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.
Geth nqp: 176960303f | (Stefan Seifert)++ | 15 files
Fix bytecode writing on big endian systems
 08:23
|Tux| Rakudo version 2018.11-22-gfd216fec4 - MoarVM version 2018.11-18-gb8ba4707c
csv-ip5xs0.910 - 0.922
csv-ip5xs-206.936 - 7.161
csv-parser21.166 - 22.479
csv-test-xs-200.432 - 0.436
test7.381 - 7.734
test-t1.732 - 1.748
test-t --race0.781 - 0.817
test-t-2030.067 - 31.514
test-t-20 --race9.263 - 9.413
 11:38
dogbert2_ .seen lizmat 13:20
yoleaux I saw lizmat 11:17Z in #perl6: <lizmat> .
lizmat you rang ?
dogbert2_ lizmat: yes 13:21
can I ask you to do a couple of things?
first, a bump to get nine++ recent Moar and NQP commits
if possible 13:22
lizmat: secondlym can you close R#2361. I'm not allowed to do that? 13:23
lizmat dogbert2_ will do
dogbert2_ +++lizmat
lizmat hmmm.... looks like nqp already got bumped ? 13:27
dogbert2_ lizmat: when did that happen? 13:39
lizmat 09:22:21 +0100 this morning
lucasb a bump? 13:40
lizmat as part of nine++ work on getting bytecode writing corrected on big endian systems 13:41
lucasb ah, sorry, lizmat, you're right
I was looking for a stand-alone commit :)
lizmat testing rakudo now 13:42
dogbert2_ but then I guess we need to bump MoarVM? 13:43
lucasb dogbert2_: MOAR_REVISION was updated in nqp in the last commit 13:44
dogbert2_ I just got the impression that MoarVM wasn't rebuilt (on my system) 13:51
shouldn't tools/build/NQP_REVISION be updated? 13:52
lizmat dogbert2_: yeah it should, after I determined it builds and tests ok :-) 13:55
dogbert2_ :-) 13:57
lucasb when MOAR_REVISION is changed as part of another commit in nqp, the "bump commit message" is missed, no? which tool generates that commit "[MoarVM Bump] Brings N commits. MoarVM bump brought: COMMIT_RANGE" ?
shouldn't all bump commits be stand-alone commits? 13:59
timotimo i think the Z tool does that
lucasb (stand-alone commits for bisectabilty purporses) 14:00
s/r// 14:01
Geth rakudo: 46d4c9fe96 | (Elizabeth Mattijsen)++ | tools/build/NQP_REVISION
Bump NQP to get nine++ latest bigendian fixes
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....3-g1769603
nine lucasb: bisectability is exactly why I included the bump in the commit. Because the commit absolutely needs the accompanying changes in MoarVM 14:05
lizmat ufobat___: added another answer to your SUB-MAIN-OPTS question 14:14
ufobat___ on SO?
lizmat yup :-)
ufobat___ let me read 14:15
smart! 14:16
lizmat it's a bit dirty, but in the context of a CLI script, I think it's ok 14:17
and you could consider it cleaner than exporting %SUB-MAIN-OPTS 14:18
ufobat___ thank you. lizmat++
lizmat OTOH, if the client has it's own my %*SUB-MAIN-OPTS, the PROCESS:: one will not be seen
so you're setting up a default at best
lucasb nine: ok 14:23
but a moarvm bump brings other commits (maybe most are associated with this specific issue), but any problem in those unrelated will be associated with "Fix bytecode writing on big endian systems", which may be non-intuitive, no? 14:25
FTR, since Geth didn't show: github.com/MoarVM/MoarVM/compare/2...gb8ba4707c 14:26
robertle nine: was away for a while and did not follow this, but do you have more commits? the bump discussion above sounds like you have a big-endian fix, but I can't see any new commits... 14:29
nine robertle: no, I'm at $work 14:40
robertle ah, ok, k! 14:47
me too, that's why I am so confused :)
dogbert2_ robertle: the bump is necessary if you build from your rakudo directory, i.e. with 'perl Configure.pl --gen-moar --gen-nqp --backends=moar' 15:06
if you build MoarVM and nqp yourself you don't need it 15:07
AlexDaniel nine: fwiw that doesn't affect bisectablity in any way for proper tools than know how to skip broken builds and stuff 15:30
if I understood the situation correctly :)
lucasb: yeah, some bumps are fatter than the others. I think there are two ways to improve it. Either force everyone to bump more often, or “fix” bisectable to know how to deal with bumps 15:32
lucasb: there's this ticket: github.com/perl6/whateverable/issues/40 15:33
I attempted to do it some time ago but then the code grew like three times or so, I didn't like that and stopped…
(talking about the whateverable build script) 15:34
if somebody wants to give it a go that'd be AWESOME 15:35
and yes, that'd also automatically make it work for whateverable-based scripts that bisect locally 15:36
and once we add “build on demand” functionality that'd also make it easier to bisect on mips :) 15:37
(not for that last issue maybe, but we did need it in the past)
nine AlexDaniel: I was talking about a plain git bisect 15:47
AlexDaniel plain git bisect is also fine if you skip properly 15:50
or exit 125 if you're doing it with a script
nine Upgrading the powerpc image to a somewhat current debian version failed...with no useful error message 16:51
lizmat notable6: weekly 17:46
notable6 lizmat, 13 notes: gist.github.com/247e86dbb302a2d2b8...5c0a850307
AlexDaniel ooh it's Monday!
