00:10 arkiuat joined 00:15 arkiuat left 00:24 arkiuat joined 00:29 arkiuat left 00:43 bingos left, bingos_ joined 00:47 arkiuat joined 01:00 arkiuat left 01:11 arkiuat joined 01:17 arkiuat left 01:25 kylese left, hulk joined 01:40 arkiuat joined 02:15 hulk left, kylese joined 03:07 arkiuat left 03:13 Aedil joined 03:14 arkiuat joined 03:18 arkiuat left 03:46 arkiuat joined 03:51 arkiuat left 04:12 arkiuat joined 04:50 kylese left, kylese joined 06:04 bingos joined 06:05 bingos_ left 06:09 bingos left 06:11 bingos joined 06:43 bingos left 06:47 Sgeo left 06:50 Sgeo joined 06:56 modula left 07:10 Aedil left 07:14 derpydoo joined 07:25 Sgeo left 07:40 melezhik joined
melezhik . 07:40
tellable6 2025-08-02T02:05:10Z #raku <wayland> melezhik The Sparrowdo link at sparrowhub.io/ links to Sparky instead.
07:43 lichtkind joined 07:50 lizmat left, lizmat joined 08:03 euandreh joined 08:45 Sgeo joined 09:14 arkiuat left 09:43 arkiuat joined 09:47 Sgeo left 09:48 arkiuat left 09:50 melezhik left 09:57 apac joined
wayland . 10:17
10:17 arkiuat joined
wayland Do we have any packages that are specialised for calling a JSON API, or do I just use a web request package and a JSON package, and roll my own? 10:19
10:22 arkiuat left
lizmat doesn't Cro come with that built in ? 10:23
10:23 euandreh left 10:51 arkiuat joined 10:53 apac left
wayland Looks like it does! Wasn't expecting Cro to be the answer for calling APIs, just for making them :) . Thanks! 10:55
10:58 arkiuat left 11:03 derpydoo left 11:11 arkiuat joined 11:16 arkiuat left
ab5tract I'd like to request a sanity check on including a distribution within a distribution. 11:20
The use case is to provide a non-trivial example of Resource::Wrangler.
ie, examples/nontrivial/l{ib, META6.json, etc.} 11:21
11:29 librasteve_ left 11:31 apac joined 11:34 stanrifkin joined 11:36 arkiuat joined 11:40 arkiuat left
ab5tract It seems to be fine šŸ¤ž 11:42
lizmat hehe 11:52
11:56 arkiuat joined 12:01 arkiuat left
Geth ecosystem/main: 385e74289e | (Elizabeth Mattijsen)++ | META.list
Remove File::Directory::Tree

It now lives on as a Raku Community module		 12:13
wayland rolls 3D6 and ... fails sanity check. And becomes catatonic! 12:14
cthulhuwiki.chaosium.com/rules/san...ness-table 12:15
lizmat there's no mention of catatonic there ? 12:16
12:20 arkiuat joined 12:25 arkiuat left 12:33 Aedil joined
wayland Apologies! I refernced it without checking! My brain was running on 5th editing, and they're now up to 7th, and may well have removed it. 12:38
wayland rolls 3D6 again, and narrowly avoids Physical Hysterics! :) 12:39
12:53 arkiuat joined 12:57 apac left 13:01 arkiuat left 13:11 DarthGandalf left 13:17 DarthGandalf joined 13:31 arkiuat joined 13:35 arkiuat left 13:45 apac joined 13:53 bingos joined 13:57 bingos is now known as BinGOs 13:58 arkiuat joined 14:02 wayland left, wayland joined, arkiuat left 14:26 melezhik joined 14:32 arkiuat joined 14:35 Guest11 joined 14:36 Guest11 left 14:37 arkiuat left 14:59 arkiuat joined, apac left 15:03 arkiuat left 15:17 apac joined 15:32 arkiuat joined 15:35 hvxgr left 15:36 hvxgr joined 15:37 arkiuat left 15:41 hvxgr left 16:00 arkiuat joined 16:04 arkiuat left 16:22 arkiuat joined 16:27 arkiuat left 16:41 arkiuat joined 16:56 Guest3298 joined 16:58 Guest3298 left 17:11 Sgeo joined 17:23 arkiuat left 17:25 arkiuat joined 17:30 arkiuat left 17:50 librasteve_ joined
librasteve_ weekly: rakujourney.wordpress.com/2025/08/...searching/ 17:50
notable6 librasteve_, Noted! (weekly)
Voldenet eh, dynamic search is pain and likely will not scale or work fine 17:59
for example when you do search, earlier requests could come later when you type text
and also, sql is really bad at searching, sadly 18:01
18:02 arkiuat joined
disbot4 <librasteve> Voldenet: sorry to disappoint on the Forms side - that's coming next week 18:02
Voldenet because it has to search the whole table, unless you use fulltext index of some sort 18:04
