šŸ¦‹ Welcome to the MAIN() IRC channel of the Raku Programming Language (raku.org). Log available at irclogs.raku.org/raku/live.html . If you're a beginner, you can also check out the #raku-beginner channel!
Set by lizmat on 6 September 2022.
[Coke] docs question for users. We have a role with 'has $.proc' - it is documented as "method proc" - is this desired/acceptable/wrong? 01:45
docs.raku.org/type/X/Proc/Async
the test to make sure the docs match the source complain, and I'm trying to decide if I fix the test to allow it or change the docs 01:46
antononcube @Coke Are the tests derived from the docs? Or are the docs derived from the tests? 01:47
These didactic questions, of course. 01:48
But I assume they might identify which one -- docs or tests -- is more likely to be "the single origin of truth." 01:49
[Coke] no, the tests call out differences and I as the author decide how to resolve the issue
01:50 jgaz left
antononcube I see -- so the docs are the single origin of truth then. (If correct.) 01:51
[Coke] that's not what I said, but yes, that's the *intention*
the docs describe the *spec*, not *rakudo*, but we use rakudo as our only other source of information about the class/role hierarchy. 01:52
antononcube Right -- I exagerate(d) for didactic / rhetorical reasons.
And that is good to know that the docs describe the spec. 01:53
I think it also means that the test has to be fixed. 01:54
ugexe the docs for that were written 10 years ago. if other objects don't list methods for their attribute accessors it should probably be changed to match whatever is more common
it being marked as a method also misses that the user can instantiate it with a proc object 01:56
01:58 hulk joined 01:59 kylese left
[Coke] I like that last point, but the spec doesn't ever deal with that attribute. 02:01
(which brings up another thing that happens reviewing the docs, do we consider that getter/setter an implementation detail or do we now open a ticket to spec it) 02:02
ugexe IO::Handle docs looks like it list attribute setter methods, but also describe them in the first sentence like "One of the attributes that can be set via .new or open" (see: docs.raku.org/type/IO/Handle#method_chomp) 02:03
i dunno if that is good or not, but it does suggest maybe there isnt a common way on how to handle getter/setters 02:05
[Coke] That points out a gap in the test, which doesn't report on chomp at all in that file. 02:06
(not even as a skipped method) 02:07
[Coke] opens github.com/Raku/doc/issues/4506 02:12
02:15 hulk left, kylese joined 02:41 Aedil joined 05:05 drakonis left, corwin left, jjatria left, sivoais left, hvxgr left, leah2 left, pony left, leont left, andinus left, jcallen left, gfldex left, tailgate left, camelia left 05:06 drakonis joined, corwin joined, jjatria joined, sivoais joined, hvxgr joined, pony joined, leah2 joined, leont joined, andinus joined, jcallen joined, gfldex joined, tailgate joined, camelia joined, leah2 left, leah2 joined, nine left, cleo left, zostay left, coleman left, mtj left, JRaspass left, snonux left, nine joined, cleo joined, zostay joined, coleman joined, mtj joined, JRaspass joined, snonux joined 05:30 Sgeo left 06:27 Chanakan left 06:37 Chanakan joined 07:44 skyesoss left 07:49 sena_kun joined 07:57 wayland76 joined 08:41 sena_kun left 08:58 alaziz joined 09:01 skyesoss joined 09:06 skyesoss left 09:08 wayland joined, wayland76 left 09:24 alaziz left
tbrowder btw, i tested roundtripping all 360 unique font files on my main debian box (using utf8-c8) and got over 100 failures. somebody smart should fix that if they want a challenge. 10:46
i'm trying to come up with an algorithm to match the two different streams of code points using gnu bin (with option "-l" (list) but haven't spent much time on it. 10:51
antononcube: it might be easy with your LLM magic 10:52
lizmat tbrowder: it's my impression that it's basically unfixable, though as some sequences produces the same "fix", and thus differ on roundtripping 10:55
tbrowder "gnu bin" should be "gnu 'cmp'"
lizmat tbrowder: if you want to roundtrip, slurp(:bin) and spurt(:bin)
antononcube Why use ā€œgnu cmpā€, why not ā€œā€˜Algorithm ::Diffā€ ? 10:57
tbrowder i tried that once and ran into problems, but i probably did it wrong. before the roundtrip i tried to only use :bin on files that were !~~ Str
so :bin *should* roundtrip? 10:58
antononcube Using LLMs to compare files is possible, of course, but, generally, ā€œexpensiveā€ and having size limitations.
lizmat tbrowder: perhaps raku.land/zef:lizmat/path-utils#path-is-text can be of help ? 10:59
tbrowder: yes, :bin *should* roundtrip
tbrowder ok, i'll try both on just the font files, thanks!! 11:02
will report back later...
