| Zoffix | [Coke]: you were wondering why people put [BUG]... This page actually says to do so: rakudo.org/tickets/ | 00:44 | |
| Zoffix edits | |||
| It also links to stuff like "easy" tickets.... But there's nothing these. If a list of easy isn't maintained, perhaps the links should be removed? | |||
| AlexDaniel | yea | 01:05 | |
| Zoffix | squashable6: status | ||
| squashable6 | Zoffix, ???? SQUASHathon is in progress! The end of the event in ?10 hours | ||
| Zoffix, Log and stats: gist.github.com/70c886a1da9002015e...38220d0b2c | |||
| Zoffix | Pretty good turn out 250+ commits. | 01:07 | |
| Zoffix goes to spend the rest of the long weekend away from humanity | 01:08 | ||
| Gonna continue learning C :) | |||
| \o | |||
| yoleaux | Zoffix: fix up `make test` along with other tests | 01:14 | |
| samcv | Zoffix, fun! | 01:15 | |
| Zoffix | hah | 01:17 | |
| yoleaux: you party-pooper! | |||
| releasable6: status | |||
| releasable6 | Zoffix, Next release in 13 days and ?17 hours. No blockers. Changelog for this release was not started yet | ||
| Zoffix, Details: gist.github.com/5123e8fce184af5169...42fd727b24 | |||
| Zoffix | .in 6d fix up `make test` along with other test (make all the tests in rakudo's t/ actually get run when make test is run) | ||
| yoleaux | Zoffix: I'll remind you on 9 Sep 2017 01:17Z | ||
| Zoffix &'s | |||
| AlexDaniel | oh wow, just 13 days | 02:14 | |
| time flies | |||
| Geth | rakudo/nom: 51c3d86ccf | (Elizabeth Mattijsen)++ | src/core/Junction.pm Make all(@a),none(@a),one(@a) aboutb 9x faster - for a 10 element array - by creating multis for the @a case - causing not to have to go through one arg handling in that case This was also done for any(@a), but this caused tests 10, 16, 85, 91, 98, 101 of S02-literals/radix.t to fail: instead of a number, a Failure was being generated. Looking at --ll-exception output, it is completelt unclear (to me at least) where the Failure is actually being generated. Some more eyes on this would be appreciated! |
12:42 | |
| rakudo/jit_nativecall: 502b0b1d66 | (Stefan Seifert)++ | lib/NativeCall.pm6 Remove unnecessary creation of an extra serialization context Turns out we don't need this and it will actively prevent spesh from optimizing as it won't be able to find the WVal for self. Furthermore this would also prevent us from JIT compiling the sub's body. |
14:51 | ||
| roast: skids++ created pull request #304: Add test for RT#124247 |
15:39 | ||
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=124247 | ||
| Geth | roast: 3ef8bb0331 | skids++ | S19-command-line-options/03-dash-p.t Add test for RT#124247 |
||
| roast: ec1e1cf2c5 | skids++ (committed using GitHub Web editor) | S19-command-line-options/03-dash-p.t Merge pull request #304 from skids/rt124247 Add test for RT#124247 |
|||
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=124247 | ||
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=124247 | ||
| rakudo/jit_nativecall: 987cd9824f | (Stefan Seifert)++ | lib/NativeCall.pm6 Fix "Object in QAST::WVal, but not in SC" for function pointers While adding the sub to a newly created SC is harmful in the normal case, it's vital when dealing with function pointers return from native code as those subs will not be added to an SC by the compiler. |
15:40 | ||
| rakudo: MasterDuke17++ created pull request #1150: Fix INTERPOLATE call with wrong args in p5metachar |
19:23 | ||
| rakudo/nom: 4c25df745c | MasterDuke17++ | src/Perl6/Actions.nqp Fix INTERPOLATE call with wrong args in p5metachar I'm not sure how to craft a regex to demonstrate the bug, but printing out the argument values in INTPEROLATE shows that (e.g., `$monkey` is always set and `$s` is what changes depending on whether a `use MONKEY` is in scope, which is wrong). Some discussion about this here irclog.perlgeek.de/perl6-dev/2017-...i_15101225 |
20:29 | ||
| rakudo/nom: 9a478b1ea1 | niner++ (committed using GitHub Web editor) | src/Perl6/Actions.nqp Merge pull request #1150 from MasterDuke17/fix_INTERPOLATE_called_with_wrong_args_in_p5metachar Fix INTERPOLATE call with wrong args in p5metachar |
|||
| MasterDuke | "INTPEROLATE", ugh. and unnecessary parens around the e.g. clause. i need an editor for my commit messages | 20:32 | |
