Geth ¦ problem-solving: autarch assigned to rba Issue Need a "Raku" download button/link for US Patent on Raku 01:55
04:25 HarmtH left 04:27 guifa2 joined
guifa2 .tell lizmat I finally figured out what was going on with the slang. Or at least, a better idea of the problem. The problem wasn't in the grammar or the actions, but rather the call to HLL::Compiler.lineof(). Commented it out and violà I could declare methods with other syms 04:31
tellable6 guifa2, I'll pass your message to lizmat
06:27 jjmerelo joined 08:20 sena_kun joined
Geth ¦ problem-solving: AlexDaniel unassigned from rba Issue Need a "Raku" download button/link for US Trademark on Raku 08:21
¦ problem-solving: AlexDaniel assigned to jnthn Issue Need a "Raku" download button/link for US Trademark on Raku
sena_kun .seen patrickb 08:34
tellable6 sena_kun, I saw patrickb 2020-05-08T19:00:28Z in #raku-dev: <patrickb> vrurg: OK. Then it's already fine. Thanks!
sena_kun AlexDaniel, ping? 08:37
AlexDaniel pong
sena_kun AlexDaniel, hi. In your opinion, is worth a point? 08:38
AlexDaniel sena_kun: how many relocatable releases did we have? :) 08:39
jjmerelo sena_kun (my own unrequested opinion). Yes, so that a rakudo star distro can be made out of it (for Windows)
AlexDaniel latest star release wos 2020.01 08:40
sena_kun Hmm, is it time for rakudo star already...
AlexDaniel was*
so yeah, I'm also wondering if anybody is going to make star out of 2020.05
lizmat hopes tyil will find the time 08:41
tellable6 hey lizmat, you have a message:
sena_kun By the way, there wasn't an email announcement as I waited for builds, d'oh. 08:42
tyil I didn't even know there was a 2020.05 for rakudo
AlexDaniel sena_kun: I think yeah, probably, if you have the time
sena_kun I seem like the worst person for this position.
AlexDaniel sena_kun: quite the opposite actually 08:43
jjmerelo sena_kun you're doing fine :-) Kudos.
tyil I was originally waiting for a rakudo 2020.04 to make a star bundle out of that
AlexDaniel humans generally suck for work like this, which is why we should automate most of it. The Sakefile is meant to send emails but I never managed to find time to implement it
then I also made undersightable6 that can do more checks 08:44
which later bitrotted and doesn't even start now
so in the absence of full automation we just need someone who is responsible, and I can't think of a better person than you, sena_kun 08:45
sena_kun Ok, let's do some things then. 1)I'll post (sadly delayed) email announcement; 2)I'll intimidate someone here to add us tests for relocatability making it a next release blocker; 3)Do a rakudo point.
jjmerelo (y) 08:46
AlexDaniel number 2 is actually very important, a point release by itself doesn't really fix the underlying issue (which is us not being able to detect relocatability regressions)
lizmat: can you perhaps leave a comment on PS#193 saying what you think? 08:49
linkable6 PS#193 [open]: [language] Need a "Raku" download button/link for US Trademark on Raku
AlexDaniel lizmat: am I just out of the loop or something? 08:50
tyil AlexDaniel: is there going to be a point release soon or can I build a rakudo-star from rakudo 2020.05?
AlexDaniel sena_kun: ↑
sena_kun tyil, there will be a point release today which fixes Windows.
tyil alright 08:51
hit me when it's done, I'll start work on star
sena_kun tyil, in a couple of hours depending on how trash my morning will be today. :)
tyil, will do, thanks!
tyil Im here all day :)
so no need to hurry on my part 08:52
nine sena_kun: will the point release change anything for Linux?
sena_kun nine, no.
nine, or, ahem... I want to cherry-pick
AlexDaniel tyil: what's the difference between and 08:53
tyil AlexDaniel: rakudo/star is no longer maintained as far as I'm aware 08:54
[Tux] Rakudo version 2020.05-55-g7a902b5e4 - MoarVM version 2020.05-10-g5fe4a8114
csv-ip5xs0.830 - 0.850
csv-ip5xs-208.157 - 8.191
csv-parser23.557 - 24.160
csv-test-xs-200.383 - 0.390
test7.281 - 7.481
test-t1.882 - 1.966
test-t --race0.977 - 0.982
test-t-2031.135 - 31.952
test-t-20 --race9.113 - 9.982
nine sena_kun: I'd guess this affects all platforms. But it's irrelevant for me as distro packages don't build relocatable anyway 08:56
sena_kun Oh, it makes sense, that's why we didn't spot in with El_Che. 08:57
08:59 jjmerelo left
AlexDaniel tyil: hmmmm… okay, okay. And now what? 09:03
tyil who knows
not sure what you want to hear
you want me to stop working on rstar because Im not pushing to the correct github repo?
