Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
00:21 anatofuz left, anatofuz joined 00:28 leont left 00:46 ZzZombo_ joined 00:49 ZzZombo_ is now known as ZzZombo
Geth_ rakudo: 1fa7ceddd4 | (Vadim Belman)++ | tools/templates/moar/Makefile.in
Fix a mistype in variable name

Did not rebuild ops dynlib.
00:59
01:00 anatofuz left 01:24 AlexDani` joined 01:25 anatofuz joined 02:23 moon_child left 02:25 anatofuz left, anatofuz joined 02:29 Elronnd joined 02:30 Elronnd is now known as moon_child, anatofuz left 02:41 anatofuz joined 03:09 anatofuz left, anatofuz joined 03:15 anatofuz left 03:25 anatofuz joined 05:06 anatofuz left 05:07 anatofuz joined 05:15 anatofuz left 05:24 anatofuz joined 05:55 ZzZombo_ joined 05:58 ZzZombo left, ZzZombo_ is now known as ZzZombo
[Tux] Yesterday: 06:07
Rakudo version 2019.07.1-395-g4df02facd - MoarVM version 2019.07.1-265-g95f69be84
csv-ip5xs0.787 - 0.833
csv-ip5xs-206.522 - 7.251
csv-parser22.525 - 23.268
csv-test-xs-200.424 - 0.442
test7.049 - 7.455
test-t1.748 - 1.769
test-t --race0.801 - 0.851
test-t-2029.339 - 30.533
test-t-20 --race9.372 - 9.470
Today:
Rakudo version 2019.07.1-396-g1fa7ceddd - MoarVM version 2019.07.1-265-g95f69be84
06:20 ZzZombo left 07:03 ZzZombo joined
lizmat Files=1274, Tests=109493, 207 wallclock secs (27.89 usr 8.05 sys + 2891.63 cusr 262.47 csys = 3190.04 CPU) 07:30
07:36 ZzZombo left
nine releasable6: status 08:09
releasable6 nine, Next release in ≈5 days and ≈10 hours. R6 is down. At least 6 blockers. Unknown changelog format
nine, Details: gist.github.com/22b742ad5dae8cfc59...e41b1d21c6
08:16 titsuki left, ZzZombo joined 08:17 ufobat joined 08:38 leont joined 08:42 leont left 09:25 anatofuz_ joined 09:29 anatofuz left 09:33 timotimo left 10:00 anatofuz_ left 10:02 anatofuz joined 10:07 anatofuz left 10:09 anatofuz joined 10:16 Guest3346 joined 10:22 anatofuz left 10:28 anatofuz joined 10:30 anatofuz left, anatofuz joined 10:32 anatofuz left 10:36 anatofuz joined 10:42 Guest3346 left 10:44 Guest3346 joined 10:45 anatofuz left 10:52 anatofuz joined 10:56 anatofuz left 10:59 anatofuz joined 11:03 Guest3346 left 11:07 anatofuz left 11:12 anatofuz joined 11:17 anatofuz left 11:18 camelia joined 11:19 ChanServ sets mode: +v camelia 11:30 AlexDani` is now known as AlexDaniel, AlexDaniel left, AlexDaniel joined
Geth_ problem-solving/master: 84 commits pushed by (Elizabeth Mattijsen)++, (Aleks-Daniel Jakimenko-Aleksejev)++, (Jonathan Worthington)++, (Vadim Belman)++
review: github.com/perl6/problem-solving/c...669a0d2f22
11:38
SmokeMachine \o/ 11:43
11:45 anatofuz joined
Geth_ rakudo: 6f086e65ee | (Aleks-Daniel Jakimenko-Aleksejev)++ | README.md
Initial Raku changes to the README

Other things will need to be adjusted later.
11:53
11:56 dogbert17 left, dogbert17 joined
AlexDaniel hm, why was it decided to keep NQP meaning something 12:11
that just makes any changes so much harder
AlexDaniel looks
lizmat: “It feels that this could stay” please read the nqp readme 12:13
lizmat: it suggests that it's Not Quite Perl 6 more than once
from what I remember, I said that I'd prefer NQP to mean nothing, so NQP is just NQP 12:14
lizmat that's my recollection as well
12:14 Guest13443 joined
lizmat so we don't need another acronym, 12:14
AlexDaniel “The acronym for NQP is Not Quite Perl” 12:16
from the PR
and it even goes to say “It stays that way, so no changes to documentation are needed” 12:17
Geth_ nqp: 2051967f25 | (Aleks-Daniel Jakimenko-Aleksejev)++ | README.pod
Some Raku changes

Still unclear what NQP acronym should mean, so further changes are needed.
