01:29 Kaiepi left 01:34 Kaiepi joined
Geth_ roast: SirBogman++ created pull request #598:
Added test for P5 regex conditional expressions.
rakudo: vrurg assigned to niner Issue make test doesn't pass t/04-nativecall/20-concurrent.t github.com/rakudo/rakudo/issues/3311
vrurg++ created pull request #3317: Allow colon syntax on $.method
03:41 nativecallable6 left, statisfiable6 left, coverable6 left, greppable6 left, quotable6 left, reportable6 left, bisectable6 left, releasable6 left, notable6 left, benchable6 left, committable6 left, shareable6 left, squashable6 left, unicodable6 left, bloatable6 left 03:42 bisectable6 joined, releasable6 joined, bloatable6 joined, coverable6 joined, nativecallable6 joined 03:43 notable6 joined, shareable6 joined 03:44 reportable6 joined, squashable6 joined, statisfiable6 joined, greppable6 joined, benchable6 joined, unicodable6 joined 03:45 quotable6 joined, committable6 joined
¦ rakudo: vrurg self-assigned "colon syntax" on $. (e.g. self) in a method does not work github.com/rakudo/rakudo/issues/3306 03:59
vrurg .ask jnthn Also, would you mind enabling `$.method: <arguments>` call syntax? It's pretty simple grammar fix in R#3317 github.com/rakudo/rakudo/pull/3317 04:07
tellable6 vrurg, I'll pass your message to jnthn
04:47 evalable6 joined
Xliff What's $*QSIGIL? 05:27
.tell vrurg What's $*QSIGIL? 05:29
tellable6 Xliff, I'll pass your message to vrurg
09:02 jmerelo joined
jmerelo The "old" perl6 documentation pages now include a notice and link to the new ones docs.perl6.org/ (I guess it'll take a while to propagate to the CDN) 09:03
They also say "Raku documentation" in the main page. Since that's hard-coded into Documentable, I kind of couldn't help it. Can change it by hand if needed, though.
Will do it, indeed.
Ah, OK, now it works. 09:06
Just a reminder: for some time, docs.perl6.org will be the "old" pages (frozen sometime in October) with a link to equivalent page in the new documents, which are hosted in docs.raku.org 09:07
Also, now some search like "raku twigils" return the document pages in the 4th position www.google.com/search?client=ubunt...p;oe=utf-8 09:08
I have been editing some stackoverflow answers to the new URLs; please purposefully add links to the new documentation there (and elsewhere) and try to avoid links to the old one.
This will stay that way for some time, until search engines pick everything up correctly; then we will redirect old pages to new ones. 09:09
rba jmerelo++ 09:11
jmerelo Please also note that the "old" docs worked with "doc" and "docs" indistinctly, with "doc" being the preferred spelling. Now only "docs.raku.org" is available
rba: BTW, docs.raku.org points to the "old" Altai-man server. I have updated it already in the new docs.perl6.org. 09:12
rba: I seem to remember you asked me to stop using that interim server that was set up; I think we can start using the new one from now on. 09:13
rba: sorry, I meant docs.perl6.org above.
rba I see. docs.perl6.org -> interim server, docs.raku.org -> hosting server 09:20
jmerelo: In case you like to move docs.perl6.org to the hosting server as well, let me know. I have to make sure let's encrpyt is creating new certificates asap. 09:21
jmerelo rba: yes, please. Content has already been copied there.
rba: I mean, when you can. 09:22
rba jmerelo: Sure :-) ready in front or my notebook...
jmerelo rba++
rba jmerelo: Q: Do we like to redirect doc.perl6.org -> docs.perl6.org? Or just server under both names? 09:23
jmerelo rba: There are a gazillion links to doc.perl6.org. I don't think there's a difference, SEO-wise. Probably redirect will save you generate a new certificate, right? 09:25
rba jmerelo: This is not a question of the certificate, more a tatical one. I propose to do a redirect. And in the meantime I will add doc.raku.org -> docs.raku.org as well. Seems to be forgotten last time... 09:26
jmerelo rba: OK. About doc.raku.org... Maybe we shouldn't
rba jmerelo: Ok... 09:27
jmerelo rba: it's generated enough confusion. Also, we don't want to split links in two domains for the time being. Let's just use a single one, and that will be docs.raku.org
09:27 ExtraCrispy joined
rba jmerelo: So I will NOT add any doc.raku.org. So DNS is telling anybody this hostname does'nt exists. Ok too. 09:28
jmerelo rba: right. Thanks!