AlexDaniel tyil: we can merge two repos so that the latest state is the one you have 09:04
tyil: this will preserve tags and everything, with the only downside of having two git roots
tyil mine is a complete rewrite, from scratch
AlexDaniel I understand 09:05
but it's still rakudo star isn't it :)
tyil I could've called it anything else, I only kept the name because that's what people expect it to be called
AlexDaniel so, any reason not to do the git trick? 09:06
tyil I don't see any added value 09:07
AlexDaniel I'm not okay with two repos. It's confusing 09:08
tyil feel free to purge the repo from the Raku namespace if it's too confusing
I'm just working on rstar because I like doing so 09:09
AlexDaniel if I merge them for you, will you be use the main repo?
tyil the "main repo" is in ~/.local/git/rakudo-star for me
I'm not sure what you understand as "the main repo" 09:10
as said in the github issue I linked, I'm only putting it up on GitHub for ease of *other* people looking for it
AlexDaniel but you do keep it in sync, yes?
tyil not rakudo/star, I don't think I have push access there to begin with 09:11
AlexDaniel well I can resolve that
tyil I keep my version in sync with the mirrors I have access to
AlexDaniel okay. I'll do the thing and will ping you later
tyil I don't think rakudo/star is the best location for it anymore anyway, since it's not solely a Rakudo project 09:12
AlexDaniel that I agree with
tyil I'm not even part of the rakudo team, nor is hankache as far as I know
AlexDaniel I'll see if there's anything I can do about it
I mean, push access isn't a problem 09:13
09:17 jjmerelo joined
nine tbrowder: 09:58
Geth roast: c2666ff0c3 | (Stefan Seifert)++ | S03-metaops/misc.t
Fix S03-metaops/misc.t depending on Rakudo

The rakudo specific Test::Helpers module is not actually needed for this test. Remove it to fix the test failing when run outside of rakudo's source directory.
roast: niner assigned to dwarring Issue Roast depends on Rakudo when it shouldn't
niner assigned to tbrowder Issue Roast depends on Rakudo when it shouldn't
tbrowder erg. i'll take a look, but before i do, can we not put that helper in roast? 10:09
nine Looks like it's only used by that one spec test file 10:11
tbrowder ok 10:14
10:16 Altai-man_ joined 10:18 sena_kun left
tbrowder since the existing test is for a rakudo env var triggering a grammar action, where should the test really be? probably with rakudo for now maybe? i want to make it triggered through a mod to the run options but i haven't solved that yet. 10:27
nine tbrowder: yes, that test doesn't belong in roast at all 10:30
Even if we ignore the environment variable, the test poses a question: are compiler arguments part of the language specification? 10:31
tbrowder but even with the run option that's still rakudo isn't it? yes, that's my question, too.
imho you could make a case for either. most c compilers i know have similare run options. why could they not be standardized? 10:33
*similar 10:34
nine It would require a language implementation to provide a command line compiler in the first place. I'm not sure there's consensus for that for our 100 year language
tbrowder this metaphysical question is way above my pay grade! 10:35
at this point isn't it "once in roast, forever in roast?" 10:37
nine Only if it's part of an actual language release. And even there' there's -errata branches for when we were just wrong when writing the test 10:39
Altai-man_ tbrowder, as long as it is not frozen in 6.*-errata branch tests can be adjusted.
AlexDaniel tyil: ok, can you take a look at this?
tyil: if it looks fine I can transfer it to rakudo-star
tyil: I'll have to delete it first, but it shouldn't affect you in any way. On your side a simple `git pull` will make it in sync 10:40
tbrowder is the grammar raku or rakudo? 10:41
tyil ugh, now it has all the submodules again 10:46
tbrowder raku i think, but it gets modified on the rakudo side during compilation doesn't it?
