🦋 Welcome to Raku! raku.org/ | evalbot usage: 'p6: say 3;' or /msg camelia p6: ... | irclog: colabti.org/irclogger/irclogger_log/raku
Set by ChanServ on 14 October 2019.
muffindrake Hewwo 01:53
lucs Gweetings 01:57
Geth doc: 581df70bcd | Coke++ | writing-docs/STYLEGUIDE.md
Fix "Perl 5"
03:28
doc: 7b450219dc | Coke++ | writing-docs/STYLEGUIDE.md
minor formatting
Geth doc: 1260934a53 | Coke++ | README.md
Remove contradictory paragraph
03:57
doc: 738991f664 | Coke++ | 6 files
Prefer "Raku" over "raku"

Both work, but Raku is the canonical version
04:06
enge vvvv 04:34
guifa2 Whoa, what happened between the last two releases of Comma? It's buttery smooth on macOS on the newest compared to the previous 05:35
jmerelo releasable6: status 11:40
releasable6 jmerelo, Next release in ≈10 days and ≈7 hours. 3 blockers. 0 out of 50 commits logged (⚠ 35 warnings)
tellable6 2021-03-09T13:38:40Z #raku <tbrowder> jmerelo jj, tweak has disappeared from the docs again
releasable6 jmerelo, Details: gist.github.com/1f13e4400eed9b82d2...342c43609a
jmerelo .tell tbrowder what? I'll have to check that. Thanks 11:41
tellable6 jmerelo, I'll pass your message to tbrowder
lucs lizmat: I like your idea of having 'r' to invoke Raku, and that is how I discovered (after a bit of head-scratching) that 'r' is a zsh built-in :) 12:58
tadzik :D 12:59
lucs I was lucky not to accidentally re-invoke something with nasty side-effects. 13:00
tadzik: Hi fellow lurker ;) 13:01
tadzik o/ ;) 13:02
my lurking was interrupted by my server being migrated to the literal cloud 13:03
cetjs2 jmerelo, hello 13:25
jmerelo, why perl foundiation isn't approved in gsoc 2021? 13:26
jmerelo cetjs2: I'd like to know why too 13:27
It simply was not.
We didn't get any more feedback than telling us there were many submissions. 13:28
cetjs2 jmerelo, (
jmerelo Right. Big disappointment. Sorry about that.
We're gonna try again with season of docs, but of course and again, no guarantee... 13:29
cetjs2 jmerelo,I can be realize this project for fun... but gnu main project rejected too 13:31
cetjs2 jmerelo, in gnu main project were better idea( 13:32
jmerelo was it? 13:33
Well, going through the accepted projects it's difficult to find out why some were accepted... while others were not.
Some note taking apps... Upcoming languages... Disjointed projects from an university. They probably put up a better application that I did, but still... 13:34
leont Their criteria have always been weird 14:17
And fairly arbitrary 14:19
leont In 2013 they accepted Parrot Foundation, even though MoarVM had already been announced and it was entirely obvious to any insider that Parrot was doomed the moment any competitor arrived on the scene 14:29
moritz well, I'd don't think gsoc was a parrot/perl insider 15:04
MasterDuke56 codesections: not related to the meat of your answer to stackoverflow.com/questions/665653...e/66566599 but if they don't need the `@ids` to be lazy, using a range (i.e., `..` instead of `...`) is even faster, and `xx` is faster still 15:24
tbrowder well, again i can't get docs.raku.org, it's intermittent (ssl protocol error, chrome), but it sure looks like a tls problem on the server end to me. i have no problems with any other sites. 17:25
tellable6 2021-03-10T11:41:11Z #raku <jmerelo> tbrowder what? I'll have to check that. Thanks
lizmat tbrowder; works for me :-( 17:26
tbrowder i know, that's what everyone says, but i think most ppl who use docs are outside us (but it may be my isp). 17:28
lizmat tbrowder: are you behind some proxy ? sometimes overloaded proxy servers can produce those phenomena
tbrowder no proxy 17:29
lizmat which provider?
tbrowder but isp is cox cable. i've talked to them and they swear there is nothing between me and the internet 17:30
lizmat hmmm
tbrowder i have tried various dns servers like google and opendns but still get intermittent probs with docs.raku.org 17:32
jmerelo tbrowder: I'll tell Roman
tbrowder thnx, jj
lizmat tbrowder: do you have a time when this problem started? 17:34
lizmat wonder if it could somehow be related to travaux.ovh.net/?do=details&id=49471& 17:35
it appears this also affected the banking app of one of the leading banks in NL
jmerelo lizmat: rba had set up a CDN. Maybe the CDN is hosted partly there... Anyhow, I've reported it to rba 17:37
raydiak fwiw, I've had the exact same intermittent tls error as tbrowder on comcast using linux chromium for at least a month or two 17:53
rba tbrowder: Try docs.rakulang.site 17:55
lizmat feels to me TLS errors are an indicating of a Man in the middle having problems with handling certificates ?
rba++
rba This is directly to the server, not via CDN.