11:13 MasterDuke left 11:28 skyesoss joined 11:29 elcaro left 11:30 elcaro joined 11:33 skyesoss left
ab5tract Neat CSS neat.joeldare.com/ 11:37
librasteve: ^^ maybe something usable for the htmx efforts? 11:38
tbrowder wow, :bin did the trick, ZERO bad comps. and thatb 11:47
*that include PostScript type 1 fonts which have text reprs. gonna try on known pdfs now. 11:48
the docs misled me a bit, but i'll look again after all works for my baby module 11:49
antononcube ā€œbaby moduleā€ -> ā€œbabe moduleā€ -> ā€œcougar moduleā€ -> ā€œcommunity moduleā€ 11:57
librasteve ab5tract: neat idea (geddit?) 12:08
I'm doing a bit of rubber ducking about Htmx and Cro here btw github.com/librasteve/raku-HTMX/issues/8 all f/back & contributions welcome 12:09
tbrowder just searched 950 pdf files and found only 50 bad ones. all of them are very old 1200 dpi scans of photograhs. 12:13
i'm going to try rescanning the bad ones using the utf8-c8 option 12:14
no, the scans were from my classmate's 1200 dpi scan of our college freshman annual, using an old windows app, circa pre-2015 12:16
all the bad pdfs were from that batch 12:17
13:07 jgaz joined
ab5tract Haha! 13:16
(I geddit) 13:17
tbrowder: interesting, I would think that image-based PDFs would have little to get wrong relative to text oriented ones with embedded fonts and extensions and other weirdness 13:19
tbrowder well, i think @lizmat mentioned OCR having differing bits somehow, as well as i think the original images were tiff fomat and i think i created the pdfs with some gnu app. 13:48
lizmat that shouldn't matter if your reading with ":bin" ?
tbrowder lots of ā€œthinkosā€ there. iā€™ll see if i can find the original creation path 13:49
i know i was using perl wrappers at the time 13:50
antononcube šŸ˜± 13:51
tbrowder gotta goā€”ā€œon the road againā€
13:56 Sgeo joined
ugexe i'm not sure what purpose round tripping binary data using slurp/spurt would serve over just copying the file 14:16
in other words: why are you not just copying the font file? 14:18
14:31 [Coke] left 14:33 [Coke] joined 15:00 skyesoss joined 15:46 skyesoss left 17:58 skyesoss joined
[Coke] anyone have a recommendation for a windows laptop for dev work (and also, I guess, gaming?) Would like to get a personal machine to work on windows related raku stuff that isn't also a worktop. 18:14
ivarec Gaming laptops are usually a sane choice. They handle well the heat. I've been lucky with my Asus ROG 18:19
19:21 skyesoss left 19:30 skyesoss joined 20:03 swaggboi left 20:08 swaggboi joined 20:47 swaggboi left 20:56 swaggboi joined 21:02 dustinm` left 21:05 swaggboi left 21:07 dustinm` joined 21:17 swaggboi joined 21:21 jgaz left
tbrowder ugexe: in order to extract resources/files from published modules 21:25
21:25 swaggboi left
tbrowder slurp the content, spurt the content to a local file. 21:26
INSTALLED modules
21:26 sena_kun joined
tbrowder didn'mean to shout 21:26
21:30 swaggboi joined 21:45 sena_kun left 22:27 skyesoss left
tbrowder [Coke]: have you tried a KVM host of windows on yr personal mac? with enough ram and disk space it can work well 22:40
23:05 calma joined
calma hello c: 23:06
i have a question
raku is a niche language?
is worth to learn?
23:13 MasterDuke joined 23:37 calma left

Powered by Cro and the RakuĀ® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!