AlexDaniel tyil: sorry, where are you seeing this? 10:47
jjmerelo tbrowder I have just moved the routine, and done a PR for it. 10:48
tyil after I pulled in rakudo/star, when I try to pull, it's warning me about all the old submodules that I never used in my version
tbrowder so what should i do? if left to me i will try to move the helper into roast.
jjmerelo tbrowder check out my PR 10:49
AlexDaniel tyil: how does the warning look like?
tyil warning: 5a9358143a4a4c209e0cdc13e4b3ed555696b5ee:.gitmodules, multiple configurations found for 'submodule.modules/file-find.url'. Skipping second one!
tbrowder thnx jj (i'm working on reply to yr kind email)
jjmerelo (y) 10:50
Altai-man_ >Users of other platforms or those who don't need relocatability should continue to consider Rakudo 2020.05 release to be latest stable. 10:51
Sounds sane?
*the latest stable
AlexDaniel tyil: interesting. I tried pulling the merged repo from your repo and didn't see any warning. Tried it again and still don't see it, hmmm. I think after the pull you should be fine given that there are no submodules in master? 10:52
tyil: well, I mean not anymore on HEAD
tyil I don't know, but I'm not going to spend my afternoon figuring out why `git push all` is spewing out 1500 warnings
Geth roast: 51f90e3e88 | (Juan Julián Merelo Guervós)++ (committed using GitHub Web editor) | S26-documentation/block-leading-user-format.t
Fix external dependency (#641)

By putting it in here. Closes #640.
linkable6 ROAST#641 [closed]: Fix external dependency
ROAST#640 [closed]: Roast depends on Rakudo when it shouldn't
jjmerelo Er... Maybe I shouldn't have said close
Let me check
AlexDaniel tyil: so Raku/rakudo-star just lost a commit, right? 10:56
lizmat tyil: I think the new module bot is AWOL
tyil I just added
tyil lizmat: I kicked it 10:57
AlexDaniel ah, even two
tyil should be back soon, I'll double-check the processes in charge of checking for updates
AlexDaniel 7159e581905 and 9600d177c
linkable6 (2020-05-10) Add GitHub mirror to README
(2020-04-19) Update list of repository mirrors in README
tyil ye, I reset --hard back to my local version and pushed that 10:58
I'm not going to look at over a 1000 lines of warnings every time I push
AlexDaniel I have my limits, you know 10:59
tyil so do I
the only reason I'm getting your complaints is because I made a repo for it in the new namespace, something I talked about about 3 weeks ago 11:00
if I would not try to include the people on github, I wouldn't have had any complaints
AlexDaniel or vice versa, if you just pushed to github it would've been easier 11:01
tyil not really
because I dont have push access there to begin with
AlexDaniel I've tried multiple git remotes thing in the past, it doesn't really work
tyil nor do I like github, because its a terrible centralized "service"
AlexDaniel well, I mean, it does, but you need to actively maintain it
tyil could also just not have a rakudo star at all, of course
AlexDaniel and what's the problem with access? It's a matter of simple “hey can I have access to …” message here 11:02
I never said no to anyone
tyil because I'd have to sign some DCO or whatever, and again, I dislike github because it's an atrocious platform
AlexDaniel so when you see the warnings, do you have a .gitmodules file? 11:03
tyil idk 11:04
AlexDaniel can you please check?
tyil I reset --hard the stuff back to just my own commits
doesn't look like it 11:05
Geth roast: JJ++ created pull request #642:
Fixes integration test
rakudo/2020.05.1-release: fe14bcb7a9 | (Patrick Böker)++ (committed by Altai-man) | tools/lib/NQP/Config/
Fix relocatable builds on the moar backend

For relocatability to work the config variable `static_rakudo_home` needs to be empty. It defaulted to the build time rakudo home path globally. Later on a MoarVM specific config variable with the same name was set to empty assuming it would overwrite the global variable. But that's not the case. In backend specific templates the empty ... (10 more lines)
rakudo/2020.05.1-release: 0d238d89e8 | Altai-man++ | docs/release_guide.pod
Add point and fix next release date
rakudo/2020.05.1-release: 40fdfee1de | Altai-man++ | 2 files
Changelog + release announcement
AlexDaniel tyil: hmm, maybe you can just ignore the warnings? You wouldn't see them later because there are no submodules in master anymore
tyil ...
