[Coke] tbrowder: how did you send in the cla? 01:46
yoleaux 20 Jan 2018 14:23Z <tbrowder> [Coke]: can please check on my CLA application? (Thomas M. Browder, Jr.; aka Tom Browder)
Geth roast/6.c-errata: 12bf2b8ae0 | (Zoffix Znet)++ (committed by Aleks-Daniel Jakimenko-Aleksejev) | S32-num/rat.t
Use correcter tests for Num.Rat conversions

The test descriptions say `==` is being checked, but the `is` routine is actually testing Str views of Rats and Nums. Change to `cmp-ok` with
  `==` as comparator, since IEEE-style zero-denominator Rats are explosive
in Str views, while non-explosive in Num views, so they will match here.
... (16 more lines)		 06:54
AlexDaniel .tell Zoffix Awesome, thanks. Committed the change to 6.c-errata. 06:55
yoleaux 20 Jan 2018 23:57Z <Zoffix> AlexDaniel: irclog.perlgeek.de/perl6-dev/2018-...i_15716502
AlexDaniel: I'll pass your message to Zoffix.
Geth rakudo: c52f323ece | (Aleks-Daniel Jakimenko-Aleksejev)++ | Configure.pl
Revert "Ensure Configure.pl uses gmake for builds on BSDs"

This reverts commit 4f865671685c6d53b53dfacf4beb9e5faf8aecf8.
  See rakudo/rakudo#1420		 07:00
synopsebot RAKUDO#1420 [open]: github.com/rakudo/rakudo/issues/1420 [regression][⚠ blocker ⚠] FreeBSD issues
Geth nqp: 9222147e47 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 3 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...ga04d1099b a04d1099b Revert "Try to use egcc when compiling on OpenBSD" 9df7c1645 Merge pull request #784 from MasterDuke17/fix_parse_jitgraph c602235ef Un-bitrot tools/parse_jitgraph.p6		 07:02
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...ga04d1099b
rakudo: c6f05001fe | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[NQP Bump] 9222147e4 [MoarVM Bump] Brings 3 co […]