raydiak docs.raku.org/ currently doesn't work for me, but docs.rakulang.site/ does 18:00
lizmat perhaps one of the servers in the CDN is having issues, hence the intermittentness of the issue? 18:01
raydiak just a stab in the dark, but one of the common causes of this error is serving http on 443 where https should be served. possible misconfig of the cdn or even just one of its edge nodes? it's not an expired cert error, which is what I assumed at first when I saw this back in jan iirc
rba For me docs.raku org works. Yet this doesn’t mean anything as the CDN has different entry points.
raydiak error is 3 lines: This site can’t provide a secure connection docs.raku.org sent an invalid response. ERR_SSL_PROTOCOL_ERROR 18:02
rba www.cloudflarestatus.com 18:03
jmerelo A couple of places in America have been rerouted... 18:05
codesections yeah, that sounds a lot like problem on the cloudflare end. Maybe it's something to open a support ticket with them about?
jmerelo Anyway, remember there's an almost up-to-date mirror at rakudocs.github.io There's at least 1 good thing of having fully static pages
codesections jmerelo: yeah, but I'm concerned about how widespread this could be. Getting an apparent server error on your first or second visit to the docs page is a pretty awful first impression for us to make 18:07
jmerelo codesections++
tonyo_ looks ok for me, in the western US
codesections There are _lots_ of good things about having a fully static site :D
raydiak even weirder is to circumvent it I've been going to http instead, then it sends me to safebrowse.io or something and says this site is dangerous (redirect is on the cloudflare end, not a browser error, confirmed this with netcat). I click proceed anyway, and then it takes me to the site with https, and it works...for a while... 18:08
codesections (works for me too, in central US)
lizmat I wonder if CloudFlare has names for each of its entry points, so one could check if one is faulty ? 18:09
for a particular site like docs.raku.org 18:10
?
rba I‘m not aware of such a detailed state info. 18:11
lizmat is there a way to escalate this ? 18:12
rba If we have too much trouble I can try turn of the cdn.
tonyo_ if you curl -v it shows a `report-to` header
shows cf-nel for me, currently in oregon
codesections there's also the `cf-request-id` header, but not sure of the details 18:13
cf-nel for me too 18:14
with the url a.nel.cloudflare.com/report?s=$long-id 18:15
but I don't think that indicates an entry point. See support.cloudflare.com/hc/en-us/ar...or-Logging 18:17
tbrowder one prob may be using css or js (or other resources) from non-https links, a definite no-no with modern browsers. i do sometimes see the docs without css.
codesections nel = NEL = network error logging
raydiak shouldn't have to do with links, this error happens before the page content is even loaded 18:18
curl gives me curl: (35) error:1408F10B:SSL routines:ssl3_get_record:wrong version number
tbrowder well, i'm glad i'm no longer crying wolf :-D 18:19
rba Changed docs.raku.org from proxified to dns only
codesections never thought you were :)
raydiak which iirc, just means it got back something that isn't a valid TLS ServerHello, not necessarily a version mismatch specifically 18:20
codesections oh, curl also gave the SSL cert:
subject: C=US; ST=CA; L=San Francisco; O=Cloudflare, Inc.; CN=sni.cloudflaressl.com
tbrowder just kidding, when i first reported it some time ago i don't remember anyone else seeing it 18:22
raydiak thanks rba++ . I still get the error, but I'd assume it takes a while for the dns changes to propogate 18:23
codesections rba: is turning off the CDN going to mean we have too much traffic/hosting costs more, or will it just not be as fast for users who aren't near our server? 18:24
(and where is our server, anyway?)
