Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2017-08-21

| Channels | #perl6-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:04 BenGoldberg joined #perl6-dev
01:33 geekosaur joined #perl6-dev
01:52 ilbot3 joined #perl6-dev
01:52 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
03:11 geekosaur joined #perl6-dev
03:19 llfourn joined #perl6-dev
03:27 pharv joined #perl6-dev
03:35 geekosaur joined #perl6-dev
06:56 brrt joined #perl6-dev
07:25 nine Regarding Windows: how am I supposed to fix issues when I cannot even test them? Why am I supposed to care if not even the people who chose that platform care enough to involve themselves? If there are so many as stmuk++ points out, why is there not even one of them available for questions or testing?
07:26 nine I personally do not care a bit about Windows. Windows is a thing that makes my life worse, not better, whenever I have contact with it. I refuse to waste my precious spare time on it.
07:27 nine Of course I would never deliberatly break it. I usually keep it in mind whenever I touch something that may be platform dependent. I also try to duckduckgo those things before committing (to) something. But at some point all I can do is try and see if it breaks.
07:28 nine And the only way for me to see it break is some CI system that fails to build half the time anyway. Or people notifying me on the channel. And the latter happened only a week later as it was.
07:35 |Tux| joined #perl6-dev
07:47 jsimonet joined #perl6-dev
07:50 jsimonet joined #perl6-dev
07:56 robertle joined #perl6-dev
08:09 robertle joined #perl6-dev
08:18 Ven joined #perl6-dev
08:27 stmuk I'm no particular fan of Windows but we do need to continue to support it.  I think the root problems are with the release and development processes (not just minority platforms such as windows).  The monthly releases are getting harder and harder (eg. 2017.04 as well as 2017.08)
08:27 lizmat does docs.perl6.org get refreshed automatically, or is some manual intervention needed?
08:28 * lizmat would like to see her updates to atomicint.pod live before the P6W publishes
08:28 stmuk lizmat: it should be automatic but was broken (and is now hopefully fixed as of a day or two ago)
08:29 lizmat well, the last update was 2017-08-20T01:52:51Z afaics
08:29 lizmat and that's more than 24 hours ago  :-(
08:34 lizmat breakfast&
10:10 Zoffix TIL there's Introducing Perl 6 book
10:10 Zoffix https://www.flickr.com/photos/wendyga/sets/72157685410027133
10:11 Zoffix Don't see it on https://perl6book.com/
10:11 lizmat Zoffix: it's not a book, it's a brochure/leaflet
10:13 Zoffix Ah
10:13 Zoffix 60 pages. That's pretty big
10:14 Zoffix lizmat: currently forcing a doc build.
10:14 Zoffix The last build log had your commit in it; not sure why that didn't propagate to the site
10:16 lizmat Zoffix++
10:20 dogbert17 joined #perl6-dev
10:20 AlexDaniel fwiw https://irclog.perlgeek.de/moarvm/2017-08-21#i_15048783
10:28 Zoffix it's curious they're not failing on master.
10:34 lizmat Zoffix: doc has arrived  :-)
10:36 Zoffix Good. Though I'm none the wiser why it didn't update of its own volition :/
10:38 Zoffix .tell [Coke] looks like cronned doc still is busted. Didn't build existing new commits for 12 hours (even tho ~/update.log had them in), but when I manually $HOME/force-update'ed it, then it just worked.
10:38 yoleaux Zoffix: I'll pass your message to [Coke].
10:43 lizmat timotimo: do we have an archive of Perl 6 weeklies somewhere from before 7 Feb 2014 ?
10:45 timotimo lizmat: did wordpress throw out any entries older than that?
10:46 lizmat https://p6weekly.wordpress.com/2014/02/07/off-we-go/  # oldest I could find
10:48 timotimo something's wrong with my browser and/or 'net conneection
10:48 timotimo oh, yeah
10:48 timotimo the ones before that were on glitchmr's blog somewhere
10:49 AlexDaniel timotimo: I'm assuming you won't be able to take a look at the spesh issue also, right?
10:49 timotimo is that only on windows?
10:50 AlexDaniel timotimo: no, linux+errata
10:50 timotimo i can look at it
10:50 Zoffix AlexDaniel: this reminds me of the bus factor discussion :)
10:51 AlexDaniel c: HEAD https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
10:51 committable6 AlexDaniel, Successfully fetched the code from the provided URL.
10:51 committable6 AlexDaniel, https://gist.github.com/a2147ac2748bedca7da8a3900b23125d
10:51 AlexDaniel timotimo: ?
10:52 timotimo you can set env vars for committable, right?
10:52 AlexDaniel commit: MVM_SPESH_DISABLE=1 HEAD https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
10:52 committable6 AlexDaniel, Successfully fetched the code from the provided URL.
10:52 committable6 AlexDaniel, https://gist.github.com/bd2f3c76c444e15a9feb6c3e7815e862
10:53 timotimo c: HEAD MVM_SPESH_BLOCKING=1 MVM_SPESH_NODELAY=1 https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
10:53 committable6 timotimo, https://gist.github.com/c015388120639ac0b8cc9a5fe1fea7a0
10:53 lizmat timotimo: did you mean this: http://blogs.perl.org/users/perl_6_reports/archives.html
10:53 timotimo that's not how
10:53 timotimo lizmat: no, they had their own blog hosted on maybe their own server
10:53 AlexDaniel timotimo: ah shit… I don't think it supports more than one
10:54 timotimo i remember it was mostly white
10:54 AlexDaniel according to the code it does
10:55 AlexDaniel MVM_SPESH_BLOCKING is simply not in the list
10:55 timotimo ah
10:56 timotimo GlitchMr's twitter looks kind of dead, last tweet 2014
10:57 committable6 joined #perl6-dev
10:57 AlexDaniel c: HEAD MVM_SPESH_BLOCKING=1 MVM_SPESH_NODELAY=1 https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
10:57 committable6 AlexDaniel, https://gist.github.com/afa92bb0a72a22f400042e8ea710bac0
10:57 AlexDaniel ah right
10:57 AlexDaniel c: MVM_SPESH_BLOCKING=1 MVM_SPESH_NODELAY=1 HEAD https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
10:57 committable6 AlexDaniel, Successfully fetched the code from the provided URL.