NQP bump brought: github.com/perl6/nqp/compare/2017....g9222147e4
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2017....g9222147e4
AlexDaniel .tell lizmat github.com/rakudo/rakudo/issues/14...-359231728 08:20
yoleaux AlexDaniel: I'll pass your message to lizmat.
AlexDaniel .tell lizmat btw, I'm looking at this commit github.com/rakudo/rakudo/commit/5c228333 but failing to see what changed practically (NaN..NaN).roll(2) still hangs for me. Is everything alright there? 08:26
yoleaux AlexDaniel: I'll pass your message to lizmat.
Geth rakudo/post-release: 5793e6fe87 | (Zoffix Znet)++ | 2 files
Put IO grant work into "archive" dir
 09:39
lizmat . 11:08
yoleaux 08:20Z <AlexDaniel> lizmat: github.com/rakudo/rakudo/issues/14...-359231728
08:26Z <AlexDaniel> lizmat: btw, I'm looking at this commit github.com/rakudo/rakudo/commit/5c228333 but failing to see what changed practically (NaN..NaN).roll(2) still hangs for me. Is everything alright there?
lizmat m: say (NaN..NaN).elems
camelia (timeout) 11:09
lizmat Files=1232, Tests=76104, 326 wallclock secs (14.93 usr 5.45 sys + 2257.27 cusr 232.09 csys = 2509.74 CPU) 11:12
tbrowder .tell [Coke] I sent the CLA via snail mail (USPS) to the address on the form. 11:18
yoleaux tbrowder: I'll pass your message to [Coke].
stmuk I'm seeing some roast failures .. should I be testing against errata at this point? 11:52
pmurias samcv: I have looked into rt.perl.org/Public/Bug/Display.html?id=132233 12:02
samcv: the bug with that is that a ligature is a single grapheme while the letters it's composed of are 2 graphemes 12:03
samcv: as such we can't use eqatim and then advancing by the length of the ligature (or for that matter the stuff it's composed of) 12:04
samcv: one fix would be to make eqat* ops return the length of the stuff that's matched 12:06
lizmat 0 no match, > 0 number of codepoints or graphemes or bytes ? 12:10
pmurias lizmat: not sure, maybe using -1 for no match would be more sane 12:24
lizmat wouldn't that be the same as 0 ? 12:25
aka false ?
lizmat worked for a long time on a system where < 0 was true and >= 0 was false
pmurias what about using a "" needle? 12:26
lizmat the advantage was that you could $cond ?? minusone !! zero !! one !! two (pseudocode)
what does a "" needle with eqat mean ? 12:27
I know it has meaning for split, but for eqat ?
pmurias does $haystack have an empty string needle at position $pos? 12:28
lizmat you mean it's the end of the string ? 12:29
pmurias nqp::eqat("xfoobar", "foo", 1) is currently 1 12:32
lizmat: nqp::eqat with a "" needle return 0 if the position is inside the string 12:33
(currently)
lizmat ah, hmmm...
and if it is not ?
pmurias sorry meant 1 12:34
it returns 0 if it's not inside the string
nqp::eqat checks for $needle being a substring of $haystack starting at $pos 12:35
samcv pmurias: yeah i had thought about returning the length of what has been matched
lizmat m: use nqp; say nqp::eqat("foo","",3) # with a strange definition of "inside string" :-)
camelia 1
samcv and i think one of the functions in moarvm internally returns that
pmurias has to leave visit grandparents 12:36
samcv but i need to go to bed
but yeah i'd thought about returning the length as well. also i'm not sure where the code is that causes the bug
pmurias m: use nqp:; say nqp::eqat("foo, "", 4)
camelia 5===SORRY!5=== Error while compiling <tmp>
Confused
at <tmp>:1
------> 3use nqp7⏏5:; say nqp::eqat("foo, "", 4)
samcv if you know where it is, can you reply to the bug with info on where it is?
pmurias m: use nqp; say nqp::eqat("foo, "", 4)
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in argument list; couldn't find final ')' (corresponding starter was at line 1)
at <tmp>:1
------> 3use nqp; say nqp::eqat("foo, "7⏏5", 4)
expecting any of:
…
pmurias m: use nqp; say nqp::eqat("foo", "", 4)
samcv: ok
samcv since it's only a bug for interpolated thinges
camelia 0 12:37
pmurias I have the exact place, will post once I come back
[Tux] Rakudo version 2017.12-277-gc6f05001f - MoarVM version 2017.12.1-62-ga04d1099b
csv-ip5xs1.132 - 1.148
csv-ip5xs-2012.602 - 13.093
csv-parser12.277 - 12.428
csv-test-xs-200.424 - 0.437
test11.072 - 11.120
test-t2.768 - 2.779
test-t --race1.154 - 1.164
test-t-2049.042 - 49.394
test-t-20 --race17.357 - 18.455
 12:47
AlexDaniel stmuk: “I'm seeing some roast failures .. should I be testing against errata at this point?” both should be green ideally 13:02
there's one know failure in v6.c-errata which is not fixed yet
everything else was fine last time I checked
(but note that releasable still won't let it through if at least something is failing) 13:03
stmuk: “I was enable to build” – sorry, “able” or “unable”? 13:06
stmuk enable 13:07
not only were the BSD changes broken but unneeded
and the reverts fixed them 13:08
AlexDaniel OK, good!
Geth rakudo: 55cb8f3d61 | (Elizabeth Mattijsen)++ | src/core/Range.pm
Fix Ranges with NaN hanging