09:30 ExtraCrispy left
rba Ok. Did the switches and invalidate CMS cache for perl6.org. 09:44
jmerelo: May you please double-check from your side?
jmerelo rba: OK 09:51
docs.perl6.org looks fine 09:52
doc.perl6.org too.
rba Hmm. Something went wrong. docs.perl6.org/type/Any is not showing content any more.
jmerelo rba: that's the nginx hack
rba: we've got to add this to the configuration github.com/perl6/doc#nginx-configuration 09:53
I can do it if you want.
rba jmerelo: fixed
jmerelo rba++
rba: we should really have a way to do all this automatically.
rba jmerelo: Question for me would be if you like to more prominent put the deprecation message with the link to docs.raku.org on top of each page? 09:54
jmerelo: Yes. I we started already working on this. 09:55
jmerelo rba: they're not so much intended for human consumption as for search engine spiders...
rba My plan is to have every website run in it's own docker container.
jmerelo rba: tht would be great :-)
rba With a traeffik in front.
This would allow us to setup all the sited redundant as well. 09:56
jmerelo rba: nginx would be inside the container then?
rba Hopefully the Cloudflare DNS then doing what I expect and just route it to the next backend server.
jmerelo: Nginx using http or the site which are using Mojolicious would be doing the serving of the page. 09:57
jmerelo rba: not mojolicious, please
rba jmerelo: traeffik will do https->http.
jmerelo: Mojo is still part of some pages...
jmerelo jmerelo: I don't really know what it's doing. 09:58
rba marketing.raku.org
jmerelo But One can only shave yaks so far. It's no longer needed in docs, at least.
rba and modules.raku.org
09:59 sena_kun joined
rba I think in modules is stil necessary. 09:59
nine Why don't we redirect doc.perl6.org to docs.raku.org right away?
jmerelo rba: so some containers will serve using Mojo and some of them will use nginx, right?
rba jmerelo: Yes, this is the idea.
10:00 ExtraCrispy joined
jmerelo rba: great! 10:00
rba jmerelo: Then using docker-compose try to put it together.
jmerelo: So for I'm still in the trial phase...
jmerelo: I hope I have more time in xmas holidays for this... 10:01
jmerelo nine: it would not be good for visibility either of Perl 6 or of Raku. We've created this roadmap github.com/perl6/doc/issues/3046
rba: so, not kubernetes for doing automatic scaling and stuff.
rba jmerelo: I'm interessted in Kubernets. Yet, not sure if this overcomplicated the things. 10:02
jmerelo: Open for some guidance from your side...
jmerelo rba: I guess that as long as you're offloading scaling to the CDN it's OK.
10:03 Altai-man_ joined
jmerelo rba: totally blushing by this. I'm just an amateur, you're the pro. 10:03
rba: containers are perfect, because we can deploy them anywhere if we so want. Heroku, Amazon...
rba jmerelo: Not a pro in this area yet...
jmerelo rba: speaking of which... Any news from Amazon? 10:04
nine jmerelo: "We need the two sets of pages for the new ones to gather SEO juice" that's not correct. Search engines would follow the redirects. One can also accelerate the process for Google by adding the new domain in the search console.
rba jmerelo: No, just an automated answere that they are reviewing it.
jmerelo nine: they would, and them slam down the results in search queries because "perl 6" would not be in any place in these new pages.
rba jmerelo: So far we just defined some base images: github.com/rba/docker-raku-infra
10:05 sena_kun left
jmerelo rba: looking good :-) If I have some time, I'll check them out. Good call, using Alpine. 10:05
rba jmerelo: Don't look to close yet. There are some placeholder Dockerfile. 10:08
jmerelo rba: I just noticed :-) The intention is clear, though :-) You might want to check out ubuntu 18.04, too. It's got very small footprint, and you get to use good old Ubuntu. 10:09
rba jmerelo: I'm more familiar with Ubuntu. Some people told me it's easiere to produce small images using Alpine. So far I'm still open... 10:10
jmerelo rba: the thing is, from 18.04 Ubuntu made their "ubuntu-minimal" image the default, and it's _really_ small. Eventually, you might want to check both and pick the smaller one. 10:11
rba jmerelo: Ok. 10:12
jmerelo: There are still many open question. Eg. if we should allow cron based github pull inside the image or rather let CI do the rebuild of the image.