12:18
AlexDaniel lizmat: see this ↑
lizmat AlexDaniel: sorry, I'm in the middle of what is going to be a ginormous Perl 6 Weekly, which is also quite emotional for me
could I think about that tomorrow? 12:19
AlexDaniel lizmat: that's alright, we now have all the time in the orld :)
w
lizmat: just a question though, will rba take care of the domain?
lizmat which domain? 12:20
AlexDaniel lizmat: currently linking to raku.org is a bit problematic because it doesn't redirect correctly
for example raku.org/community/ leads to the wrong place
so instead of fixing that, can we just go straight to raku.org doing the right thing, and perl6.org redirecting? 12:21
this is what rba should be able to do, from what I understand
or at least they should know who to ping to get it done
lizmat atm raku.org/foobar goes to perl6.org
AlexDaniel yep, but should go to perl6.org/foobar
MasterDuke didn't pmichaud express a strong desire for NQP to stay "not quite perl(6?)" 12:23
AlexDaniel ah, maybe that was the reason, yeah 12:24
rba Let me know if I can do something. I‘m not sure what would be the best to redirect vs. host on both names. 12:25
AlexDaniel maybe pmichaud can then tweak the README as they see fit
because right now it doesn't make much sense after my changes, and honestly it made even less sense before my changes 12:26
12:26 anatofuz left
AlexDaniel (considering the rename, of course) 12:26
rba: wasn't there a SEO issue with hosting on both names? 12:27
jnthn On NQP, I can also take a look, though not until later today (or mebbe tomorrow)
MasterDuke without really having a stake in the game or knowing how difficult the various options are, i would vote for whatever (redirect, host on both) has the least impact on anybody going to either (at least for the short term)
AlexDaniel rba: so what google for example sees is that someone (on raku.org) is simply copying the contents of a well-known website (perl6.org)
rba: and so they'd penalize raku.org, or at least that's my understanding 12:28
rba AlexDaniel: Yes, that‘s what jjmerelo said.
MasterDuke i.e., copying all content to both would be preferable to someone going to either and it not working
and i would vote for taking a short-term seo hit to make sure that content is available
12:28 anatofuz joined
rba The question for perl6.org would be anyway if we like to freeze the site as it is, make a clone of the repo and start changing there hosted as raku.org 12:29
MasterDuke i wouldn't want the news to get out and people try to visit a website (either, i don't care) and get 404s
lizmat I'd like to think about site issues tomorrow 12:31
AlexDaniel I personally see no problem just redirecting to raku.org
lizmat Also, I'd rather we work on a new raku.org site rather than just cat-licensing the current one
and when that new site is ready, *then* swap and make perl6.org a redirect to raku.org
AlexDaniel who's going to work on new site 12:32
and how long that'd take?
lizmat as long as it needs 12:33
12:34 Geth__ joined, Geth_ left
AlexDaniel this is just a test… 12:34
didn't work 12:35
12:38 anatofuz left
Geth__ geth/master: 10 commits pushed by (Patrick Spek)++
review: github.com/perl6/geth/compare/6b72...32d9da6bc8
12:41
12:47 Geth_ joined, Geth__ left 12:57 anatofuz joined
AlexDaniel ok now it works! :) 13:00
tyil \o/ 13:10
I'll let Geth exist in #perl6 and #perl6-dev for now, since the cost of being in a channel is pretty low: p
13:10 timo1 joined
tyil but someday I'll drop the #perl6-* channels in favour of the #raku-* channels 13:10
13:11 timo1 is now known as timotimo 13:15 lucasb joined
AlexDaniel tyil: my understanding is that soon there'll be a +f on this channel 13:21
tyil oh, then it'll solve itself 13:22
13:27 patrickb joined
patrickb o/ 13:27
tellable6 2019-10-11T23:21:18Z #perl6-dev <Kaiepi> patrickb, i don't think it does. you'd have to loop over the list of groups getgroups returns or st instead
13:29 Geth left
patrickb Just a quick question wrt the plan of execution for the rename. Do we try to do a centralized rename on one set date or do we just start renaming stuff? 13:29
If we try to do the former, then I'd like the websites to not change or redirect to raku.org or something before we are ready in general. 13:30
tyil Geth has just been updated to use #raku and #raku-dev
nine Looking through the release blockers, we actually seem to be in a pretty good shape. They're all in the "please re-test" stage or similar, except for the hand full of blin regressions - which are not all certain regressions 13:40
13:52 anatofuz left 13:53 anatofuz joined 13:55 anatofuz left, anatofuz joined 14:02 Kaiepi left, Kaiepi joined
AlexDaniel nine: well, Unicode::GCB is definitely broken and just needs a PR I think 14:21
I don't remember what's up with Math::FFT::Libfftw3 14:22
but yeah, it's looking pretty good
patrickb If we want to make this release the first that provides pre compiled distributions, we still need to change the website to have respective download buttons. 14:24
That's not a blocker though
nine pre compiled distributions? 14:31
patrickb nine: Binary releases. A compiled MoarVM + Rakudo + zef + some documentation. See github.com/rakudo/rakudo/blob/mast...-binary.md 14:34
In .zip for Windows and .tgz for MacOS + Linux.