Affected were any methods that would call .elems, since .elems was hanging. Methods such as .pick / .roll etc.		 13:10
lizmat AlexDaniel: ^^^
AlexDaniel stmuk: thanks! 13:11
lizmat: great! Thanks. What about github.com/rakudo/rakudo/issues/14...-359231728 ? 13:12
lizmat: like, do we have to revert 729caa8f1 also or not?
lizmat I already commented on github.com/rakudo/rakudo/commit/72...978a14ff16 that it should also be reverted :-) 13:13
I guess I put the comment at the wrong place :-)
sorry
AlexDaniel hmmmm… 13:18
lizmat ?
guess I was not awake yet: saw your question, looked at the issues, followed link to commit and commented 13:19
AlexDaniel lizmat: actually, I think it was “autoreverted” in my previous revert
lizmat without going back to the issue
ok, then we should be good right ? 13:20
AlexDaniel yea
yeah, looks right
lizmat TimToady: wouldn't you say that Tuple is basically an immutable subclass of IterationBuffer? 13:21
AlexDaniel lizmat: ok, now about github.com/perl6/roast/commit/904a...2468a6bc97 which I assume was the result of github.com/rakudo/rakudo/commit/202459ce 13:25
lizmat: what's the current status of this? Do we go with changing the tests in errata or doing something with that rakudo commit?
lizmat looks 13:26
yeah, I think we need to change that test in 6.c-errata as well 13:28
the checks in 6.c were too specific
shall I change them? 13:29
lizmat changes them 13:32
Geth roast/6.c-errata: b688d39d49 | (Elizabeth Mattijsen)++ | S02-types/WHICH.t
Handle existence of ValueObjAt in 6.c

Basically redo of github.com/perl6/roast/commit/904ae98f034e7
Please revert if not appropriate after all.		 13:34
AlexDaniel OK 13:42
thanks
jnthn: here? 14:23
jnthn: anyway, your dangling whenever compile-time check causes a problem here: github.com/ufobat/HTTP-Server-Ogre...ol.pm6#L30 14:24
jnthn AlexDaniel: Sort of :)
AlexDaniel which, if I get it right, is then used here: github.com/ufobat/HTTP-Server-Ogre...6#L92-L106
any thoughts on this? 14:25
jnthn Ah, somebody did rely on it.
I actually thought it already did check that. There were no spectests ever covering such a thing.
huh, the code in there looks like it's largely copied from Cro 14:26
Anyway, if we can't do that check, then we can't have any of the opts originally intended (but only very recently intended) for supply/react blocks 14:27
s:2nd/intended/implemented/ 14:28
AlexDaniel that's the only module that is currently failing because of that 14:30
jnthn OK. Then I think we can consider it as having relied on a bug. 14:32
Was this somehow missed in the last toaster run? 14:33
Clearly the change was merged by then, because I fixed a regression that run showed up 14:34
stmuk 07a-tables.t, 07b-tables.t, 07c-tables.t are absent but present in spectest.data.6.c 14:36
AlexDaniel it was missed, yes. I shrugged it off because I see this in the output right now: gist.github.com/AlexDaniel/08946e5...62cb6d3be7
stmuk absent from the roast errata branch I mean
AlexDaniel and so far I have no idea what that means :)
jnthn It's odd, there is a META6.json in the repo at least 14:40
timotimo maybe a CATCH that's too broad caught a different exception and thought the file was missing
jnthn Ah, could be
Dunno
AlexDaniel I can't get that error locally. At least, didn't get enough hands on it yet to do it 14:41
jnthn Anyway, summary is that many of the opts that went in rely on the "no whenever outside of supply/react" check in order to be safe.
Reverting the check thus implies we'd have to revert a whole bunch of other things 14:42
It may be possible to require a `use v6.d.PREVIEW;`
And only call the optimized paths in those cases, and thus only enforce the check in that case too 14:43
Geth rakudo: 5cbd55281b | (Aleks-Daniel Jakimenko-Aleksejev)++ | t/spectest.data.6.c
07a/b/c-tables.t files are not part of 6.c-errata

