🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
00:55 MasterDuke joined
Geth rakudo: vrurg++ created pull request #5480:
Minor cleanups for #5477
linkable6 RAKUDO#5477 [closed]: github.com/rakudo/rakudo/pull/5477 Move closer to complete generics instantiation
06:08 MasterDuke left 09:24 sena_kun joined 09:42 sena_kun left 12:02 [Tux] left 12:04 [Tux] joined
[Coke] DOCS#1234 13:07
linkable6 DOCS#1234 [closed]: github.com/Raku/doc/pull/1234 Docs for ≤ ≥ ≠ unicode ops
[Coke] huh.
I assume that's an AlexDaniel++
lizmat yeah, many of the bots are functional again 13:08
like 17e3d0e6c75863ee090d2
linkable6 (2023-11-23) github.com/rakudo/rakudo/commit/17e3d0e6c7 RakuAST: don't apply pragma semantics to native types
Geth rakudo/revert-5477-complete-generic-instantiation: 1b73d18d85 | (Vadim Belman)++ (committed using GitHub Web editor) | 30 files
Revert "Move closer to complete generics instantiation"
rakudo: vrurg++ created pull request #5481:
Revert "Move closer to complete generics instantiation"
rakudo/main: 1b73d18d85 | (Vadim Belman)++ (committed using GitHub Web editor) | 30 files
Revert "Move closer to complete generics instantiation"
rakudo/main: 73470580c9 | (Vadim Belman)++ (committed using GitHub Web editor) | 30 files
Merge pull request #5481 from rakudo/revert-5477-complete-generic-instantiation

Revert "Move closer to complete generics instantiation"
16:14 donaldh joined
donaldh I was looking at docs.raku.org today and sadly I found it pretty darn unusable. 16:19
There's a significant delay between each char as I try to type into the search box.
When searching for Str, the result for the Str class is not visible until you scroll far down the results. I followed the first Str result which took me to docs.raku.org/routine/Str and left me very confused. 16:21
There seem to be 2 scrollbars and trying to page down after selecting a search result only moves the outer scroll region. I thought there was no more content until I clicked to focus on the inner region and was able to scroll further down. 16:22
If I type "string" in the search box, I don't get any results for the Str class. 16:24
If I didn't already know there was a page for Str with docs for all its methods, I would have given up. I'm a reasonably experienced Raku user so I persevered. I think this would be a bit horrifying for a newcomer. 16:26
nemokosch This is the usual "fix something and something else breaks" situation 16:29
currently, all search results are fetched, therefore anything very short will cause a hang
previously, only the first N results were fetched and there was no way to even access the rest 16:30
I think on average it's not that bad - as long as you can accept that the first 1-2 characters will always cause a couple seconds freeze 16:31
donaldh Maybe shouldn't send a search for a single character. It's also blocking the browser from showing the typed chars, which should be avoidable.
nemokosch If you are searching for something that has a gazillion results, that's always going to be bad; ideas are welcome for that
I was thinking about running a timer before fetching anything so that intermediary searches can be prevented 16:32
the thing is, knowing the little human resources the doc site is running on, I didn't really want to bring it up
kind of a "patches welcome" situation
donaldh Yeah. I realise there are valid single-char searches when searching for raku operators, but a time delay or a min 3 chars before sending any search could be a workable solution. 16:33
nemokosch there should be some solution for sure 16:34
you could also bring this up at #raku-doc or open an issue in the Raku/doc-website github repo
you know, the current position is a result of steps that always seemed to be in the right direction. Knowing that "from the A to the Z", I'm kind of content with the way it works, knowing the current states and the tradeoffs. Freezing for a couple of seconds is more acceptable to me than a chaotic order or bad navigability of the results 16:38
donaldh Sure. I've not been active in the raku community for a few years so I just jumped on here 'cos it's still in my IRC. I'll join #raku-doc and if I have time I might be able to contribute a fix or two. 16:39
nemokosch Sure, it could be better but that's how it went the whole time: changing something in a direction that "we" (whoever was more involved in a particular change) thought was better. This is how it goes normally.
Not having a lot of people with UX and frontend development experience, it's kind of a stumbly process 16:40
And as much as the problem is legit, if you understand why it happens, I would say it doesn't seem so dramatic. If there was a full site load, you'd have to wait a similar amount of time without responsivity 16:41
full matches should be prioritized over something that is just contained, though 16:43
that's for sure 16:44
after all, the order of the results is the main value provided by a search
donaldh The search UX is not my main concern – it's more that the docs for Str seem buried and hard to "discover". I think there are a couple of things that would help: 1. Add a page for strings that would appear at the top of the search results under "Language". It should match for "Str" and "string. 2. Add more links from pages to the page for class Str. 16:46
nemokosch I wouldn't say "hard to discover", not to something like that 16:48
it's not something you accidentally type, you probably know what you are looking for
and if you do, it's pretty easy to find; actually, probably easier without search than with it 16:49
donaldh If I were a new user to the language and I wanted to learn how to work with strings, I would type "string" into the search box.
The results are not helpful.
nemokosch the problem is, the "doc site" is not one thing 16:50
it's two at least
donaldh > probably easier without search than with it
nemokosch it is a reference on one hand, and I'd expect that somebody who types "Str" wants that use
donaldh I agree, the Types page gets me there quicker. But it's still buried at the bottom of the page and that takes me back to the scroll problem. Page down does not work. 16:51
nemokosch and it is kind of a guide on the other hand, and that is really hard to look up indeed, but it's also pretty hard to mitigate
my usual stance on the guide usage is that there should simply be a lot of cross-referencing 16:52
it helps more than trying to make it into some lookup-table style
page down does work... if you click into that view. This is another "problem caused by a solution" situation 16:54
I would naively assume, though, that the middle, "content" view could be just made smaller, and then the whole site should fit into one page, leaving only one scroll bar 16:55
anyway, I'm already late for something, gotta go 😅 16:56
donaldh 👍
lizmat donaldh please make a doc issue with your findings and suggestions 16:57
donaldh I will do. I'll hopefully manage to contribute some PRs as well. 17:05
lizmat donaldh: would you be interested in writing an Advent post ? 17:07
donaldh lizmat: hmm, not sure I have a topic in mind, or the time to do something worthy. 17:33
lizmat donaldh: no worries... thanks for at least considering :-) 17:34
donaldh I see there are quite a few days free. 17:36
El_Che donaldh: it can be yours for the right price! 17:37
[Coke] added a feature request for doc-website so that if someone searches for, e.g. "string", that "Str" would be considered a match as well. (and to make it configurable so we can guide other searches as well) 18:29
Feel free to follow up on that in raku-doc
18:45 sena_kun joined 22:37 sena_kun left