reportable6: 2018-12-03T00:00:00Z 2018-12-10T00:00:00Z 17:47
reportable6 AlexDaniel, OK, working on it! This may take up to 40 seconds
AlexDaniel, gist.github.com/86ea3e1c0b6a885e1d...59cc844bb1 17:48
AlexDaniel lizmat: ↑ 17:49
lizmat AlexDaniel: got it :-)
AlexDaniel releasable6: status
releasable6 AlexDaniel, Next release in ≈5 days and ≈1 hour. 4 blockers. 0 out of 23 commits logged
AlexDaniel, Details: gist.github.com/0e371a5cada3cb39a6...4b17c34e38
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/12/10/...li-landed/ 18:56
notable6: weekly reset 18:58
notable6 lizmat, Moved existing notes to “weekly_2018-12-10T18:58:07Z”
releasable6 Next release in ≈4 days and ≈23 hours. 4 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
lizmat moritz: ^^ 19:02
lucasb I saw these recent issues labeled with @LARRY and an idea occurred me. We all would like his presence all the time, but he is not always available. So, unfortunately, seeing that label gives me the unconscious feeling that these are the issues that will never be dealt with. 19:41
What the label is used for? Design decisions? So, can that label be renamed to "design-decisions" so that you, the language design team, can make decisions on those issues? 19:42
lizmat well, Larry by himself is $LARRY, that's why the label is @LARRY :-) 20:02
but yes, perhaps we should be more assertive on these matters
and renaming the label might be a first step 20:03
lucasb wow, #WeAreAllLarries :)
lizmat
.oO( I am Spartacus! )
AlexDaniel lucasb: I agree that it should be renamed 20:30
gfldex I got a 7% speed up with my json-munch-benchmark from about 2 weeks ago.
AlexDaniel that's not the first time I see that question, yes it's easier to just rename it than explain every time
gfldex More of the same please! :)
Geth nqp/master: 6 commits pushed by (Paweł Murias)++ 21:17
moritz lizmat: tweeted 21:24
lizmat moritz++ 21:25
Geth roast: 27476f9f2e | (Elizabeth Mattijsen)++ | S12-introspection/attributes.t
Add tests for R#2521
 21:33
timotimo jnthn: i'm not sure what the safe way would be to find the right $tap object in the exception handler for LAST inside !run-supply-code; could it be it should really go somewhere else instead? 21:39
oh, !run-supply-code is actually called by the block inside add-whenever, and that has the $tap available 21:46
jnthn Dunno, I'm way too tired to look at that :) 21:48
timotimo that's fair :) 21:51
i'll have to manually figure out if there's phasers for LAST inside the whenever block, right?
jnthn Guess so 21:58
Well, we must run them somewhere :) 21:59
timotimo aye 22:01
well, anyway, right now it's misbehaving :D 22:02
Geth nqp: b4de55eda6 | (Elizabeth Mattijsen)++ | docs/ops.markdown
Document the new readint/writeint ops
 22:09
lizmat nine: I assume you did not implement the view/viewdim ops, right ? 22:10
timotimo don't think so
nine no 22:12
Geth rakudo: 890d628a24 | (Timo Paulssen)++ | src/core/Supply.pm6
Support "last" And "LAST" In Whenever Blocks
 22:54
timotimo jnthn: committed :) 22:55
jnthn timotimo: Cool. Remember tests ;) 22:57
timotimo hum, "tests", what was that again ... 22:58
jnthn sleep & 23:01
evalable6 Use of uninitialized value of type Callable in numeric context
in block <unit> at /tmp/T…
jnthn, Full output: gist.github.com/6233b79f46e6a3beb6...21bcce405a
timotimo :D 23:02
we'll possibly want to support redo too at some point? 23:04

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!