IRC channel logs

2017-08-16.log

back to list of logs

<manumanumanu>Good morning everyone! I have spent the whole day gardening
<manumanumanu>Such a fantastic hobby, which sadly my whole family considers a waste of time
<manumanumanu>planning a garden is fun
<nee`>I took a short look at guile-pg and why it doesn't build with guile2.2. It uses the scm_port struct which was renamed to scm_t_port and had it's definition removed from libguile/ports.h. Is there some alternative, or should a pointer to a custom struct be restored in SCM_STREAM now?
<wingo>nee`: does guile-pg implement a port type?
<wingo>in summary tho the change is documented in NEWS and the port API is documented in the manual: https://www.gnu.org/software/guile/manual/html_node/I_002fO-Extensions.html#I_002fO-Extensions
<nee`>wingo: yes for a postgresql connection I think http://git.savannah.nongnu.org/cgit/guile-pg.git/tree/src/libpq.c
<wingo>hoo, that code is pretty old
<wingo>if you can, you probably want to get rid of the mark function for the "xc" data type
<wingo>see https://www.gnu.org/software/guile/manual/html_node/Foreign-Object-Memory-Management.html#Foreign-Object-Memory-Management for a discussion
<nee`>I've been using it to connect to postgres. I'm now considering wether it's less work to upgrade it, or to just switch to guile-squee (which is pure guile code).
<wingo>yeah i feel you. in your place i would look at switching to guile-squee but i understand the pain of migrating working code :/
<wingo>if you want to look further regarding guile-pg, civodul has a commit to gnutls where he makes the gnutls guile bindings use either old or new C port types
<nee`>okay thanks for the hints
<rekado_>manumanumanu: oh, a fellow gardener!
<ennoausberlin>Hello, Guix
<ennoausberlin>I wonder if there is a way to get the result of a grouping in a guile regexp
<ennoausberlin>Lets say: I have a string like "Starting point: 'xxxxx' end point: 'yyyyy' Time: zzzzz" where xxxxx, yyyyy and zzzzz arent fix. How do I get them as result of a match:string
<ijp>okaky, so I think I know my issue from yesterday which is that I don't think the module was "beautified"
<cmaloney>morning
<ijp>afternoon
<dustyweb>o/
<manumanumanu>rekado_: it's a shame I am got started so late. We have had a house like... forever.
<stis>Hey all!
<stis>The irc logs does not work for me with chrome, some problems with sequrity it seams. Anyone knows what happened?
<lfam>stis: The site's TLS certificate expired
<stis>thanks for the info ifam
<daviid>nee`: if you switch to guile-squee, checkout the devel branch, because I autotool chained there (in the devel branch) and for some reason, christopher did not merged these mofifs yet ... - also note that guile-squee is very experimental and does not compare, by no means to guile-pg in terms of functionalty
<daviid>nee`: https://notabug.org/cwebber/guile-squee/src/devel
<zv>so, https://gnunet.org is down with an expired certificate
<zv>also uses HSTS -- can't that *permanently* lock you out of your site?
<zv>luckily w3m still works :)
<lfam>zv: Once the certificate is valid again, then the site will work again.
<zv>lfam: isn't the thing with HSTS that it delivers TOFU in addition to a CA-verified certificate?
<zv>oh, actually that's HKPK
<zv>err, HPKP
<lfam>HSTS tells the client (like your browser) that the site should only be accessed over HTTPS. It's a way for site operators to disable unauthenticated HTTP with compliant clients.
<lfam>As you found, it's possible to write a client that doesn't handle it
<lfam>Or, you can just tell your client that you will accept bogus TLS
<lfam>One shouldn't enable HSTS unless they are confident they can manage their TLS certificates
<lfam>zv: I'm curious, where did you get your w3m? From a distro?
<zv>lfam: yeah
<lfam>The project was abandoned upstream and Debian seems to be the most active development site
<zv>of course there is no way to search the logs the day I need them :)
<zv>ijp: is there any way you remember a conversation about something having to do with some scoping issue or something like that about 6 months back?
<zv>SICP-related?
<ijp>not off the top of my head, but if you have some more context...
<zv>There was a bug in guile, or at least you said it was a bug
<zv>it was let/letrec related??
<ijp>I vaguely remember it
<zv>you angrily insisted it was a bug when i brought it up later?
<zv>s/angrily/firmly
<zv>well i can't search logs so this is why i'm bugging you
<zv>maybe there was a guile patch, i'll have to take a look
<ijp>okay, july 14th?
<ijp>2016...maybe this is the wrong one
<zv>maybe just search "sicp" if you're keeping logs
<ijp>is this it http://shift-reset.com/tmp/zvlog?
<ijp>whoops wrong encoding :/
<ijp>I think my "firm insistence" relates to the question about symbol-set! vs module-set!
<ijp>dustyweb: assuming I'm not accidentally invoking the wrong continuation (which is always a possibility), then apart from a few lines which call resolve-module (<= 6 of them) we have boot-9
<dustyweb>ijp: woohoooooo!
<dustyweb> ASCII DANCE PARTY
<dustyweb>
<dustyweb> /_o_/ \\ / \\o_ o
<dustyweb> ( _|_ ) //)
<dustyweb> /\\ / o \\ /| /|
<dustyweb>
<dustyweb> *BMCH BMCH BMCH BMCH*
<bavier`>lol
<ijp>that was very confusing for a few lines
<dustyweb>;)
<dustyweb>nice work ijp
<dustyweb>I know that's a hell of a lot of hacking
<dustyweb>now bee arr bee, lunch
<ijp>I owe my success to dustyweb, the good people at Tetley and 1980s rock music
<janneke>dustyweb: you ever read `the dancing men?'
<dustyweb>:)
<dustyweb>janneke: I have not, what is it?
<janneke>dustyweb: it's a short Sherlock Holmes story by Sir Arthur Conan Doyle
<janneke>your ascii dance party has some resemblance to it
<janneke>more than some
<dustyweb>aha :)
<janneke>:-)
<zv>thats it ijp
<cmaloney>I'm getting an error on this line in Fibers: https://github.com/wingo/fibers/blob/master/fibers.texi#L826
<cmaloney>Should that be @code{sup} ?
<cmaloney>Introdced in this commit: https://github.com/wingo/fibers/commit/2a4f707827105d041668f34b62daaff871c22c07#diff-b2b20f2adfe41d88efea699432cd3096
<cmaloney>or is that supposed to be something like superscript?
<cmaloney>Feh, naturally that was added after ubuntu 14.04 shipped. ;)
<cmaloney>I hacked the doc so it uses @math{2^{32}} instead. Seems to work on older versions of Texinfo