| AlexDaniel | is there anybody here suffering from Malformed UTF-8 issue? | 00:32 | |
| :'| | 01:09 | ||
| MasterDuke | that seems to have dropped off the radar, you just started seeing it again? | 01:10 | |
| AlexDaniel | I have a file here that's failing on 257570th line, you know… | ||
| if I delete any file before that, it stops failing there | |||
| any line* | |||
| and of course it fails elsewhere, like 100000 lines further… | 01:11 | ||
| samcv | do you know if it's valid? | ||
| AlexDaniel | samcv: I don't know, but given that rakudo is able to parse it under certain conditions, I guess it is valid | 01:12 | |
| for example, I tried sed s/[A-Z]/_/; and it stopped failing | 01:13 | ||
| or, I tried piping into it in different ways | |||
| like from sed or something | |||
| and for the same input it may fail in different places | |||
| samcv | hm | ||
| AlexDaniel | “perl6 -ne '…' < golfed” fails | 01:15 | |
| same thing with “perl6 -ne '…' golfed” | |||
| “cat golfed | perl6 -ne '…'” and… obviously… this one is fine | 01:16 | ||
| except when you run it again, then it fails | |||
| does anybody want to play with it? | 01:17 | ||
| BenGoldberg | What if you give the program to a non-perl-based utf8 validator? | 01:18 | |
| Or maybe I should say, non-perl6-based validator ... | 01:19 | ||
| cat golfed | perl5 -ne '...' | |||
| AlexDaniel | iconv -f UTF-8 golfed -o /dev/null; echo $? # 0 | ||
| isutf8 golfed; echo $? # 0 | 01:20 | ||
| anyway, here's the file: files.progarm.org/golfed.gz | 01:22 | ||
| perl6 --ll-exception -ne 'say $++' < golfed | |||
| (say $++ makes it a bit slower, but it's easier to see what's happening this way) | |||
| this thing is driving me crazy actually | 01:23 | ||
| (when piping into it, make sure to do a couple of runs to see how inconsistent it is) | 01:24 | ||
| geekosaur | sounds more likely to be I/O related than unicode related | 01:33 | |
| like it's dropping bytes on the floor and if the utf8 decoder was in (or lands in) the middle of a sequence, boom | |||
| AlexDaniel | sure | 01:35 | |
| [Coke] | malformed utf-8 never stopped seeing it. | 01:40 | |
| AlexDaniel | \o/ | 01:41 | |
| this gives hope! | |||
| eh, ok: RT #131961 | 01:56 | ||
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=131961 | ||
| [Coke] | thanks | 01:57 | |
| Skarsnik | . | 07:32 | |
| Geth | rakudo/nom: 753c9a5ead | (Elizabeth Mattijsen)++ | src/core/Str.pm List.join is not guaranteed to return Str in the future |
08:17 | |
| nine | Zoffix: I fixed require.t and possibly other flappers by having S11-modules/perl6lib.t no longer create an empty Test.pm6 where other tests may find it. | ||
| lizmat | aaaaahhhh :-) | 08:19 | |
| Skarsnik | m: use NativeCall; sub foo(num $a) is native {}; | 08:27 | |
| camelia | 5===SORRY!5=== Error while compiling <tmp>?Cannot resolve caller infix:<==>(NQPMu, Int); none of these signatures match:? ($?)? (\a, \b)? (Real \a, Real \b)? (Int:D \a, Int:D \b)? (int $a, int $b)? (Num:D \a, Num:D \b … | ||
| Skarsnik | can someone push github.com/rakudo/rakudo/pull/1133 to fix this? (the build error is only for win32 when rakudo was not building on it) | 08:28 | |
| Geth | rakudo/nom: 4077842c87 | (Sylvain Colinet)++ | 2 files Fix NativeCall signature check routine dieing on unsuported native type (int, num). Change the tests to catch up this kinda of error |
08:35 | |
| rakudo/nom: 3518b13bc8 | lizmat++ (committed using GitHub Web editor) | 2 files Merge pull request #1133 from Skarsnik/nc-sig-check-fix Fix NativeCall signature check routine dieing on unsuported native type |
|||
| Skarsnik | thx :) | ||
| lizmat | Skarsnik: no, thank you! | 08:39 | |
| :-) | |||
| Skarsnik | dunno why suddendly it get an undef and not 0 | 08:40 | |
| m: say Num.^nativesize | 08:41 | ||
| camelia | No such method 'nativesize' for invocant of type 'Perl6::Metamodel::ClassHOW'? in block <unit> at <tmp> line 1?? | ||
| Skarsnik | m: say num.^nativesize | ||
| camelia | No such method 'gist' for invocant of type 'NQPMu'. Did you mean 'isa'?? in block <unit> at <tmp> line 1?? | ||
| Skarsnik | should it return Nil or just not be defined at all§ | 08:43 | |
| m: num.^nativesize == 0 | 08:45 | ||