AlexDaniel I'm curiouus why you saw them anyway, because I definitely can't reproduce it 11:08
tyil I should bend over backwards to please you, but not the other way around
got it
reverting back to just my commits
feel free to just remove the raku/rakudo-star repo in that case
AlexDaniel lizmat: I'll let you take it from here
tyil so we can stop this
lizmat: bot logs indicate it should've announced your Rake module 11:16
lizmat 0.02 ? or 0.01 11:18
Geth rakudo: Altai-man++ created pull request #3680:
2020.05.1 release
Altai-man_ AlexDaniel, if you're around... It is true I can just prepare rakudo archive for the point branch and no need to make a nqp point if only rakudo was changed, right? 11:39
AlexDaniel Altai-man_: interesting question 11:43
Altai-man_: you don't need a release for NQP, that's right 11:44
Altai-man_: ah, I see
Altai-man_: there's `all-rakudo-only` target in the sakefile 11:45
Altai-man_ AlexDaniel, yes, but it has `sake nqp` target as well as `human-2` which includes `nqp-sign rakudo-bump-nqp rakudo-bump` which seems redundant to me. 11:46
oops, rakudo-bump is nice, of course. 11:47
AlexDaniel Altai-man_: yeah but
Altai-man_ Right.
AlexDaniel so yes it will run those steps and will create a tar for NQP :D
but you can just discard it 11:48
Altai-man_ will just go for rakudo tar
Ok, thanks.
AlexDaniel yeah, and normally I'd double check the generated commits and stuff to make sure it did what I asked it to do 11:49
but I definitely used this target before, so it should work :)
Altai-man_: btw later people will ask why there's rakudo .1 but no nqp or moar point release
which is fine :)
Altai-man_ AlexDaniel, better than questions like "So you had a point, then delayed a release and now you have another point, what's your CV again so we'd blacklist it in our company?". :P 11:52
AlexDaniel Altai-man_: do you know about 2017.04.3? 11:53
Altai-man_ yesh
AlexDaniel the one after which we said “ok it is still broken but we're not doing more point releases this month” 11:54
AlexDaniel &
Altai-man_ o7
Thanks for your help!
12:07 MasterDuke left 12:17 sena_kun joined 12:18 jjmerelo left, Altai-man_ left 12:33 MasterDuke joined
Geth rakudo/2020.05.1-release: 002acb1be2 | Altai-man++ | VERSION
[release] Bump VERSION to 2020.05.1
rakudo/master: 5 commits pushed by (Patrick Böker)++, Altai-man++ 12:52
sena_kun tyil, 2020.05.1 release tarballs are released. Thank you very much for working on star so far! 12:53
tyil sena_kun: ack, I'll start on the 2020.05.1 star bundle 13:00
sena_kun tyil++
tyil sena_kun: 13:07
I'm going to install it on all my machine and wait for gitlab-ci's test results, if those are all acceptable I'll announce it to perl6-users for testing for a week 13:08
if no issues pop up, I'll drop the -rc1
and announce it as a proper release
I'll send an email to hankache too for the windows build 13:09
sena_kun tyil, thanks! Is there anything I should help with from now on? 13:22
tyil download, install and see if anything breaks in your standard workflows, or module tests
timotimo oh hey tyil, can the irc/discord bridge be kicked again? i think my access isn't working right at the moment 13:53
tyil timotimo: sure
13:54 discord6 joined, discord6 left, discord6 joined
El_Che dot releae out? time to release pkgs :) 14:05
sena_kun El_Che, not necessary for non-windows. 14:10
And given Linux packages are not very helpful on Windows... :) 14:11
El_Che sena_kun: relocatable is also broken on linux 14:14
sena_kun El_Che, what, what?
El_Che I wrote it above
let me look it up
ttps:// 14:15
I looked up the Windows issue, did the same test en relocable linux binaries
14:16 Altai-man_ joined
El_Che I could test it because besides native pkgs, rakudo-pkg also has compiled tar.gzs for amd64 and i386 14:16
14:18 sena_kun left
El_Che Altai-man_: about releases. My 2c that is easier to have one version marked as the latest (the latest dot release) instead of a matrix per platform OS. Builds are automated anyways 14:18
Altai-man_ El_Che, sorry, what are you referrying to? 14:19
El_Che, you mean on the website or? 14:20
El_Che Altai-man_: 14:21
Altai-man_ I think the point of this line in announcement was to not bother users who don't have a breakage with "Huh, do I need to update again?". If we didn't have a general breakage, I mean. 14:22
El_Che ok 14:23
I am now building relocatable on Linux to see if the problem is fixed there as well 14:53
it should be done soon
done. The problem is fixed for relocatable linux as well. 15:00
.tell patrickb: The problem is fixed for relocatable linux as well. patrickb++
tellable6 El_Che, I'll pass your message to patrickb
15:39 jjmerelo joined 15:45 jjmerelo left 15:46 jjmerelo joined 15:52 AlexDaniel left 16:02 lichtkind joined 16:16 sena_kun joined 16:18 Altai-man_ left
timotimo lizmat: do you think native array's reverse function should return a Seq with intposref so that assigning into it changes the original like it happens for regular arrays? i think so. there'll be a not so great performance hit of course :< 17:11
lizmat timotimo: that's the semantics that we have for non-native arrays as well 17:15
m: my @a = ^10; my int $b = 42; $_ = ++$b for @a.reverse; dd @a
camelia Array @a = [52, 51, 50, 49, 48, 47, 46, 45, 44, 43]
timotimo right, that's why we should make it so 17:16
will just be much more expensive than just cloning the int_array or whatever and pushing values directly 17:17
lizmat well... making it return containers makes for different versions between List and Array
timotimo i ran into it while implementing .reverse for the hyper codegen optimization 17:18
lizmat so personally, I wouldn't mind if .reverse (or .rotate for that matter) would decont
timotimo we don't have native lists, so that shouldn't be an issue?