10:57 committable6 AlexDaniel, https://gist.github.com/4d4d899043ce01a1f3d71508f4270a00
10:58 AlexDaniel timotimo: ? :)
11:02 Geth ¦ rakudo/nom: 0de25a8f4c | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/announce/2017.08.md
11:02 Geth ¦ rakudo/nom: Release announcement draft for 2017.08
11:02 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0de25a8f4c
11:02 AlexDaniel mmmm… I c
11:07 timotimo if the spesh log weren't typically gigantic, i'd ask for it to be includable d)
11:08 timotimo :)
11:09 AlexDaniel well, you can run this file with your own perl6
11:10 timotimo of course
11:12 lizmat AlexDaniel: I realise it's copy/paste from previous announcements
11:12 lizmat but looking at it now, I think "It includes bugfixes and optimizations on top of
11:12 lizmat the 2015.12 release of Rakudo, but no new features."
11:12 AlexDaniel hahahaha no new features
11:12 lizmat is a bit untrue on the "no new features" part  :-)
11:12 AlexDaniel I'll fix that. I have to fix markdown formatting anyway
11:14 nine lizmat: I don't think augmenting Grammar in the 6.d setting is gonna work. That setting will be loaded when any part of the program at all uses v6.d.PREVIEW. The augmentation will then affect all code.
11:14 timotimo "but doesn't change the language"?
11:15 lizmat nine: isn't that also the case with the new await behaviour then?
11:16 nine Also I remember jnthn arguing that augmenting a class in the 6.d setting is a bad idea. I tried that in my original language_versions branch.
11:16 nine lizmat: await is a sub. It's not a method that's added to an existing class via augmentation
11:17 lizmat but it masks the original sub await?
11:17 nine Yes. And that's fine. Same with the masked DateTime class. But augment changes the _existing_ class and thus the change is visible globally.
11:18 lizmat hmmm...
11:18 nine I really wish we would just revert the whole business and actually take the time to do this right.
11:19 AlexDaniel but we still can
11:19 lizmat revert Gramnmar.parse returning a Failure at all you mean
11:19 lizmat at least for 2017.08 ?
11:19 nine yes
11:19 lizmat AlexDaniel: what are your feelings about that ?
11:19 nine We have lived with parse returning Nil for years. One more month won't make all that much difference.
11:20 lizmat nine: I would agree with that
11:20 AlexDaniel lizmat: if it doesn't work, then what's the point of keeping it?
11:20 nine Us breaking existing code for no good reason however will cost us in the long term.
11:20 AlexDaniel again, there's no rush :)
11:20 lizmat I guess it's now up to AlexDaniel to ok a revert
11:20 AlexDaniel I think reverting it is the right thing to do
11:20 AlexDaniel better safe than sorry
11:21 lizmat ok, I'll revert
11:21 nine I would love for us to take the opportunity at the Swiss Perl Workshop to figure out how to do this language version thing for real.
11:22 AlexDaniel my position was that we definitely should *not* include it in 2017.08. When it was moved to v6.d, I thought that there is no problem anymore
11:22 lizmat feels like a plan to me
11:22 AlexDaniel if there are any suspicions, then surely it has to be reverted, especially given that we have time for it :)
11:23 nine Maybe what we need is a mechanism that is like augment, but will actually create a copy of the original class with the same name + the augmentation.
11:23 jnthn And then they all won't type match 'cus they're not the same class. :P
11:24 nine But that won't actually help all that much in Grammar's case. Because what's important is not the lexical scope where the grammar is defined, but the scope of the _caller_ of the .parse method.
11:24 lizmat basically use the existing class as a Role ?
11:24 jnthn This is a case of "solution worse than problem"
11:24 nine jnthn: I'm just thinking out loud here :) The type matching could be solved by not copying the contents of the original class but making the new class a subclass.
11:25 nine jnthn: right now, I'm just curious if this could be made to work (and I think it actually could - and would expose some weird bugs), not if it's a good thing to do :)
11:26 nine But again, it won't solve anything here anyway.
11:26 nine What would solve parse's problem is if there was a way to get the caller's language version.
11:26 nine Or a very low tech solution: just provide a different method with the new semantics.
11:27 timotimo class Grammar:ver<v6.d> is Grammar:ver<v6.c> { method parse { ... } }
11:28 * lizmat spectests removal
11:29 * lizmat also removes a paragraph from the Perl 6 Weekly
11:30 nine I also wonder how we could prevent more people from falling into the augment trap
11:31 Ven`` nine: what you propose reminds me of ruby's refinements
11:31 Ven`` they're lexically-scoped augments/monkey-patching
11:33 AlexDaniel lizmat: I'll remove it from the changelog
11:34 Geth ¦ rakudo/nom: 465d91abdf | (Elizabeth Mattijsen)++ | 4 files
11:34 Geth ¦ rakudo/nom: Remove Grammar.parse returning a Failure on fail
11:34 Geth ¦ rakudo/nom:
11:34 Geth ¦ rakudo/nom: We cannot really have a versioned solution to the problem yet.  So
11:34 Geth ¦ rakudo/nom: either this will have to wait for a real 6.d, or we need to find a
11:34 Geth ¦ rakudo/nom: better solution to versioning.