If I got it right, this is the file for the errata tests. It was edited by accident I assume.		 14:45
AlexDaniel stmuk: thanks! This ↑ should resolve it I think
jnthn: uhhh… I don't know. I get the point about optimizations, but using whenever like this does indeed look rather innocent to break it like that 14:57
so I'll let that sink while I do something else…
jnthn AlexDaniel: It was always intended, and only ever tested/documented, that whenever had to be textually inside of a react/whenever 14:58
Yes, it's a pity that wasn't enforced prior to now 14:59
I could live with it being put behind 6.d.PREVIEW. 15:00
uh, inside of a react/supply, I meant above 15:02
jnthn afk for a while 15:06
AlexDaniel I'll be away for an hour or two so I created this ticket for others to comment on: github.com/rakudo/rakudo/issues/1428 15:08
if somebody wants to :)
AlexDaniel afk for a bit 15:17
AlexDaniel afk for a byte*
stmuk I'm seeing the missing META6.json warning but can't reproduce 15:44
it's a recent addition to zef
AlexDaniel ah 15:51
9 days ago
so at least it's not a rakudo regression causing this message to pop up
good
stmuk github.com/ugexe/zef/issues/225 15:59
its just a harmless warning I was confused by the log message 17:20
Geth nqp: jstuder-gh++ created pull request #395:
Don't die if QAST parent doesn't have dump_extra_node_info method		 18:11
DrForr hurr, test failed in Cro:Tools::CroFile. 18:36
Might be permissions - I'll just force install and see if something works. 18:38
mst DrForr: or, the CroBar approach 20:21
travis-ci Rakudo build passed. Aleks-Daniel Jakimenko-Aleksejev '07a/b/c-tables.t files are not part of 6.c-errata 20:29
travis-ci.org/rakudo/rakudo/builds/331470683 github.com/rakudo/rakudo/compare/5...bd55281b9c
DrForr mst: Yeah. At the moment I'm bringing my poor VM to its knees building Angular. 20:34
timotimo mst: *snrk* 21:11
jnthn I've long wondered what I'll end up using the CroBar pun for... :P 21:45
lizmat hehe...
ok, so I want to call gethostname() with NativeCall
and I'm too lazy to look up the exact incantation
jnthn I've so far resisted putting "Now that's something to Cro about" in our release announces :P
nativecallable6: int gethostname(char *name, size_t len) 21:46
nativecallable6 jnthn, No! It wasn't me! It was the one-armed man! Backtrace: gist.github.com/33b6540a144645ffdb...5efc3328e2
pmurias samcv: I'll attempt to change eqat* ops myself as this doesn't seem to be too hard
jnthn Come on dude, that's a soft one 21:47
pmurias samcv: the bug occurs without interpolation too, it's just that the tests aren't good enough
jnthn sub getnostname(Str $name, size_t $len --> int32) is native {*} # or something
lizmat jnthn++
jnthn You may have to `use NativeCall :types` to get size_t, I forget 21:48
lizmat looking good so far :-)
hmm... nor working :-( 22:08
anyways, did we ever think about having an API that would allow the functionality that $#foo offers in Perl 5 22:09
a .set_elems(42) if you will ?
jnthn .end does that as an r-value 22:11
lizmat yeah, but I need it as a left value ?
in perl 5 $#foo is a left value
jnthn Don't think there's anything for that. Though, perhaps given we have sized arrays, I've not yet felt a need. 22:15
Buf.allocate exists for doing an initial size from the start
samcv .ask pmurias where does it fail? can you show/tell me? thanks! 22:22
.tell pmurias where does it fail? can you show/tell me? thanks! 22:23
yoleaux: help
geekosaur yoleaux only lived once?
lizmat jnthn: it also as Buf.reallocate 22:33
but really what I think is missing, is something like .set-elems
the fact that reallocate is actually calling nqp::setelems sorta gives that idea :-)
jnthn I'd rather we have consistent names across Buf/Array 22:36
(And so allocate/reallocate on Array also) 22:39
Though it'll be a bit...interesting...to implement in the case that the Array has a $!todo 22:40
lizmat it would force a reify, just like .elems 22:45
and if it's lazy, it will fail ?
Zoffix jnthn: are you still around? What did you say needs to be done on here: github.com/rakudo/rakudo/issues/14...-359277998 23:26
Guess not :) Oh well 23:34
jnthn Zoffix: For a little bit 23:42
And only half paying attention :P
Oh...uh...I think we read lizmat's comment the opposite way :) 23:43
lizmat sorry I wasn't more clear
jnthn I'm now not quite sure how I read it :P 23:44
Probably "in the way that makes the most work for me possible" :P
Zoffix :D
lizmat and I was trying to keep it simple :-)
jnthn Yeah, now I read it again, I think you are advocating "do nothing, make the module fix it"? 23:45
lizmat jnthn: yes 23:48
lizmat calls it a nigh& 23:49
t
jnthn o/ 23:51

Powered by Cro and the Raku® Programming Language.

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

Thank you!