10:13 ExtraCrispy left
jmerelo rba: I would definitely go for CI. The thing is, we would need to set up hooks in every repo. Probably not a bad thing, though. 10:13
rba jmerelo: Thinking as well if I like to give azure pipelines a try: blog.urth.org/2019/11/18/my-new-ci...-for-perl/ 10:16
jmerelo: Are you now beeing able to doing delta builds for docs? Or how did you solve the long build issues? 10:17
jmerelo rba: it's probably easier to set up GitHub Actions. You can even host the images there github.com/perl6/doc/issues/3046
rba: yes; the new Documentable takes a few minutes to regenerate just the pages needed 10:18
rba jmerelo: Very nice.
jmerelo rba: a full build takes between 350 and 450 seconds, an update takes IIRC a fifth of that 10:19
rba: it's still got some issues, like generating the type graphs every single time... but it's a great improvement over the 15 minutes it took every single time before.
rba: check out my immensely useful Daleksay (like cowsay, but with daleks) container hosted in GitHub github.com/JJ/docker-daleksay/packages 10:21
rba: I've still got to get the hang of using github actions to create it, but it'll arrive eventually. I use them for other purposes, anyway. 10:22
rba: do you have to pay for Azure Pipelines? Or host the stuff in Azure too? 10:23
rba: I'm going to create an issue in your repo, if you don't mind, mainly to remind myself to do that 10:24
rba jmerelo: LOL
azure.microsoft.com/en-us/services...s/?nav=min 10:26
10 free paralell jobs for open source projects... 10:27
jmerelo rba: ah, that's cool :-)
rba jmerelo: Just to understand. The github action is doing the docker build an then the docker registry on github is used to make it available? 10:31
jmerelo rba: right 10:32
rba: isn't it sweet?
rba jmerelo: yes, it is.
jmerelo rba: so fat I've done it separately, I'd still have to streamline it in a single action. Shouldn't be too difficult, though. 10:33
rba jmerelo: Thanks for the talk. I think I will be mostly afk for the rest of today... 10:40
jmerelo rba: have fun :-) 10:44
nine jmerelo: thanks for enlightening me :) 10:54
10:58 ExtraCrispy joined
jmerelo nine: the issue is open for comments, anyway. Any suggestion is always welcome. Help, too :-) 11:29
12:04 sena_kun joined 12:05 Altai-man_ left 13:05 jmerelo left 13:39 Guest38485 left 13:46 Guest38485 joined 14:02 Altai-man_ joined 14:05 sena_kun left 15:34 lucasb joined
Geth_ nqp: 58996e370a | (Stefan Seifert)++ | tools/templates/MOAR_REVISION
Bump MoarVM to get NativeCall negative number fix
rakudo: fc30941910 | (Stefan Seifert)++ | 3 files
Test JIT compiled native calls returning negative numbers
nine One of 3 issues I uncovered when trying to use the JIT compiled native call even on the first call (where we currently use the old "interpreted" way) 15:50
dogbert17 Hmm, is the build broken again 15:53
ogbert@dogbert-VirtualBox ~/repos/rakudo $ make 15:54
Makefile:962: *** missing separator. Stop.
looks like a missing tab 15:57
nine Wasn't there a fix for something like that today or yesterday?