11:34 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/465d91abdf
11:35 lizmat nine: I think it could be solvable by having $*PERL be actually scoped
11:35 Geth ¦ rakudo/nom: fd82e3be1d | (Aleks-Daniel Jakimenko-Aleksejev)++ | 2 files
11:35 Geth ¦ rakudo/nom: Add Markdown code blocks where needed
11:35 Geth ¦ rakudo/nom:
11:35 Geth ¦ rakudo/nom: There are other places where this can be done, but I don't think we
11:35 Geth ¦ rakudo/nom: want every other word to be highlighted.
11:35 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/fd82e3be1d
11:35 Geth ¦ rakudo/nom: 44542c5666 | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/announce/2017.08.md
11:35 Geth ¦ rakudo/nom: Remove untrue statement
11:35 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/44542c5666
11:35 Geth ¦ rakudo/nom: d5aec61f54 | (Aleks-Daniel Jakimenko-Aleksejev)++ | 2 files
11:35 Geth ¦ rakudo/nom: Remove reverted change from the changelog
11:35 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d5aec61f54
11:36 AlexDaniel lizmat++
11:37 nine lizmat: yes, that'd be really nice
11:37 lizmat m: say $*PERL; { use v6.d.PREVIEW; say $*PERL }; say $*PERL   # expected 6.c / 6.d / 6.c
11:37 camelia rakudo-moar 0de25a: OUTPUT: «Perl 6 (6.d)?Perl 6 (6.d)?Perl 6 (6.d)?»
11:37 AlexDaniel timotimo: so. What about the spesh issue? If you can look at it only a few hours later, just tell me, and I'll take a break too :)
11:38 pmurias joined #perl6-dev
11:39 lizmat m: say &await; { use v6.d.PREVIEW; say &await }; say &await   # same issue
11:39 camelia rakudo-moar 0de25a: OUTPUT: «sub await (| is raw) { #`(Sub|52187256) ... }?sub await (| is raw) { #`(Sub|52187256) ... }?sub await (| is raw) { #`(Sub|52187256) ... }?»
11:39 lizmat I mean, why can't we treat "use v6.d" just like any other lexically scoped module loading ?
11:40 nine lizmat: not exactly the same. The 6.d setting is loaded for the whole computation unit, which is what your await test shows. But $*PERL.version is changed for all code that runs while compiling the current computation unit. Even loaded modules
11:40 lizmat nine: well, that's my point: should it be for the whole compilation unit?
11:40 lizmat any other use of "use" *is* lexical
11:41 lizmat so maybe we should have at least have another word for it?
11:41 timotimo i'll have a look
11:42 timotimo just building a fresh rakudo now
11:42 lizmat also: it gets worse with precomp, no
11:42 lizmat ?
11:42 lizmat if a  module is precomped with 6.c, and 6.d is done, the precomp would still use 6.c semantics
11:42 timotimo "6.d is done"?
11:43 nine The $*PERL.version thing? Yes. I bet no one could really answer without trying, what $*PERL.version will print in all possible cases. It's acting really weird.
11:43 lizmat whereas if the module would need to be recompiled for whatever reason, it suddenly would have 6.d semantics
11:43 lizmat timotimo: is released
11:43 timotimo well, when your rakudo changes all your modules will be precomped again
11:44 lizmat but thinking about it further now, I realise by that time we would still have the same issue
11:44 timotimo so you won't have a few 6.c and a few 6.d
11:47 timotimo waitwhat
11:47 timotimo it dies under DEPRECATED
11:48 lizmat que?
11:50 timotimo https://gist.github.com/timo/d3ef42d54e5c60ea46918935e01e484b
11:51 * lizmat checks
11:51 timotimo it grabs a backtrace and goes through its "interesting index"es
11:51 timotimo i believe when it goes to grab $bt[$index] it does the Any thing
11:52 lizmat aha, next-interesting-index can retiurn Nil
11:52 timotimo this can explain the reliance on spesh
11:52 timotimo because spesh is able to inline stuff
11:53 timotimo it should probably reconstruct the original backtrace when one is created though?
11:54 timotimo though throwing exceptions probably shouldn't always deoptimize inlines?
11:55 lizmat timotimo: I'm going to change the code so that it will keep the last valid index seen
11:55 timotimo c: MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 HEAD use Test; is_approx 1, 2; is_approx 2, 3; is_approx 3, 4; is_approx 4, 5;
11:55 committable6 timotimo, https://gist.github.com/14521c1b27f8d122b2392308735bc57a
11:55 timotimo looks like that's not quite enough
11:56 timotimo c: MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 HEAD use Test; for ^100 { is_approx $_, $_ + 1 }
11:56 committable6 timotimo, https://gist.github.com/d83a2c092ff7f2db3ae336650c5c5085
12:01 timotimo took me a bit to be able to open the gist, but here's my golf
12:02 * lizmat waits
12:02 dogbert17 there seems to be something wrong with the rakudo build
12:02 timotimo oh? what is?
12:02 AlexDaniel dogbert17: okay, more info?
12:03 dogbert17 well, I wanted the new changes that lizmat made, so I did 'git pull' in my rakudo dir followed by 'perl Configure.pl --gen-moar --gen-nqp --backends=moar'
12:04 timotimo lizmat: i'd say put your fix for DEPRECATED in
12:04 lizmat timotimo: spectesting now
12:04 dogbert17 I'm then told to run 'make' which I did and it worked well
12:04 timotimo good
12:04 dogbert17 but 'make test' fails unless I do 'make install' first and that's not right
12:05 lizmat hmmm...  isn't that a known issue ?
12:05 dogbert17 if it is then it's relatively new I'd say
12:06 timotimo yeah
12:06 timotimo that's because we changed the PERL6LIB stuff
12:07 timotimo isn't it?
12:07 timotimo perl6 used to always look under the current path for its things
12:07 timotimo which is a security risk
12:08 dogbert17 so the generated makefile is incorrect then ...
12:09 dogbert17 do we want it to work?
12:09 timotimo isn't our problem that we can't pass env vars cross-platform from makefiles?