| camelia | WARNINGS for <tmp>:?Useless use of "==" in expression ".^nativesize == 0" in sink context (line 1)?Cannot resolve caller infix:<==>(NQPMu, Int); none of these signatures match:? ($?)? (\a, \b)? (Real \a, Real \b)? (Int:D \a, Int:D… | ||
| Skarsnik | bisectable6, num.^nativesize == 0 | ||
| bisectable6 | Skarsnik, Bisecting by exit code (old=2015.12 new=3518b13). Old exit code: 0 | ||
| Skarsnik, bisect log: gist.github.com/93e0e96d65525813b8...0f4c69f552 | |||
| Skarsnik, (2017-04-01) github.com/rakudo/rakudo/commit/af...b555485618 | |||
| Skarsnik | m: uint.^nativesize == 0 | 08:47 | |
| camelia | WARNINGS for <tmp>:?Useless use of "==" in expression ".^nativesize == 0" in sink context (line 1)?Cannot resolve caller infix:<==>(NQPMu, Int); none of these signatures match:? ($?)? (\a, \b)? (Real \a, Real \b)? (Int:D \a, Int:D… | ||
| Geth | rakudo/jit_nativecall: ac9ea4fe7c | (Stefan Seifert)++ | 2 files Test Pointer return values in argless function |
09:12 | |
| rakudo/nom: 6cec6b7218 | (Elizabeth Mattijsen)++ | src/core/Hash.pm Fix for RT #126951 Sorry, this appears to have fallen through the cracks. Please note though that a FLATTENABLE_HASH cannot represent all object hashes correctly, only the ones that happen to have Str for keys. If the key is not a Str, it will be coerced to Str in the returned hash |
09:31 | ||
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=126951 | ||
| nqp: c368470ddf | (Samantha McVey)++ | 2 files Add new --ignore-errors flag to ./Configure.pl Will allow you to compile nqp even if MoarVM/JVM versions are not high enough. May ignore certain other errors as well. It could be extended to allow functioning when other errors occur, but that will need to be added as needed, since the script often doesn't pass information along when failing conditions are encountered. |
11:16 | ||
| timotimo | nice | 11:40 | |
| nine | Regarding 6.d, I'd like to talk about whether to prohibit a use v6.d; anywhere but in the top most scope of a computation unit and maybe even anywhere but as the very first statement there. | 11:51 | |
| Also if we want to continue working in the direction of methods being able to get the caller's language version, it may be useful to have Language objects (stored in a $?LANGUAGE or something like that in the comp units outer lexical scope). | 11:54 | ||
| If those are singletons (one per language/version), a comparison might become pretty cheap. Except of course if you need to compare with > | 11:56 | ||
| Geth | nqp: e5e5a3f03c | (Samantha McVey)++ | Configure.pl Fix a bug in the last commit, that could cause configure errors Previous commit added a line to print out errors that wouldn't get printed out otherwise if you pass in --ignore-errors. Doing this though would cause builds without --ignore-errors to fail as often the messages may not be fatal messages. Correct this. |
12:38 | |
| rakudo/nom: a5014fd085 | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm Fix RT #131962 - Make sure we accept Mu as a value - Use IterationEnd as a sentinel value on JVM (untested) |
13:16 | ||
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=131962 | ||
| rakudo/nom: e8137b45c5 | (Elizabeth Mattijsen)++ | src/core/Junction.pm Streamline Junction.defined|Bool|ACCEPTS - HLLize the value in only one place - Remove unnecessary parens - use nqp::isfalse($a) instead of nqp::if($a,0,1) |
13:24 | ||
| rakudo/jit_nativecall: 85c1e8fd39 | (Stefan Seifert)++ | 2 files Test argless native functions returning strings |
13:26 | ||
| stmuk | I'm having problems with IO::Path::ChildSecure (dependency for Cro) | 13:35 | |
| I don't seem to be able to install on .07 or .08 | 13:36 | ||
| hmm probably why --/test is suggested | 13:54 | ||
| dogbert11 | how's the conference going? | 14:18 | |
| lizmat | watching Sue Spence atm | 14:29 | |
| dogbert11 | aha, web crawling | 14:34 | |
| TimToady | nine: the introspection approach kinda smells, and doesn't solve caller's caller; the approach suggested at irclog.perlgeek.de/perl6-dev/2017-...i_15070774 doesn't involve introspection, but has other issues if you rely on the contructor from a library that upgrades to 6.d | 16:45 | |
| I suppose a constructor could be written to introspect and decide whether to return a 6.c or a 6.d object, but then that would stick via the type, and every method wouldn't have to introspect, just constructors | 16:47 | ||