oh, you mean they should all behave like list does 17:19
lizmat well, yes, but that would break some spectests 17:20
17:23 El_Che left
lizmat do we have a way to change an `nqp::list` to an IterationBuffer ? 17:50
they are both BOOTArrays afaik
and both repr("VMArray") 17:51
18:04 jjmerelo left 18:16 Altai-man_ joined 18:18 sena_kun left
timotimo lizmat: maybe a nqp::rebless can do it? but i don't know if that works for any repr besides p6opaque 18:21
and BOOTArray is at the same "level" as IterationBuffer i believe, i.e. both are types
18:22 discord6 left
lizmat m: use nqp; dd nqp::rebless(nqp::list,IterationBuffer) # alas 18:22
camelia This REPR cannot change type
in block <unit> at <tmp> line 1
18:22 raku-bridge joined, raku-bridge left, raku-bridge joined
timotimo it would be incredible if we could rebless from an int64 list to an int8 list and back 18:31
Altai-man_ 18:35
timotimo so should i make a copy before doing the rest of the calculation 18:37
for now i just removed the support for reverse 18:46
no i'm not so certain what i should do next 18:51
MasterDuke nativecast it? 19:25
lizmat hmmm... looks like Geth is awol again 19:36
tyil ^^ 19:37
20:17 sena_kun joined 20:18 Altai-man_ left
Geth rakudo: ef0c457bab | (Elizabeth Mattijsen)++ | 10 files
Remove return constraint from identity methods

These would be the methods that always return "self". By removing this constraint, these methods became about 1.5x as fast. Now, they didn't do much to begin with, so in absolute numbers this probably won't matter much. But in some cases, such as Str.gist, this may actually be noticeable in some applications that do a lot of "say" on strings.
timotimo you can't nativecast a VMArray
rakudo: 88d98225a4 | (Elizabeth Mattijsen)++ | 5 files
Normalizing IO::Handle.say/put/print/note, part 2

The &say sub now performs all of the .gist + newline logic and then dispatches to .print. The advantage of this is that we do not need any additional logic / checks for modules that pretend to be $*OUT.
This also removes the IO::Handle.say(Str:D) candidate: even though a ... (12 more lines)
rakudo: 627b39b849 | (Elizabeth Mattijsen)++ | src/core.c/IO/Handle.pm6
Oops, forgot to remove a proto
rakudo: 438c1d2fa6 | (Elizabeth Mattijsen)++ | 4 files
Normalizing IO::Handle.say/put/print/note, part 3

Basically rince / repeat for .put, like in 88d98225a4
rakudo: 483cac3ab3 | (Elizabeth Mattijsen)++ | 2 files
Normalizing IO::Handle.say/put/print/note, part 4

Basically rince / repeat for .note, like in 88d98225a4, except that no changes were needed for IO::Handle, as note is just say, but writing to $*ERR instead of $*OUT.
rakudo: 2b38666f98 | (Elizabeth Mattijsen)++ | 4 files
Normalizing IO::Handle.say/put/print/note, part 5

Basically rince / repeat for .print, like in 88d98225a4
lizmat and that concludes my hacking for today 22:06
22:16 Altai-man_ joined 22:18 sena_kun left
guifa2 lizmat ++ 22:49
timotimo : re int64 to int8 um, yes. I'd love that.
(i'm assuming you mean int64.elems = 8 * int8.elems type coercion) 22:50
22:54 Kaiepi joined 22:56 lichtkind left 23:01 Altai-man_ left 23:58 raku-bridge1 joined