12:11 timotimo i think this was discussed recently
12:11 timotimo making test depend on install seems like a bad idea
12:13 dogbert17 if someone follows the given instructions and it doesn't work that's a bit suboptimal
12:13 timotimo of course
12:14 dogbert17 I guess the spesh problem is a bigger issue :-)
12:15 lizmat dogbert17: I've just done a fresh clone of rakudo, and a fresh build without install, but "make test" runs fine for me?
12:16 dogbert17 hmm, interesting
12:16 lizmat macOS fwiw
12:17 dogbert17 might try to nuke my install a little later
12:17 dogbert17 timotimo: spesh problem alternative 'MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 ./perl6 t/spec/S03-operators/set_precedes.t'
12:17 nine make test is supposed to work even without a make install
12:18 nine That's why the generated ./perl6-m will still look in '.' for modules and nqp libs. Only the installed one (we generate a new one in the install target) gets rid of those
12:18 Geth ¦ rakudo/nom: 1d69ebb9c2 | (Elizabeth Mattijsen)++ | src/core/Deprecations.pm
12:18 Geth ¦ rakudo/nom: Be a more careful about interesting indexes
12:18 Geth ¦ rakudo/nom:
12:18 Geth ¦ rakudo/nom: Backtrace.next-interesting-index *can* return Nil, so we should guard
12:18 Geth ¦ rakudo/nom: against that by remembering the last valid index.
12:18 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1d69ebb9c2
12:18 nine I consider the make test thing a definite release blocker as it will break packaging scripts.
12:18 lizmat timotimo: ^^^
12:20 * dogbert17 tries 'make test' again with lizmat's latest commit ...
12:20 lizmat lunch&
12:20 AlexDaniel c: 1d69ebb9c2 HEAD https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
12:20 committable6 AlexDaniel, ¦1d69ebb: «Cannot find this revision (did you mean “1247986”?)»
12:21 dogbert17 /usr/bin/perl t/harness5 t/01-sanity t/04-nativecall
12:21 dogbert17 t/01-sanity/01-literals.t ................... Dubious, test returned 1 (wstat 256, 0x100)
12:21 dogbert17 No subtests run
12:21 dogbert17 etc
12:21 Zoffix c: 1d69ebb9c2,HEAD https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
12:21 committable6 Zoffix, Successfully fetched the code from the provided URL.
12:21 committable6 Zoffix, https://gist.github.com/714501a090cea8b2d402a01f5b8d9750
12:22 AlexDaniel it's just not ready yet
12:23 timotimo dogbert17: i get the same error as for the other file
12:23 dogbert17 yup, but now lizmat seems to have fixed it
12:24 AlexDaniel c: 1d69ebb9c2 HEAD https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
12:24 committable6 AlexDaniel, ¦1d69ebb: «===SORRY!=== Error while compiling /tmp/kCfwe4gXRp?Confused?at /tmp/kCfwe4gXRp:1?------> HEAD https:?//raw.githubusercontent.com/perl6/roast/?    expecting any of:?        colon pair «exit code = 1»»
12:24 AlexDaniel c: 1d69ebb9c2 https://raw.githubusercontent.com/perl6/roast/6.c-errata/S32-num/complex.t
12:24 committable6 AlexDaniel, Successfully fetched the code from the provided URL.
12:24 committable6 AlexDaniel, https://gist.github.com/fc21c9c47eab5dfd0cfece883dce1369
12:24 AlexDaniel “Block <unit> (from unknown) seen at:” :/
12:25 AlexDaniel but that's good enough
12:25 pmurias shouldn't we get rid of NQP_LIB on moarvm altogether
12:25 pmurias ?
12:28 pmurias timotimo: the security fix wasn't rakudo loading Perl 6 modules PERL6LIB but even installed rakudo prefering to load it's subparts from the current working directory
12:29 nine FWIW make test runs just fine here without a make install
12:30 nine m-test6, too
12:31 dogbert17 hmm, If I'm the only one having this problem then perhaps by rakudo is busted, let me start over
12:32 dogbert17 s/by/my/
12:37 timotimo lizmat: is there a good reason we still test these deprecated operators?
12:38 timotimo well, i suppose they're still part of the language?
12:38 AlexDaniel yup
12:39 AlexDaniel dogbert17: ping me when it's done
12:42 dogbert17 it's done, my bad, sorry for the noise, my build env was obviously busted :(
12:42 Zoffix The deprecated message is from the is_approx itself, which is now is-approx
12:43 Zoffix We converted master to is-approx, but left 6.c-errata as is, 'cause the argument was that 6.c-errata should be more-or-less runnable by earlier rakudos
12:43 timotimo right
12:49 Zoffix AlexDaniel: are you gonna be using the release bot?
12:49 Zoffix I've no idea what its opinion on already-existing release announcements.
12:50 Zoffix AlexDaniel: also dogbert17 pinged you above when done :)
12:50 dogbert17 only forgot the ping part :)
12:50 NeuralAnomaly joined #perl6-dev
12:50 Zoffix NeuralAnomaly: steps
12:50 NeuralAnomaly Zoffix, Use releasable6 bot instead. e.g. try: releasable6: status
12:50 Zoffix screw you
12:50 Zoffix releasable6: steps
12:50 releasable6 Zoffix, I cannot recognize this command. See wiki for some examples: https://github.com/perl6/whateverable/wiki/Releasable
12:52 Zoffix well, it's a separate step that can be skipped https://github.com/zoffixznet/na/blob/master/lib/NA/ReleaseScript/Rakudo.pm6#L43-L69
13:03 AlexDaniel Zoffix: no, not going to use a bot
13:04 AlexDaniel I have a sakefile here that cuts releases locally
13:04 AlexDaniel and then I can decide to push if it looks right
13:04 Zoffix cool
13:05 timotimo .o( sake is japanese make? )
13:06 Zoffix "Curious if anyone knows of any good training for Perl?"   https://www.reddit.com/r/perl/comments/6uywyh/online_courses/
13:06 Zoffix We have a bunch on https://perl6.org/resources/
13:07 stmuk maybe they mean actual physical training?
