<daviid>wingo: my appologies, commit 19f38a3 fully fixes bug#47084 - i mistakenly thought it didn't, because in my wip-3.0 branch, I forgot to revert my own temporary patch for the re-export-public-interface macro - and the good news is ... G-Golf now works with 3.0! <daviid>wingo: this said, it triggers 'new' warnings - like WARNING: (guile-user): imported module (g-golf) overrides core binding `map', for-each' and `connect', i am not sure how to best deal with this problem - hints welcome <daviid>so, wrt these 'new warnings' - no such warnings when using guile 2.2 - related to module-use! (you need guile >= 3.0.6.1-19f38a to try) that replace core binding(s), I don't see any possibility to 'prevent' them, and the only way to silent those is to call default-duplicate-binding-handler and remove warn-override-core and warn from the list <daviid>here is an example, will paste in a min <wingo>daviid: glad things are working for you :) regarding the warnings, see "Add #:re-export-and-replace argument to `define-module'" in NEWS ***apteryx is now known as Guest15199
***apteryx_ is now known as apteryx
<meo>any reason why I shouldn't override a constructor generated by sfri-9 define-record-type? <meo>i have a record of foo bar baz cat, but baz and cat during initialization depend entirely on foo and bar <rekado>meo: you can override them. What I normally do, though, is to leave the plain constructor be, don’t export it, and export a ‘smart’ constructor that wraps the plain constructor. <meo>is it the convention that record accessor names must mention full record type name? <leoprikler>but if you have a different convention there is nothing keeping you from using that instead <meo>and how does one generally deal with deeply nested structures? my guess would be macros? <rekado>for nested structures you may find guile-lens to be of interest <abralek>Hi, I am trying to test an actor I wrote with 8sync. I created another probe actor to send messages to the main one. A problem I am having is that when I want send a *cleanup* to the main actor (server) from the probe, I am getting a backtrace from 8sync/agenda.scm with (select (#<closed: file 7fac07ff6460>) () () #f #f) <abralek>Am I doing something wrong here or it's a bug? I kinda see that we not suppose to give closed files to select. <abralek>How I could or should properly initiate a graceful shutdown? <lampilelo>since streams can be infinite why do they cache all computed values? wouldn't it fill the memory eventually? <lampilelo>in the spec there's only one period between words "infinite" and "cached" <lampilelo>am i doing doing it wrong if i run into corner cases every time i try something new? <leoprikler>lampilelo streams can still be infinite if you discard (and gc) the heads <davexunit>it would be cool to see a stream implementation that kept the api the same but did some magic under the hood to avoid boxing every single value. too much overhead and garbage. <leoprikler>don't you have that boxing as a part of delay/force though? <davexunit>yeah I dunno how to do it but as-is streams are mostly a fun toy that can't be used for anything that needs to perform well <davexunit>I bet if you compared them to generators in other languages they wouldn't stack up too well <dustyweb>did (compile ...) disappear from the standard environment? <dustyweb>(compile '(+ 1 2) #:from 'scheme #:to 'tree-il) <leoprikler>davexunit: come to think about it, how do they perform compared to generators in Guile? :P <lampilelo>test-out-of-memory fails for me on guile 3.0.6, what's up with that? <davexunit>dustyweb: try importing (system base compile) ? <daviid>wingo: ok, thanks, I was aware of #:re-export-and-replace, but didn't think i had to use it in a module that calls module-use! (my mistake) ... but if i do that in foo, that module-use! (srfi srfi-1) - as an example - and say re==export-and-replace for-each, then a module or a file bar.scm that import foo and calls for-each 'stops' triggering the warning indeed <daviid>wingo: the question is how to keep these modules def so users may either use 2.0, 2.2 or 3.0 ... <dustyweb>davexunit: yeah, for whatever reason I thought it was in the default environment <civodul>lampilelo: hi! test-out-of-memory fails in a non-deterministic fashion it seems <lampilelo>civodul: yeah, it complained it can't finalize a thread or something because it has no memory so i disabled it too to check the rest <lampilelo>i guess it's tricky to test memory limits because there's no memory for the test to recover <civodul>well i didn't investigate much, but i suppose that's what happens <lampilelo>thanks for confirming i'm not the only one having this issue <dsmith-work>Ah, sorry. Have a habit of saying "the bot" instead of the "sneek". So the bot doesn't mistake my command for a command.