***catonano_ is now known as catonano
<wingo>just popping in to say that i find the guix discussion of #t/#true extraordinarily amusing <civodul>heheh we're really talking about the color of the bikeshed <civodul>but the color is the first thing passersby see <wingo>yeah to me it's a fine change, either way, i am just amused there is opposition to it <wingo>and that the argument is essentially about confusion, which i find weird <wingo>or about what guile does, and that's precisely the reverse of how it should go: guile can change *wingo has not had any guile time recently :P <wingo>how is guix going these days? seems pretty good but hard to tell <civodul>and now there are so many big changes compared to the previous one <civodul>people have been preparing the on-line Guix Days too, that's great <wingo>yeah that looks pretty cool :) <civodul>i think we're at the stage where we're starting to reach out to people who are much less on the "geek" side of things, so to speak <civodul>and i can feel that things we never bothered properly documenting before now need to be documented, for instance <civodul>BTW, we'll have to do Guile 3.0.5 someday to enjoy the new case dispatch! <brendyyn>civodul: i was thinking for such code style changes, are there any code refactoring tools for scheme/guile or would you just find a way with emacs? <civodul>Emacs or even sed should be good enough for that task :-) <ArneBab>wingo: short experience about Guile I had the past weeks: I’ve written a quick static site generator for a family member, and due to some design choices (ordered permutations of keywords in URL) it had grown somewhat slow. Since switching to Guile 3, it is easily fast enough again. You made a real impact here. <wingo>i am a dinosaur and i still do perl -ipe 's/foo/bar/' $file <wingo>glad things are going well :) i like it when compiler improvements can paper over algorithmic problems <wingo>like "i know i'm doing this wrong but i am happy it works" ;) <ArneBab>Same for me :-) — though I’m not actually doing it wrong. The decision „we want gallery pages for all ordered permutations of these keywords“ enforces an expensive algorithm (to the best of my current knowledge). This was no problem while the content conformed to the original decision (5 keywords). Then new keywords happened … and luckily the improved compiler was ready to compensate ;-) <ArneBab>I said „yes, we can make that, up to X keywords, and yes, this will be clean“ <ArneBab>lay-people just cannot judge how the cost changes if you add a single element to a permutation. I’s just one keyword, right? ;-D <ArneBab>maybe I should add a check that errors out at >8 keywords :-) — but still: that we could go up to 8 is great! <ArneBab>(if we go higher, the filesystem will become a limiting factor for page access) <ArneBab>⇒ with the improvements in Guile 3, it provides sufficient performance that the filesystem will be a bottleneck before Guile. And that’s awesome! *ArneBab is aware that filesystem layout optimizations can fix the filesystem bottleneck, but then the site will require a migration path, and this still is a side-project for family :-) <lloda> i do ctrl-r on the guile prompt and instead of the regular readline reverse search i get a bizarre bck? prompt. Anyone seen this? <leoprikler>Probably (substring/read-only (string-copy s) 0) is what I want <mwette> lloda: did you (activate-readline) ? <lloda>mwette: yeah, it's in my .guile <lloda>even after doing it manually on the repl i still get the bck? prompt :-\ <lloda>anyway without readline i wouldn't get a bck? prompt, i'd just get ^R (i think?) <mwette>hmm there is a c function rl_get_keymap_by_name <lloda>think i switched modes inadvertently <lloda>don't know how to switch back tho. I'm used to having to ctrl-e now and then, but that doesn't work for this <lloda>i've also lost tab completion... <lloda>actually i get tab completion for directory files in cwd even in the guile repl <lloda>i don't remember writing it myself, but maybe i did (?) <str1ngs>lloda: okay that should not effect anything then. And you are using guile from command line and not say geiser? <str1ngs>lloda: does M-r work instead of C-r? <lloda>gnome-terminal, M-r doesn't work <lloda>guile works fine by itself, it's my embedding of it that has the problem <lloda>didn't ever happen before today tho <lloda>think i'll just reboot everything :D <str1ngs>strange, C-r should show (reverse-i-search)`': then. the only time I've seen a back prompt is when using geiser. in which case you need to use M-r <str1ngs>can you past the actual prompt you get with C-r? <lloda>bck: appears when i press C-r <str1ngs>lloda: does bind -P | grep history-search-backward output anything? <str1ngs>lloda: see if bind -P show that C-r is bound. <lloda>re-read-init-file can be found on "\C-x\C-r". <lloda>reverse-search-history can be found on "\C-r". <lloda>revert-line can be found on "\e\C-r", "\er". <lloda>^ that's what i get from bind -P | grep C-r <str1ngs>I've never seen bck: that's just really odd. are any of the vi bindings bound when you list them with bind -P lloda ? <lloda>bind -P | grep vi | grep -v 'not bound' <lloda>previous-history can be found on "\C-p", "\eOA", "\e[A". <str1ngs>reverse-search-history is not actually bound for me. though yours is bound to C-r which is correct anyways. <str1ngs>you can try reloading .inputrc see if it helps with C-x C-r <str1ngs>but I suspect it's not an issue with your ~/.inputrc <str1ngs>anything in ~/.guile that changes readline behavior at all? <lloda>i have (import (ice-9 readline)) (readline-set! history-length 50000) (activate-readline) that's all <lloda>C-x C-r didn't do anything btw <str1ngs>C-x C-r is silent so maybe start a new terminal just encase <lloda>gtg, thanks for the help str1ngs <str1ngs>lloda: ahhh this only happens when you embed. need to think more on this