13:09 timotimo what, you mean like lifting? or track&field?
13:10 Zoffix .oO( Pushups for Perl )
13:11 Zoffix AlexDaniel: so I wanna leave for a week. You're good with the release, right? No questions? I can just go?
13:12 AlexDaniel Zoffix: sure
13:12 Zoffix Sweet
13:12 Zoffix See y'all in a week
13:12 AlexDaniel Zoffix++ take care
13:12 lizmat Zoffix++  have a good R&R !
13:13 timotimo seeya Zoffix :)
13:13 Geth ¦ rakudo/nom: b686f41312 | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
13:13 Geth ¦ rakudo/nom: Reflect actual date, claim next release
13:13 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b686f41312
13:14 jnthn Zoffix: Good relax :)
13:15 [Coke] Thanks everyone (esp. AlexDaniel) for making sure the release is awesome
13:15 yoleaux 10:38Z <Zoffix> [Coke]: looks like cronned doc still is busted. Didn't build existing new commits for 12 hours (even tho ~/update.log had them in), but when I manually $HOME/force-update'ed it, then it just worked.
13:15 AlexDaniel So I'm going to kick off the script in a few minutes, if anybody has any critical issues please feel free to ping me
13:16 lizmat all clear from me
13:17 [Coke] lizmat: is your doc issue temporarily resolved?
13:17 lizmat [Coke]: yes
13:17 AlexDaniel ok I see something in #moarvm
13:18 AlexDaniel ah “we *won't* need another release”
13:18 [Coke] Looks like the build is hanging on the sync.
13:20 AlexDaniel by the way, maybe to do it 200% right I should have bumped NQP/Moar *before* the release
13:20 AlexDaniel so that people who really want to test it on Windows have easier time doing this
13:21 [Coke] yes. waiting until the last second to bump versions is part of the problem.
13:21 [Coke] This isn't you, it's a long running issue.
13:24 AlexDaniel Maybe we should be publishing beta releases or something? Right now I have the whole process automated locally, it doesn't push tags, bump commits and tarballs, but other than that it is identical to the normal release process
13:24 AlexDaniel so I get tested tarballs on the output
13:25 dogbert17 got a flop in t/spec/S29-os/system.t
13:25 dogbert17 can someone else run that?
13:26 [Coke] (doc build) - there was a hung copy of htmlify waiting for a lock. killed that, we'll see if that fixes the issue. :|
13:26 AlexDaniel “Some issues can be quickly found by users, even if uncaught by tests” is one of the things Zoffix++ mentioned in his blog post about 2017.04 releases
13:26 timotimo aye
13:27 dogbert17 it's test 35: not ok 35 - :cwd(...) changes the directory processes are spawned in
13:28 AlexDaniel c: https://raw.githubusercontent.com/perl6/roast/master/S29-os/system.t
13:28 committable6 AlexDaniel, I cannot recognize this command. See wiki for some examples: https://github.com/perl6/whateverable/wiki/Committable
13:29 AlexDaniel c: HEAD https://raw.githubusercontent.com/perl6/roast/master/S29-os/system.t
13:29 committable6 AlexDaniel, It looks like a URL, but for some reason I cannot download it (HTTP status line is 503 Service Unavailable).
13:29 AlexDaniel ?
13:29 AlexDaniel c: HEAD https://raw.githubusercontent.com/perl6/roast/master/S29-os/system.t
13:29 committable6 AlexDaniel, It looks like a URL, but for some reason I cannot download it (HTTP status line is 503 Service Unavailable).
13:30 AlexDaniel “We are investigating reports of elevated error rates”
13:31 AlexDaniel hey, I need a little bit of github for the release process :)
13:31 AlexDaniel yup, MoarVM clone just failed due to timeout
13:33 dogbert17 AlexDaniel: got the test working, had a PATH issue
13:33 AlexDaniel dogbert17: great
13:34 [Coke] here's a thing: doc build kicked off. did a git fetch; the fetch failed (github issues); the lock file was NOT removed.
13:35 [Coke] (this effectively disables any future updates)
13:38 timotimo aaah, dang
13:38 timotimo that's a bad thing :)
13:38 timotimo that's why lock files have the pid of the process in them so they can be removed when the pid in question is found to be dead?
13:39 [Coke] this lock file is empty.
13:39 [Coke] ... (urk. ok, with the job running every five minutes, I am now longer sure which instance of cron this one is from.: )
13:39 [Coke] s/cron/the job/
13:39 timotimo heh.
13:40 timotimo there's a few things i find strange about the introducing perl6 booklet
13:40 timotimo isn't "people claiming moore's law is ending" a meme %)
13:40 lizmat timotimo: please direct your queries about that booklet to mdk  :-)
13:41 timotimo oops, this page calls perl6 a "moden" language %)
13:43 timotimo oh crap, the PDF has an example "sin( /2)" which "will return the sine of  /2" "which, as you know, is 1."
13:44 moritz uhm, which PDF?
13:47 timotimo "introducing perl6", https://drive.google.com/file/d/0B7Iymye3xMqjOEdnb0dwS2QtaTQ/view?usp=sharing
13:47 AlexDaniel fwiw I'm going to wait out github outage instead of trying to make it through connection errors and timeouts
13:48 [Coke] ok. the last logged run had nothing to do - one update I made to the script is that now when it has nothing to do, it explicitly calls "exit"; could this have changed the behavior of flock, which calls the script?
13:48 [Coke] (and that run didn't remove the lock file)
13:49 timotimo how do i best reach mdk?