disbot4 <librasteve> this is a raku implementation of the HTMX Active Search example ... htmx.org/examples/active-search/#demo
18:05 melezhik left
disbot4 <librasteve> I guess you are right about scalability ... I know sh*t about SQL search 18:06
<librasteve> [just trusting that SmaokeMachine did it right]
18:06 arkiuat left
Voldenet the problem is that it isn't even ORM's fault if it doesn't do search text well 18:08
it's sql's fault
actually, I have no idea if htmx search has the problem I described, docs aren't precise on how hx-trigger works 18:11
but I've never seen any framework deal with it too well 18:12
worse part, searching for `ab` is going to return more results (and be slower) than `abc` so in practice this data race is easily reproducible with bigger db
18:29 arkiuat joined
Voldenet so, nevermind: I've checked how it works and it enqueues searches 18:34
so when you type in "ab" only one request is executed, "abc" is executed after "ab" is finished
not the best UI in the world, but it isn't buggy 18:35
or rather, when input is changed while things load, it sends another /search (even if it two queries are the same) 18:38
disbot4 <librasteve> sorry was &afk 19:30
<librasteve> :hx-trigger<keyup changed delay:500ms, search>, yeah ... so listens for keyup in the input and only fires every 1/2 sec 19:32
<librasteve> and I see from your note that will only fire again after the new fragment has loaded in the DOM 19:33
<.landyacht.> it would appear docs.raku.org is down for me 19:35
<.landyacht.> anybody else?
<librasteve> I usually say "HTMX is good for smaller scale sites" - but actually I think the limit is more on the level of interactivity eg don't use HTMX to make Facebook or Google Maps
<librasteve> @.landyacht. docs.raku.org is OK for me
<librasteve> I can imagine (in the limit) a Cro::HTTP::Middleware handling only search route over many servers and employing some shared eventually consistent text cache (eg. no SQL sharded) --- I think that in that model HTMX is immaterial performance wise and that the Raku Cro middleware will be the hot path so likely need to use more servers per user than eg JS 19:39
<.landyacht.> it's back for me 19:40
<librasteve> ;-)
<.landyacht.> I'd rather Google be down than my beloved rakudocs... 19:41
<.landyacht.> Also I think there is a genuine bug in Buf's version of splice... What I'm seeing does not line up with the example in the docs, nor expectations 19:42
<.landyacht.> I have to step away, but I'll make an issue when I get back
20:00 MasterDuke joined 20:20 arkiuat left 20:33 arkiuat joined 20:37 arkiuat left
Voldenet well, on basic level htmx doesn't integrate with anything but html, so anything needing microservices will quickly need rewrite with htmx frontend - so you would end up implementing backend react 20:40
the other part about search is probably not a big deal - if you have large dataset you'll quickly find out that it's laggy and why you need n-grams in indexes 20:42
it will work for up to around 1000 records just fine
disbot4 <librasteve> interesting ... setting aside HTMX SSE, I suppose you are saying that it's somehow hard for microservices APIs to serve HTML fragments (I can believe this and that the tooling around react microservices is pretty substantive) 20:47
20:47 melezhik joined
disbot4 <librasteve> I do recall that the naming idea all along was my Cro $service; and aiui Cro can serve HTML fragments and JSON and websockets ... 20:50
20:59 arkiuat joined 21:05 arkiuat left
Voldenet actually you could kinda serve microservices html, but that'd very clearly go beyond htmx 21:22
since every microservice would need to introduce css scoping
it'd be a lot easier to have all rendering compiled into one file out of many 21:23
21:34 arkiuat joined 21:38 arkiuat left 22:02 arkiuat joined 22:06 arkiuat left 22:13 apac left 22:18 lichtkind left 22:30 arkiuat joined 22:34 arkiuat left 23:02 arkiuat joined 23:07 melezhik left 23:10 wayland left 23:24 MasterDuke 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!