lizmat not in Strasbourg, presumably :-) 18:25
codesections raydiak: curl now shows me connecting directly, so the dns has updated for me anyway
tbrowder jmerelo: ref gsoc, remember sean morrison at brlcad.org? they got accepted for gsoc again this year (for the zillionth time). one thing sean did some years ago was partner with other cad orgs and entered as the leader of a group of orgs. 18:27
jmerelo tbrowder: my buddies at chaoss.org got accepted again this time over... They always are. 18:28
tbrowder it provided a wider menu of interests for students as well as maybe have more mentors
jmerelo tbrowder: but I really don't know. It's true that last year the proposal was not so good, but I really tried to make a good case this year. Wider menus might be it, but then it might not 18:29
tbrowder: I'll try with season of docs, anyway
BTW, I will share the proposal with you
tbrowder hm, healthcare, lots of points there. raku vs go, maybe not so many points. of course we know google is completely unbiased 18:30
jmerelo tbrowder: but then, "Pharo" was funded. 18:32
tbrowder what is pharo
jmerelo tbrowder: exactly
tbrowder heh
jmerelo Due to lack of feedback, it's really complicated to "improve" proposals from one year to the next. We don't know if it's the track record, the language itself and whatever people are lately saying about it, or the proposal. 18:33
raydiak codesections: working from my server in central california, but not yet for me locally up in washington state. wouldn't be surprised if there's some DNS caching on my ISP's end
jmerelo Anyway, we'll keep trying. When funded, it's a good thing for the language and community. It takes some time, but well, ROI is good if achieved. 18:34
tbrowder jmerelo: i suggest you see if sean morrison might critique your application. it looks good to me but sean is expert at gsoc 19:33
his track record is of olympic std 19:34
tbrowder idea: 19:39
jmerelo: why not add the one or two alternate doc sites to the raku.org homepage. i have never had a problem landng there 19:40
jmerelo tbrowder: OK; I'll see if I can do that. Thanks for the suggestion :-) 19:42
codesections but we should be careful how we present it. We don't want it to look like we need multiple backups to keep the docs online :D 19:44
(not that I think we'd put it like that :-)
raydiak I do think it's still broken. my local machine and my server are both resolving to the direct IP address. but on my local machine, the error persists. firefox gives the slightly more informative error "SSL_ERROR_RX_RECORD_TOO_LONG". according to wireshark, after it sends the ClientHello, it receives a tcp payload of 256 FF bytes instead of a TLS ServerHello. docs.rakulang.site/ works great everywhere. 19:49
MasterDuke codesections: using `method get_timestamp() { use nqp; (nqp::time_n() * 1000e0).Int; }` (note lack of return and num literal), `my @ids = $worker.get_id() xx 10000000;` is ~7s faster than mapping over the Range 19:50
codesections MasterDuke: Interesting, ty. I get why the `xx` is faster, but why is `(num * Num).Int` faster than `(num * Int).Int`? 19:52
or are our literals smart enough that 1000e0 is a num rather than a Num? 19:53
MasterDuke well, i'll admit i haven't benchmarked that change very much 19:54
lizmat codesections numeric literals are almost all both in teh AST 19:56
if I remember correctly :-) 19:57
MasterDuke benchable6: HEAD compare use nqp; my num $n = nqp::time_n(); my $i; $i := ($n * 1000).Int for ^1_000_000; ||| use nqp; my num $n = nqp::time_n(); my $i; $i := ($n * 1000e0).Int for ^1_000_000;;
benchable6 MasterDuke, starting to benchmark the 1 given commit
MasterDuke, No new data found
codesections also, it seems like (int * num) should return a num, but right now it returns a Num. Do you think it'd be worth adding an infix:<*>(int \a, num \b --> num) multi? (and a matching (num \a, int \b) one)?
right now, num * int falls back to Num * Num 19:58
MasterDuke benchable6: compare HEAD use nqp; my num $n = nqp::time_n(); my $i; $i := ($n * 1000).Int for ^1_000_000; ||| use nqp; my num $n = nqp::time_n(); my $i; $i := ($n * 1000e0).Int for ^1_000_000;;
benchable6 MasterDuke, starting to benchmark the 1 given commit
lizmat afk again&
benchable6 MasterDuke, ¦HEAD: «Benchmark: ␤«timed out after 10 seconds»»
codesections lizmat: ty good to know
MasterDuke benchable6: compare HEAD use nqp; my num $n = nqp::time_n(); my $i; $i := ($n * 1000).Int for ^100_000; ||| use nqp; my num $n = nqp::time_n(); my $i; $i := ($n * 1000e0).Int for ^100_000;;
benchable6 MasterDuke, starting to benchmark the 1 given commit
MasterDuke, gist.github.com/27ebba85f0a5cf5752...fdc027241f
MasterDuke yeah, num literal is a bunch faster 19:59
codesections MasterDuke++ that makes sense. I didn't know our literals were that smart. 20:00
tonyo_ is benchable6 using Bench ? 20:03
MasterDuke yeah 20:04
github.com/Raku/whateverable/blob/...ble.p6#L66 20:05
tonyo_ nice 20:39
rba codesections: afaik there are mainly the images cached on the cdn side. So far the traffic to the server itself haven't been an issue. 20:45
codesections rba: mainly the images on the docs site? So, pretty much just Camilla, then? :D 20:55
or, wait, does that include the "Type Graph" SVGs? Because those don't load correctly for me about 30% of the time; if they were on the CDN, I suppose that could be related… 20:56
rba codesections: So, as the CDN is turned of atm, you might give me some feedback in the next days? 22:33
codesections I'd be happy to. But (and this is part of what makes me wonder if the issues could be related) as with the issue tbrowder mentioned, I'll have long-ish periods where the type graphs display fine, followed by a few days when they consistently don't. So if I don't see any issue in the next days, that won't really prove that it was the CDN. But I'll keep you posted 22:38
codesections (I also can't remember the last time the type graphs didn't display; I don't pay a whole lot of attention to them most of the time) 22:39
tbrowder i'm trying to extend DateTime with a new "new" method by inheriting from and adding the new "new" but not having much luck so far. 22:41
after getting to see the docs again, i don't see a clear example of that situation, so i filed a doc issue. 22:43
i'll try to start with a simpler case for now and see if i get that working. if i can't, i'll post some code for help. if i do get it working, i will take ownership of the issue. 22:45