13:49 [Coke] oh. I misunderstood how flock works. the presence of the lock file means nothing. :|
13:50 [Coke] I... think the doc build is probably fine at the moment, then. :)
13:50 timotimo oh, it actually acquires a lock on the lock file
13:52 [Coke] yup, seems fine ish. Wonder what hung earlier that was giving lizmat/zoffix issues. :|
13:56 AlexDaniel heh, this release is tough. First someone removing their modules from the ecosystem and getting in the way, now github outage :)
14:31 brrt joined #perl6-dev
14:44 dogbert17 timotimo: did you forget a 'break,' here? https://github.com/MoarVM/MoarVM/blob/master/src/spesh/optimize.c#L2139
14:46 timotimo looks like i did. fortunately it's no problem there
14:46 timotimo should be put in, though
14:47 dogbert17 do you want to do it or should I do a PR?
14:48 timotimo i can do it real quick
14:49 dogbert17 cool, I'll continue looking at your latest Coverity Scan instead then :)
14:55 AlexDaniel alright, “Everything operating normally.”
15:03 raschipi joined #perl6-dev
15:29 Skarsnik joined #perl6-dev
15:33 AlexDaniel .oO( https://www.youtube.com/watch?v=Cimh1LY1NKs )
15:35 Skarsnik AlexDaniel, do you think I should use Whateverable as a base to do a bot that NativeCallise C code?
15:37 AlexDaniel Skarsnik: it depends. If it's in whateverable project, I'll be maintaining, fixing, refactoring, and writing tests for it for an extended period of time. If that's what you want, then go for it.
15:38 AlexDaniel that's basically the only difference
15:38 Skarsnik I was more thinking from a pratical pov, since it's already have stuff to read from gist and produce gist x)
15:39 AlexDaniel Skarsnik: then yeah
15:40 Skarsnik I just need to rewrite the run-stuff method since I need to ignore stderr x)
15:40 AlexDaniel maybe you want to use “run” directly?
15:41 Skarsnik probably. I just need stdout (from GPTrixie)
15:41 Skarsnik or I could add a silent mode to gptrixie
15:42 AlexDaniel try run, it should work
15:43 AlexDaniel (unless you want it to time out if it's stuck and stuff like that)
15:44 AlexDaniel Skarsnik: by the way, are there any security implications when throwing untrusted inputs into gptrixie?
15:44 Skarsnik gccxml (gcc4.9) will probably complain
15:45 AlexDaniel (because if so, maybe it should go through run-snippet as it will do it in a more or less secure way once I finally implement it)
15:46 Skarsnik basicly the bot will run gptrixie --all generatedheader.h and generatedheader.h is the user feed. gptrixie will do a run 'gccxml', options.. 'generatedheader.h'
15:46 Skarsnik so I don't think you can inject stuff
15:52 AlexDaniel cool
15:58 * AlexDaniel feels like he is not doing TEST_JOBS=8 or -j 8 somewhere, because this took a bit longer than I wanted
15:59 Skarsnik damn nice error message
15:59 Skarsnik Function 'Type' needs parens to avoid gobbling block (or perhaps it's a class that's not declared or available in this scope?)
15:59 Skarsnik it's exactly that ^^
16:00 Skarsnik This is really obonxious/annoying when working with huge header : Times -- gccxml: 0.5213915 sec; xml parsing: 18.6218791 sec; magic: 1.63685878
16:16 AlexDaniel joined #perl6-dev
16:17 * AlexDaniel had to reboot :D
16:18 Skarsnik x)
16:19 dogbert2 joined #perl6-dev
16:23 Geth ¦ nqp/refs/tags/2017.08: 01001f19e1 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
16:23 Geth ¦ nqp/refs/tags/2017.08: [release] Bump MoarVM revision to 2017.08.1
16:23 Geth ¦ nqp/refs/tags/2017.08: review: https://github.com/perl6/nqp/commit/01001f19e1
16:23 Geth ¦ nqp/refs/tags/2017.08: d143da5d20 | (Aleks-Daniel Jakimenko-Aleksejev)++ | VERSION
16:23 Geth ¦ nqp/refs/tags/2017.08: [release] Bump VERSION to 2017.08
16:23 Geth ¦ nqp/refs/tags/2017.08: review: https://github.com/perl6/nqp/commit/d143da5d20
16:23 Geth ¦ rakudo/refs/tags/2017.08: 0127c4238f | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
16:23 Geth ¦ rakudo/refs/tags/2017.08: [release] Bump NQP revision to 2017.08
16:23 Geth ¦ rakudo/refs/tags/2017.08: review: https://github.com/rakudo/rakudo/commit/0127c4238f
16:23 Geth ¦ rakudo/refs/tags/2017.08: 69a91d62a4 | (Aleks-Daniel Jakimenko-Aleksejev)++ | VERSION
16:23 Geth ¦ rakudo/refs/tags/2017.08: [release] Bump VERSION to 2017.08
16:23 Geth ¦ rakudo/refs/tags/2017.08: review: https://github.com/rakudo/rakudo/commit/69a91d62a4
16:23 AlexDaniel this didn't push the commit itself :/
16:24 Geth ¦ rakudo/nom: 0127c4238f | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
16:24 Geth ¦ rakudo/nom: [release] Bump NQP revision to 2017.08
16:24 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0127c4238f
16:24 Geth ¦ rakudo/nom: version bump brought these changes: https://github.com/perl6/nqp/compare/2017.07-58-gd4d77b66c...2017.08
16:24 Geth ¦ rakudo/nom: 69a91d62a4 | (Aleks-Daniel Jakimenko-Aleksejev)++ | VERSION
16:24 Geth ¦ rakudo/nom: [release] Bump VERSION to 2017.08
16:24 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/69a91d62a4
16:24 Geth ¦ rakudo/nom: version bump brought these changes: https://github.com/perl6/nqp/compare/2017.07-58-gd4d77b66c...2017.08
16:24 Geth ¦ nqp: 01001f19e1 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
16:24 Geth ¦ nqp: [release] Bump MoarVM revision to 2017.08.1
16:24 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/01001f19e1
16:24 Geth ¦ nqp: version bump brought these changes: https://github.com/MoarVM/MoarVM/compare/2017.07-467-gf22f8d7c...2017.08.1
16:24 Geth ¦ nqp: d143da5d20 | (Aleks-Daniel Jakimenko-Aleksejev)++ | VERSION
16:24 Geth ¦ nqp: [release] Bump VERSION to 2017.08
16:24 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/d143da5d20
16:24 Geth ¦ nqp: version bump brought these changes: https://github.com/MoarVM/MoarVM/compare/2017.07-467-gf22f8d7c...2017.08.1
16:28 AlexDaniel that's it
16:28 AlexDaniel (almost)
16:29 Skarsnik AlexDaniel, I can't find your rt with the missing possibility to pass the source ip for a socket
16:30 AlexDaniel Skarsnik: RT #131766
16:30 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=131766
16:31 Skarsnik damn need to patch 2 things them xà
16:31 Skarsnik async and inet x)
17:14 * AlexDaniel celebrates with an inappropriate amount of fun
17:15 AlexDaniel honestly, this was hard. I'll what I can do to make it easier for everyone
17:17 AlexDaniel special thanks to samcv++, jnthn++, lizmat++, Zoffix++, nine++, pmurias++, timotimo++ and everybody else who did the awesome work for the release
17:18 AlexDaniel I'll see*
17:20 Skarsnik Hoo the release is done? Contratulation!
