now known as zxq9
<davexunit>wingo's new post on HN front page
<wingo>prediction: 2 comments :)
<davexunit>one will be from one of us ;)
<zch>the only type of articles that make it to the front page of HN are business type or something about paul graham
<taylan>only thing I read about Paul Graham in a long time is :P
<davexunit>I like some paul graham's lisp articles, but when it comes to economics and politics: booooooo
<davexunit>some of*
<zch>davexunit: where is he on the political spectrum?
<zch>I dont really remember
<zch>but isnt he a fan of trickle down economics? :/
<davexunit>zch: he is a rich venture capitalist, soooo
<zch>ah, yeah
<peterbrett_work>My feelings about PG are not suitable for this channel :)
<zch>aw, no fun peterbrett_work
<zch>im going to read some his political stuff now, see how he is :p
<calher>I want to be a libertarian but I don't want to let all the mean people completely loose.
<calher>So Green Party it is.
<zch>oh, hes talking about inequality, this should be good:
<calher>"...was one of the founders of a company called Y Combinator..." and didn't release the source code to HN!
<zch>'I think rising economic inequality is the inevitable fate of countries that don't choose something worse.'
<calher>zch: What is 'worse'?
<zch>calher: Eliminating great variations in wealth
<peterbrett_work>ACTION is disappointed that match-lambda is undocumented
<peterbrett_work>Mainly because I want to use it and have no idea how it works…
<mark_weaver>peterbrett_work: (match-lambda (<pattern> <body> ...) ...) expands to (lambda (x) (match x (<pattern> <body> ...) ...))
<mark_weaver>match-lambda* is the same but does (lambda x ...) instead of (lambda (x) ...), so the resulting procedure can accept any number of arguments, and the pattern matches the entire list of arguments.
<mark_weaver>see module/ice-9/match.upstream.scm
<mark_weaver>sorry for the lack of docs. patches welcome.
<peterbrett_work>mark_weaver: Thanks, that's very helpful
<mark_weaver>you're welcome!
<peterbrett_work>mark_weaver: Okay — so the "match" docs, which *are* in the 2.0.11 manual, actually provide all the info I need
<mark_weaver>ah, good
<mark_weaver>there are some features of 'match' that are missing from our docs.
<mark_weaver>the comments at the top of module/ice-9/match.upstream.scm are more complete
<mark_weaver>e.g. the optional (=> failure) bits
***mikey is now known as Guest71697
<daviid>helo guilers!
***gjanssens_ is now known as gjanssens
<daviid>wingo: it would be awsome if you could drop an eye on this problem: and help me to understand what's going on and how to fix it
<kristofer>I'm trying to compile davexunit sdl2. I've run into this problem: no code for module (ice-9 hash-table))
<davexunit>kristofer: you must be running an old guile.
<davexunit>which means that I should fix my autoconf check.
<kristofer>I'm on guile 2.0.9
<davexunit>that would explain it.
<davexunit>upgrade to 2.0.11
<davexunit>I'll either bump my autoconf check to enforce that, or snarf the procedure I use from (ice-9 hash-table)
<daviid>davexunit: the former rather, my 2c :)
<davexunit>it's a *very* small procedure, and I wrote it!
<davexunit>so I don't mind snarfing the module and just adding it to the build in the case that the module isn't available.
<davexunit>that way older versions of guile will work.
<daviid>davexunit: but other things improved drastically, math related, mark_weaver would tell you better, and it would be better to recommend users to use always the latest stable,imo
<davexunit>daviid: I can *recommend* that they use newer, but forcing that is not something I really want to do.
<davexunit>on the other hand, guile 2.0.11 is 2 years old
<davexunit>(2.0.12 when?)
<daviid>don't know, but I'd rather wait goops related bugs solved before to get 2.0.12 out, imo again
<davexunit>there will always be bugs to fix.
<davexunit>ideally, guile will just release more often so it's not a big deal.
<daviid>but these [at least 2 of these] are years old and just break goops [any 'decent' use of goops]
<kristofer>guile 2.0.9 is the latest release for ubuntu
<daviid>i agree, we need more maintainers :)
<daviid>kristofer: that is an ubuntu 'issue' not a guile issue
<kristofer>I understand.. I'm just surprised it's 2 years out of date
<davexunit>kristofer: distros aren't good about updating guile
<davexunit>anyway, you can use this:
<davexunit>you can install that on your guile load path somewhere
<davexunit>in $moduledir/ice-9/hash-table.scm
<kristofer>I was about to build guile from git :)
<daviid>kristofer: if you are 'serious' about using guile, I really recommend you to clone and use the latest stable-2.0 branch, it has tons of updates, tons ... compared to the latest tarball
<davexunit>kristofer: you can do that as well, but be sure to build the stable-2.0 branch unless you want the.
<davexunit>er, unless you want the braaaaand new guile.
<daviid>kristofer: yeah do that! after git clone, checkout the stable-2.0 branch, it is the bleeding edge stable version of guile, the pre 2.0.12
<kristofer>davexunit: I'm having trouble cloning anything from - I would guess it's a certificate issue..
<davexunit>kristofer: your system must be configured to accept certs from Let's Encrypt
<kristofer>okay I'll investigate.
<kristofer>I'm excited to tinker with sly
<davexunit>use git:// instead
<davexunit>you are brave.
<davexunit>Sly is one of those things you should build from master.
<davexunit>which means building guile-sdl2 from master is also a good idea.
<kristofer>yeah, I'm at the hospital with my daughter and that port is blocked
<davexunit>so you can just skip the cert check for now
<davexunit>GIT_SSL_NO_VERIFY=true git clone ...
<kristofer>perfect, thank you!
<kristofer>I did not expect compile time to be this long ha! how long does compiling guile from source usually take? I'm on a pretty old machine
<davexunit>kristofer: did you check out stable-2.0?
<davexunit>or are you building master AKA unstable?
<davexunit>master takes a significantly longer time to build than stable-2.0
<davexunit>and it's also the wrong branch to be building if you want a stable version of guile
<kristofer>yeah I'm compiling master I believe
<davexunit>both daviid and I explicitly mentioned stable-2.0 above.
<daviid>kristofer: git checkout --track -b stable-2.0 origin/stable-2.0
<mark_weaver>daviid: on non-ancient git, "git checkout stable-2.0" works.
<kristofer>git branch now says (detached from origin/stable-2.0)
<daviid>I like to explcitly list these options actually
<kristofer>wow, that compiles considerably faster
<davexunit>kristofer: explanation of why master is much slower:
<kristofer>alright, weird question.. should I remove the system package for guile before I make install?
<daviid>kristofer: what --prefix did you pass to autogen/configure ?
<daviid>if any
<daviid>that's bad, imo
<kristofer>lol, I can reconfigure it.. I wasn't sure if the packages would coexist or not
<daviid>i recommend you clean, distclean, and redo using --prefix=/opt or another, but _not_ the system and packagers location
<daviid>it is the route to hell :)
<kristofer>okay cool, thanks for the advice
<mark_weaver>the default prefix is /usr/local, which is good for most systems
<mark_weaver>and more convenient I find because /usr/local/bin is usually already in PATH and /usr/local/lib is usually already in the search path.
<kristofer>that's a good point
<daviid>I prefer /opt,but it's opinion, matter of taste ...