dogbert17 yep, and it's busted :( 15:59
at least on my system 16:00
dogbert17 wipes the rakudo install 16:02
16:03 sena_kun joined 16:05 Altai-man_ left
lucasb regarding R#3316, I tried search where that P5 extended pattern (?...) was recognized in github.com/perl6/nqp/blob/master/s...rammar.nqp , but couldn't find the spot 16:37
is that the right place to look for the code that implements the rx:P5/.../ regex flavor, right? 16:38
IOW, Raku's "P5" regex flavor doesn't implement full compatibility with Perl 5's regex (?) 16:40
### Perl 5 regex conditional expressions not working. #3316 github.com/rakudo/rakudo/issues/3316
also, a time to remind of R#2624 16:43
github.com/rakudo/rakudo/issues/2624 - Should we remove/deprecate :P5 regexes? 16:44
(where's the "synopsebot" of the issues numbers?) 16:45
AlexDaniel lucasb: it'll come back soon, I have a rewrite 16:50
17:33 Xliff left
lizmat lucasb: FWIW, I think the rx:P5 implementation is from Perl 5.8 or thereabouts 17:48
functionally speaking
lucasb I may have misunderstood something. Can't say if Raku supports this feature or not 17:58
m: say <a b c>.map: { so rx:P5/(?(1)a|b)/ }
camelia (False True False)
lucasb there appears to already exists some tests for it in S05-modifier/perl5*.t files 17:59
I was expecting rx:P5/(?(1)a|b)/ to be equivalent to rx:P5/a/ but it ain't 18:02
18:02 Altai-man_ joined
lucasb reference: perldoc.perl.org/5.30.0/perlre.htm...-pattern_) 18:03
lol, I should have read the docs. Disregard my interpretation, I'm wrong :) 18:05
18:05 sena_kun left 18:29 lizmat left 18:30 lizmat joined 19:40 lizmat left 19:42 lizmat joined 20:03 sena_kun joined 20:06 Altai-man_ left 21:09 patrickb joined
AlexDaniel damn I'm so tired today, will try to get the release going tomorrow, sorry 21:12
lizmat AlexDaniel:hope to be able to announce on the RakudoWeekly tomorrow... no pressure :-) 21:14
AlexDaniel lizmat: we'll see how it goes but it's probably going to happen later in the day :( 21:15
lizmat I can postpone a bit into the evening :-)
AlexDaniel sounds like a plan :) 21:19
Kaiepi what sort of rules are there wrt closures in nqp? 21:31
i tried moving some code in a method of a role to a sub in the role and calling that from the method but it threw an error saying it couldn't get its static coderef 21:33
though ig it should be in the method since no other method uses it
vrurg AlexDaniel: BTW, testing the build on FreeBSD and OpenBSD. Seems ok so far. 21:43
tellable6 2019-11-24T05:29:59Z #raku-dev <Xliff> vrurg What's $*QSIGIL?
AlexDaniel nice
vrurg Running spectests now, just in case. 21:44
21:55 robertle left 22:03 Altai-man_ joined 22:05 sena_kun left
Geth_ roast: 3aa913b8b8 | (Vadim Belman)++ | S10-packages/precompilation.t
Unfudge a TODO now passing on moar

Fixed by MoarVM/MoarVM#1209
vrurg Xliff: $*QSIGIL is coming from Perl6::QGrammar. Simply put, it's often tested to see if we're in a quotation context. 22:34
tellable6 vrurg, I'll pass your message to Xliff
vrurg tellable6: thanks!
tellable6 vrurg, It's my pleasure!
vrurg Nice to talk to a polite... er... entity. 22:35
MasterDuke huh, wish i'd know about the `--jobs=<n>` option to `git pull` a while ago 22:39
patrickb . 22:46
tellable6 2019-11-23T09:54:39Z #raku-dev <nine> patrickb: please don't change release file names just for esthetic reasons. There are scripts used by packagers that expect a certain naming scheme
23:04 Altai-man_ left
patrickb I have arrived at a checkpoint for my rakudo.org refactor. 23:04
rba: Now I'm waiting for the deployment for wider testing. I'm not in much of a hurry though. So don't feel pressured, please. 23:06
rba patrickb: not helpful :-) 23:07
patrickb nine, AlexDaniel: I'll leave the rakudo source releases alone, for good reasons. But I'm still thinking about renaming all of the rakudo-star packages. What's your opinion on that? 23:10
Is the idea to add 'perl6' as a synonym to the 'raku' tag on stackoverflow still a thing? From what I understand this is what should be done... 23:12
23:20 lucasb left 23:33 entonian joined 23:35 entonian left 23:39 ExtraCrispy left
patrickb o/ 23:48
23:48 patrickb left 23:55 Kaiepi left 23:58 Kaiepi joined