17:27 timotimo <3
17:31 jnthn AlexDaniel++ # release
17:31 AlexDaniel releasable6: status
17:31 releasable6 AlexDaniel, Next release in 26 days and ?1 hour. No blockers. Changelog for this release was not started yet
17:37 timotimo AlexDaniel: will there be a tweet i can retweet?
17:38 AlexDaniel timotimo: I don't have twitter :/ should I have one?
17:38 timotimo not important
17:38 timotimo moritz has the perl6 news account
17:43 Skarsnik Do I need to make && make install when I change something in rakudo?
17:43 timotimo yup
17:44 AlexDaniel Skarsnik: are you working on the ticket mentioned above?
17:44 AlexDaniel MasterDuke: fwiw ?
17:48 [Coke] Skarsnik: you might even need to reconfigure, depending on what you changed.
17:49 timotimo yeah, but mostly just if you add new files to the core setting
17:57 nine Or work on the build system :)
18:02 Skarsnik AlexDaniel, yes x)
18:09 Skarsnik I am just working on s::inet for now
18:13 Skarsnik hm I added an arg to connect (in MoarVM and Socket.pm) but I get a Stage mast       : Arg count 4 doesn't equal required operand count 3 for op 'connect_sk' and not sure where this is defined
18:16 moritz Skarsnik: src/core/oplist
18:17 Skarsnik Oh it's in Moar so I have to remake in Moar?
18:19 timotimo yeah, if you change the oplist you need to run the generator tool
18:19 timotimo the instructions are at the top of oplist
18:20 moritz timotimo: https://twitter.com/perl6org/status/899697722781040640
18:26 raschipi Any position on implementing 'use v6.d.PREVIEW' as a command line flag to make it easier for people to run modules tests with it to detect upcoming problems?
18:28 Skarsnik hm I ran the update_ops script (with another rakudo perl 6) redid make && make install in Moar and make && make install rakudo and it's still fail
18:29 moritz raschipi: do you mean like -E in perl5?
18:30 Skarsnik Do I have to update a file in nqp?
18:30 raschipi Now that the Failure returns for failed matches were reversed, the toaster isn't detecting it anymore. People got a list of modules with problems and they're fixing it, but if it's introduced anew in another package, we won't know.
18:31 raschipi I don't mean it for normal usage, just for testing the upcoming changes to the language.
18:33 raschipi Like a RAKUDO_MONKEY_TESTING environment variable or something.
18:33 timotimo Skarsnik: you then also need to rebuild nqp
18:34 timotimo because nqp grabs the definitions of the moar ops from the installed files
18:34 Skarsnik this is so confusing xD
18:34 timotimo sorry
18:34 ugexe so make clean?
18:35 Skarsnik make remakeuniverse
18:36 Skarsnik Unhandled exception: Missing or wrong version of dependency '/home/skarsnik/perl6/socket/rakudo/install/bin/../share/nqp/lib/MAST/Nodes.nqp' (from 'src/Perl6/Pod.nqp')
18:36 Skarsnik duh
18:41 Skarsnik make clean worked
18:43 Skarsnik victory. skarsnik@devel:~/perl6/socket/rakudo$ ./install/bin/perl6 --ll-exception -e 'my $conn = IO::Socket::INET.new(:host<localhost>, :port(3333), :source-address<locahost>);'
18:43 Skarsnik Source address : locahost
18:48 AlexDaniel \o/
18:50 Skarsnik I just display the extra arg in moar for now x)
18:51 Skarsnik hm
18:51 Skarsnik can you use a hostname for a source-address
18:52 geekosaur it will have to resolve it to a single address; if it has multiple, it'd take the first applicable (i.e. ignoring ipv6 for an ipv4 socket)
18:54 lizmat AlexDaniel++  :-)
18:56 AlexDaniel \o/ thanks
19:00 Skarsnik Do I get NULL string if an argument is not given?
19:02 nine raschipi: why would you need to test modules?