| that might be the Least Worst solution | |||
| but the intent has always been that types have shortnames and longnames, and that shortnames are a lexically scoped alias to the longname for convenience | 16:49 | ||
| b2gills | I just ran my shell script I use for pulling and compiling Rakudo, and it just went below 3 minutes for what I think is the first time. | 16:52 | |
| AlexDaniel | releasable6: status | 16:56 | |
| releasable6 | AlexDaniel, Next release in 21 days and ?2 hours. R6 is down. Changelog for this release was not started yet | ||
| AlexDaniel, Details: gist.github.com/dbaedb3f2317e71c1e...36cd7c8d59 | |||
| AlexDaniel | heh, instead of showing its guts it says “R6 is down”, cool bot | 16:57 | |
| .tell Zoffix so what's up with perl6.fail? I think you were planning to use another domain for it? | |||
| yoleaux | AlexDaniel: I'll pass your message to Zoffix. | ||
| timotimo | TimToady: what's the obvious reason we have scalar containers just cloned and turned RO when passed as arguments? | ||
| you know, the one that just escaped my mind | 16:58 | ||
| TimToady | because you can avoid even that if you can prove the container isn't written to, and just do things by reference | 16:59 | |
| timotimo | i think also because we need to conserve "this is an item" for anything passed in a $foo | 17:02 | |
| that's probably the reason i almost thought of | |||
| TimToady | another thing that can be optimized away if provably not used | 17:04 | |
| timotimo | right. most likely blocked by us not speshing into inlined spseh graphs yet | 17:05 | |
| i'll close my laptop to take part in the social meeting | 17:07 | ||
| Geth | roast: 1089a7f9b7 | (Moritz Lenz)++ | S05-metasyntax/lookaround.t RT #131964: tests for ^, ^^, $, $$ in <?after> |
18:12 | |
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=131964 | ||
| samcv | woot! \o/ got stacktraces printing a full backtrace on travis now travis-ci.org/MoarVM/MoarVM/jobs/2...1486-L1545 finally can figure out this phantom segfault only happening on master MoarVM. but not on travis if it's my build or a PR build | 18:49 | |
| no clue how that could be but.. | |||
| AlexDaniel | samcv++ # p6daily :) | 19:12 | |
| samcv | heh :) | ||
| not sure how often segfaults happen in nqp or rakudo branches if we want that in non-moarvm projects as well | 19:13 | ||
| Zoffix | AlexDaniel: it expired. New URL is fail.rakudo.party/ | 19:36 | |
| yoleaux | 16:57Z <AlexDaniel> Zoffix: so what's up with perl6.fail? I think you were planning to use another domain for it? | ||
| AlexDaniel | alright | 19:37 | |
| Zoffix | stmuk: likely failing in Temp::Path used by its tests. It's flopping and I haven't determined why yet | 19:53 | |
| Geth | rakudo/nom: f097e551bd | (Samantha McVey)++ | tools/build/create-moar-runner.pl perl6-valgrind-m: remove --full-cleanup, add ENV conf. var Remove --full-cleanup because it causes confusing valgrind output, (currently not working properly). Add a MVM_VALGRIND_OPTS environment variable so you can add options to valgrind easily. |
23:07 | |
| Skarsnik | interesting | 23:09 | |
| AlexDaniel | m: my @x[0]; | 23:21 | |
| camelia | Illegal dimension in shape: 0. All dimensions must be integers bigger than 0? in block <unit> at <tmp> line 1?? | ||
| AlexDaniel | m: my int @x[0]; | ||
| camelia | ( no output ) | ||
| AlexDaniel | m: my int @x[0]; @x[0] = 42 | ||
| camelia | Index 0 for dimension 1 out of range (must be 0..-1)? in block <unit> at <tmp> line 1?? | ||
| AlexDaniel | m: my int @x[-1]; | ||
| camelia | MoarVM panic: Memory allocation failed; could not allocate 18446744073709551608 bytes? | ||
| AlexDaniel | it feels like we had this already | 23:24 | |
| arrays with negative size | |||
| m: my int @x[-2**63]; say @x.shape | |||
| camelia | (-9223372036854775808)? | ||
| AlexDaniel | RT #126800 | 23:25 | |
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=126800 | ||
| AlexDaniel | hmmm | ||
| I created a new ticket instead of reopening the old one: RT #131966 | 23:43 | ||
| synopsebot6 | Link: rt.perl.org/rt3/Public/Bug/Display...?id=131966 | ||
| samcv | AlexDaniel, the problem i've diagnosed to that it's probably a garbage collection problem | 23:51 | |
| causing corruption | |||
| AlexDaniel | samcv: you mean Malformed UTF-8 thingy? | ||
| samcv | yeah that |