AlexDaniel patrickb: isn't it something that rakudo star should do? 14:38
I'm not sure
and even if we do it, it does feel like these should be separate from main source tarballs, right?
patrickb They should be a separate tarballs for sure. 14:39
14:39 anatofuz left
patrickb I don't see a strong reason why we shouldn't do them for every compiler release, except for the additional work. 14:40
14:41 anatofuz joined 14:45 anatofuz left
Guest13443 patrickb: can I ask you a question about perl6-valgrind-m and perl6-gdb-m ? 15:07
patrickb Go ahead
Guest13443 If I run those command from my rakudo dir they don't work since --libpath=/home/dogbert/repos/rakudo/install/share/nqp/lib is missing 15:09
it is not missing from perl6-m however
it used to work not too long ago 15:10
patrickb Sounds wrong. 15:13
Guest13443 I get the feeling that the above libpath gets lost somehow 15:16
patrickb I'm not sure they are even intended to work from the build dir anymore. I thought they would only be generated on install, but I might be mistaken. 15:19
I'll be back in a few minutes... 15:21
15:21 patrickb left 15:29 patrickb joined
Guest13443 perhaps people are not supposed to run this stuff from the build dir but if that's the case then why does perl6-m work? 15:33
nine Why would I only get valgrind or gdb support with an installed rakudo? 15:35
timotimo nine: it's the compiled launchers maybe? 15:40
what do you mean "support"?
15:49 anatofuz joined 16:00 Kaiepi left 16:02 Kaiepi joined
nine I just mean that it'd be odd to have to make install just to run valgrind or gdb 16:02
16:07 anatofuz left 16:22 anatofuz joined 16:26 anatofuz left
patrickb Looking at the makefile seems to indicate they are actually meant to be there. 16:29
16:34 anatofuz joined 17:01 anatofuz left 17:03 anatofuz joined 17:08 anatofuz left 17:23 anatofuz joined
lucasb The question of whether the rename was a backward compatible change was raised in #81. I *do* believe it is one of the most backward incompatible change this project could ever made, since eventually every user will have to touch their code if they want it to keep working on newer Raku versions. And YET, everything else is *frozen*. 17:24
There's lot of little details that such rename will need (file extensions, version number, maybe others) that were made in a rush. And there are other little details that have already accumulated in Perl 6 itself... So, may I re-raise my suggestion: *lift the backward compatibility burden completely* and say this project will be delivering a "v1.0" in about a year or so?
17:24 leont joined
lucasb IOW, turn this moment in "before christmas" again :) 17:27
lizmat there's something to be said for that
17:27 anatofuz left 17:29 anatofuz joined
vrurg lucasb: do you consider the limited resources we actually have? 17:31
vrurg knows what it would take to change the version numbering.
lucasb vrurg: My suggestion is to relax the backward compatibility constraint, now that the rename path was chosen. I wrote "v1.0" in quotes because I don't care what the new versioning scheme would look like. IMO, Raku v6 don't make sense, but I'm fine with it. And "v1.0" of this project was already released in 2015 by the same limited resources this project has. This new again first version would be about polishing it. If 17:47
there's nothing outstanding to change, then fine, Raku is perfect, and let current decisions crystalize themselves.
vrurg lucasb: The renaming PR was written based on the assumption that no other breakage except for unavoidable is to take place. Then, again, in 2015 there was nothing to break about the versioning. And, actually, it wasn't v1.0 anyway. The way I currently see it is that we gonna have another debate on versioning and if it ever going to change then I'd be rather in favor of moving on to v7.0. All the historical reasons, amount of changes, backward compat – 17:54
everything would be in favor of this.
18:13 sortiz joined 18:14 anatofuz left 18:20 anatofuz joined 18:25 anatofuz left 18:36 anatofuz joined 18:40 anatofuz left 18:44 anatofuz joined
releasable6 Next release in ≈4 days and ≈23 hours. R6 is down. At least 5 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
19:10 ufobat left 19:29 anatofuz left 19:38 anatofuz joined, Kaiepi left 19:39 Kaiepi joined
patrickb Guest13443: Do you know whether gdb / valgrind / lldb are disturbed by the process to be debugged exec-ing? 19:42
Guest13443: If yes, we could either change the gdb / valgrind / lldb runners to call the perl6-m native executable in the build directory or just remove them entirely.
19:45 anatofuz left 19:55 anatofuz joined 20:40 anatofuz left
MasterDuke would this be relevant for rakudo (or moarvm)? tokio.rs/blog/2019-10-scheduler/ 20:42
20:45 anatofuz joined 20:49 anatofuz left 21:00 anatofuz joined 21:02 Kaiepi left 21:03 Kaiepi joined 21:04 Kaiepi left 21:05 Kaiepi joined, anatofuz left 21:09 anatofuz joined
jnthn MasterDuke: Yes 21:25
MasterDuke: Rakudo mostly, I think, since our scheduler is implemented there
21:49 anatofuz left 21:56 anatofuz joined 21:58 anatofuz left, anatofuz joined 22:16 titsuki joined 22:45 patrickb left 22:47 sortiz left 22:58 anatofuz left 23:06 anatofuz joined