19:03 AlexDaniel hm, I see that the lines are wrapped a bit in the release announcement that I've sent
19:09 AlexDaniel yup. Next time will make sure not to exceed 72 characters
19:10 Skarsnik is that 1980? xD
19:17 Geth ¦ rakudo/nom: 5db5b1dbfa | (Elizabeth Mattijsen)++ | src/core/Match.pm
19:17 Geth ¦ rakudo/nom: Fix for RT #131936
19:17 Geth ¦ rakudo/nom:
19:17 Geth ¦ rakudo/nom: Provide a HLL version of Match.(made|ast) that returns a Nil instead
19:17 Geth ¦ rakudo/nom: of a lowlevel NQPMu.  Wish there was a better way to test for NQPMu
19:17 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=131936
19:17 Geth ¦ rakudo/nom: though, as this will prevent type objects being properly propagated
19:17 Geth ¦ rakudo/nom: by "make".
19:17 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/5db5b1dbfa
19:19 geekosaur joined #perl6-dev
19:21 Skarsnik m: use NativeCall; sub foo(num $a) is native;
19:21 camelia rakudo-moar 69a91d: OUTPUT: «5===SORRY!5=== Error while compiling <tmp>?A unit-scoped sub definition is not allowed except on a MAIN sub;?Please use the block form.?at <tmp>:1?------> 3e NativeCall; sub foo(num $a) is native;7?5<EOL>?»
19:21 Skarsnik m: use NativeCall; sub foo(num $a) is native {};
19:21 camelia rakudo-moar 69a91d: OUTPUT: «5===SORRY!5=== Error while compiling <tmp>?Cannot resolve caller infix:<==>(NQPMu, Int); none of these signatures match:?    ($?)?    (\a, \b)?    (Real \a, Real \b)?    (Int:D \a, Int:D \b)?    (int $a, int $b)?    (Num:D \a, Num:D \b …»
19:26 Geth ¦ rakudo/nom: 8a0f6ac110 | (Elizabeth Mattijsen)++ | src/core/Junction.pm
19:26 Geth ¦ rakudo/nom: Simplify setting up auto-threading.
19:26 Geth ¦ rakudo/nom:
19:26 Geth ¦ rakudo/nom: We don't need to pollute the setting with two subs, if we can set them
19:26 Geth ¦ rakudo/nom: up using a pointy block.  This is spectest clean.  I'm assuming the
19:26 Geth ¦ rakudo/nom: named subs were needed at some point in the past.
19:26 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8a0f6ac110
19:27 Skarsnik hm interesting Rakudo does not close connection on exit
19:28 moritz the OS does
19:29 Skarsnik How I can test if this code work now x)
19:31 lizmat am I the only one not seeing the 2017.08 announcement here? https://www.nntp.perl.org/group/perl.perl6.compiler/2017/08/msg16371.html
19:32 Skarsnik it's empty
19:32 lizmat ok, so you see the same thing as me
19:32 lizmat :-(
19:33 lizmat do we have another place I could link to for the compiler release announcement ?
19:34 lizmat I guess https://github.com/rakudo/rakudo/blob/nom/docs/announce/2017.08.md will have to do
19:35 moritz that's what I used on da twitters
19:35 lizmat moritz++  # that's where I copied it from  :-)
19:57 Skarsnik I wonder if I should just add a socket_listen and expose socket_bind in nqp (and remove listen in bind)
20:26 Skarsnik hm interesting
20:29 Skarsnik I am not sure how to handle/fix that ./install/bin/perl6 -e 'my $conn = IO::Socket::INET.new(:host<127.0.0.1>, :port(80), :source-address<localhost>); $conn.print("GET index.html\r\n"); say $conn.recv()'
20:29 Skarsnik this does not work because it bind localhost to ipv6
20:29 ugexe moritz: can you change the appveyor setup to rakudo org so rakudo org members can restart builds?
20:29 Skarsnik but if I put 127.0.0.1 this work
20:31 Skarsnik AlexDaniel, This is apache log entry : (left is ip source) 127.0.0.1 - - [21/Aug/2017:22:28:28 +0200] "GET index.html" 400 301 "-" "-"
20:31 Skarsnik 10.0.2.15 - - [21/Aug/2017:22:30:37 +0200] "GET index.html" 400 301 "-" "-"
20:32 moritz ugexe: how would I do that?
20:35 ugexe https://www.appveyor.com/docs/team-setup/#setting-up-appveyor-account-for-github-organization moritz: looks like creating an "organizational" account
20:43 moritz ugexe: I'm trying it. Current problem: confirmation emails aren't delivered, so I can't confirm the organization account
20:45 moritz just asked them via twitter
20:45 moritz will try sleep next :-)
20:47 ugexe cool, thanks for trying
20:48 Skarsnik good night people :)
21:04 AlexDaniel “is that 1980?” no, but somebody had this wonderful idea to make it silently change the content of your message when you click “send” (thunderbird)
21:04 AlexDaniel this only happens with plaintext messages I think
21:06 lizmat and yet another Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2017/08/21/2017-34-going-atomic/
21:06 AlexDaniel ( or maybe it only matters with plaintext messages… :% )
21:29 travis-ci joined #perl6-dev
21:29 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Fix for RT #131936
21:29 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=131936
21:29 travis-ci https://travis-ci.org/rakudo/rakudo/builds/266918795 https://github.com/rakudo/rakudo/compare/69a91d62a477...5db5b1dbfa0b
21:29 travis-ci left #perl6-dev
21:29 buggable [travis build above] ? All failures are due to timeout (0), missing build log (0), GitHub connectivity (1), or failed make test (0).
22:41 MasterDuke joined #perl6-dev
22:46 MasterDuke AlexDaniel, Skarsnik: cool. i'd gotten as far as creating a MoarVM branch and checking that adding a uv_tcp_bind did in fact do something useful, but then got distracted by adding info to the `failed to resolve hostname` exception and haven't done anything else, so i'll leave it to Skarsnik
23:02 AlexDaniel cool. I'm actually surprised that people have any interest in this ticket
23:03 AlexDaniel I was thinking that I'm the only person who needs this practically
23:25 BenGoldberg joined #perl6-dev

| Channels | #perl6-dev index | Today | | Search | Google Search | Plain-Text | summary