| AlexDaniel | MasterDuke: just configure spell checking in vim? | 20:57 | |
| MasterDuke | i should do that... | 21:01 | |
| [Coke] | so, cro recommends installing sans tests. I ignored this, and no doubt, an SSL test failed. | 21:08 | |
| [Coke] tries upgrading rakudo | 21:09 | ||
| jnthn | [Coke]: A Cro::SSL test, or a IO::Socket::Async::SSL test? | 21:18 | |
| [Coke] | jnthn: I retried, am currently stuck here: github.com/jnthn/p6-io-socket-async-ssl/issues | 21:20 | |
| just cloned it to see if I can get more deets. | 21:21 | ||
| t/meta.t | |||
| jnthn | It fails...t/meta.t? :/ | ||
| *sigh* | 21:22 | ||
| jnthn decides that t/meta.t is always going to be xt/meta.t for all of this modules, or just not going to exist at all | |||
| [Coke] | using 2017.08-66-g9a478b1ea | ||
| jnthn | What's the failure mode, ooc? | ||
| [Coke] | Could not find META6:ver<0.0.4..*> at line 66 in: | ||
| jnthn | *sigh* | ||
| I forget the ins and outs of that | 21:23 | ||
| But effectively, some version of one of the Test::META dependencies had an issue, and having an old version of that can cuase issues | 21:24 | ||
| [Coke] | one sec. | 21:26 | |
| I have Test::META:ver('0.0.9'):auth('github:jonathanstowe') | |||
| jnthn | Well, IO::Socket::Async::SSL no longer depends on it, as of my latest push :) | 21:27 | |
| [Coke] | ... zef upgrade Test::META says "are already at their latest versions", then shows "0.0.13", but 'zef info Test::META' shows 0.0.9 | 21:28 | |
| jnthn | That may or may not be a now-fixed zef bug where version comps were done wrong | 21:29 | |
| [Coke] | ... This is why I never end up doing any interesting work. :) | ||
| jnthn | Yes, well, I'd like to, and I've wasted too many hours of my time on Test::META related shenanigans, so it's banned in any module I maintain from here on in. :) | 21:30 | |
| [Coke] | ok. upgrading zef. | ||
| [Coke] goes back to 'zef install cro' to see what's next. | 21:33 | ||
| pretty sure zef has cached that version of IO::Socket::Async::SSL (since there was no version bump). seeing if I can force it. | 21:35 | ||
| jnthn | I did a version bump a few minutes ago | 21:37 | |
| [Coke] | zef was still trying to install 0.2, doing an install from a git clone. | ||
| jnthn | Since there was a few other PRs merged since the last bump also :) | ||
| But I guess it takes a bit of time for that to be picked up on | |||
| [Coke] | are the tests very slow? | 21:38 | |
| ... ran it once. very slow, nothing happened, killed it, re-ran it, boom done. | 21:40 | ||
| Zoffix: now getting the same Test::META error on github.com/zoffixznet/perl6-IO-Path-ChildSecure | 21:42 | ||
| is jonathanstowe on here? | 21:44 | ||
| moving over to toolchain | 21:45 | ||
| geekosaur | pretty sure you're talking to him | ||
| jnthn | geekosaur: Heh, no, I'm not that Jonathan :) | 21:46 | |
| Zoffix | that jonathan is RabidGravy | 21:50 | |
| FWIW, I've just pushed a commit removing Test::META from IO::P::ChildSecure deps; since it's an author test. Gonna do the same for all my modules later on. | |||
| [Coke]: though I think that error is cause your zef's too old. Try zef --force --/cached install github.com/ugexe/zef/archive/master.zip | 21:52 | ||
| Zoffix & | |||
| [Coke] | Zoffix: my zef is 0.1.27 | 21:56 | |
| doesn't --force also disable tests? | 21:57 | ||
| jnthn: ok: back to Cro::SSL test failure. | 21:58 | ||
| # Failed test 'Can connect to a Cro::SSL::Listener with ALPN set up with ALPN from client' | |||
| clone-ing. | 21:59 | ||
| Unhandled exception in code scheduled on thread 12 | 22:01 | ||
| Cannot locate symbol 'SSL_CTX_set_alpn_select_cb' in native library 'libssl.dylib' | |||
| jnthn | Ah. | 22:03 | |
| Please file that one. We should be detecting when the current SSL library version doesn't support ALPN. | 22:04 | ||
| I suspect some tests aren't doing that. | |||
| (the detection function is already there, just perhaps not being called where it should be) | |||
| [Coke] | jnthn: github.com/croservices/cro-ssl/issues/1 | 22:05 | |
| jnthn | [Coke]++ | 22:06 | |
| [Coke] | rakudo-j dead? | 23:41 | |
| gist.github.com/coke/884246d2b85ad...bea76def61 | 23:43 | ||
| MasterDuke | [Coke]: i believe it was the commit (nqp or rakudo, not sure) removing . from some lib path | 23:47 |