Geth roast/grisu: 8a8ecf89b1 | (Zoffix Znet)++ | S02-names-vars/perl.t
Remove trailing whitespace
 01:11
rakudo/no-unneeded-java: 6419a7e53a | (Aleks-Daniel Jakimenko-Aleksejev)++ | .travis.yml
Don't install jdk when building for moar

When building with backends=moar there is no need to install oracle-java8-installer, and installing it regardless may result in 404 failures because the 3rd party source is sometimes unavailable.
Resolves #1646.		 01:18
roast/grisu: 39d4ac68ef | (Zoffix Znet)++ | S03-metaops/hyper.t
Fix incorrect nodality tests for reduce/produce

Closes github.com/rakudo/rakudo/issues/1633
The tests' description say they're testing that reduce/produce
  are nodal, and it makes sense for them to be nodal, however, the
  $expected specified in the tests indicates the non-nodal behaviour.
Likely, the $expected was simply copy-pasted from the output of running the $got code and plugged into the test, without noticing that the result is incorrect.		 01:21
roast/grisu: 07b82763db | (Zoffix Znet)++ | S03-metaops/hyper.t
Merge branch 'master' into grisu
rakudo: AlexDaniel++ created pull request #1650:
Don't install jdk when building for moar		 01:22
AlexDaniel ah, I see 01:26
Geth rakudo/no-unneeded-java: de53040e71 | (Aleks-Daniel Jakimenko-Aleksejev)++ | .travis.yml
Don't install jdk when building for moar

When building with backends=moar there is no need to install oracle-java8-installer, and installing it regardless may result in 404 failures because the 3rd party source is sometimes unavailable.
Resolves #1646.		 01:28
rakudo/no-unneeded-java: 958613eb41 | (Aleks-Daniel Jakimenko-Aleksejev)++ | .travis.yml
Retrigger (cuz no other way)

This commit won't be merged.		 01:34
AlexDaniel gah! OK now it's rolling 01:35
Geth roast/grisu: 6a2e0399e8 | (Zoffix Znet)++ | S32-num/rounders.t
Use correct test routines for Rat.round tests / Unfudge JVM

1) Compare the numbers as tests actually claim to compare:
   using numeric semantics, not string semantics which `is` uses
2) For .round(1e-5) test, use approximately equals operator:
   it was fudged on JVM and passing on MoarVM for the wrong reason.
   The result of rounding is a Num (value 123.45679000000001e0),
... (9 more lines)		 02:53
roast/6.c-errata: f6e070bff2 | (Zoffix Znet)++ | S32-num/rounders.t
Use correct test routines for Rat.round tests / Unfudge JVM

1) Compare the numbers as tests actually claim to compare:
   using numeric semantics, not string semantics which `is` uses
2) For .round(1e-5) test, use approximately equals operator:
   it was fudged on JVM and passing on MoarVM for the wrong reason.
   The result of rounding is a Num (value 123.45679000000001e0),
... (9 more lines)		 02:55
roast/grisu: a4a3c76205 | (Zoffix Znet)++ | integration/advent2013-day15.t
Remove trailing whitespace
 03:05
roast/grisu: e1d14f4006 | (Zoffix Znet)++ | integration/advent2013-day15.t
Fix incorrect Num sequence test / Unfudge JVM

The way JVM stringifies is accurate and the reason more stringifies the way the author of the test tested it is due to loss of precision bug.
While MoarVM's current stringification is used by `python2` and `perl`, I don't believe it's correct for our purposes: ... (19 more lines)		 03:18
roast/6.c-errata: f0d6e49049 | (Zoffix Znet)++ | integration/advent2013-day15.t
Fix incorrect Num sequence test / Unfudge JVM

The way JVM stringifies is accurate and the reason more stringifies the way the author of the test tested it is due to loss of precision bug.
While MoarVM's current stringification is used by `python2` and `perl`, I don't believe it's correct for our purposes: ... (19 more lines)		 03:20
roast/grisu: 4546862703 | (Zoffix Znet)++ | S32-exceptions/misc.t
Loosen Num-in-sink warning text message match test

Mostly to make master roast pass with Grisu3 Num stringification on MoarVM. After the change, the value comes out as
  `6.0221408999999994e+23`. My C-only impl of the Grisu3 algo doesn't
have this issue, and I suspect the cause is the Num parsing bug and/or nqp::div_In issue with denormals is to blame for this. ... (6 more lines)		 03:28
roast/grisu: dba0ce1656 | (Zoffix Znet)++ | S02-names-vars/perl.t
Temp-fudge Num roundtrip test for Moar

Fixing Num stringification bugs on MoarVM made this test more sensitive to the Num parsing bugs we had.
I'm fairly confident I can fix the parsing bugs before the release, so I'm fudging this test on `master` so stringification bug fixes could land in to rakudo/MoarVM master.		 03:41
nqp: a116cfb6df | (Zoffix Znet)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 5 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...7-gb8d7082 b8d7082 Merge pull request #823 from MoarVM/grisu af2eb8a More post-Grisu3 Num renderer polishing 49b7389 Generalize mp_get_double() 8841c42 Mod post-Grisu3 Num renderer… 067c059 Stringify Num using Grisu3 algo		 04:14
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...7-gb8d7082
rakudo: a2a2a745c4 | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] a116cfb [MoarVM Bump] Brings 5 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....8-ga116cfb
MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...7-gb8d7082 b8d7082 Merge pull request #823 from MoarVM/grisu af2eb8a More post-Grisu3 Num renderer polishing 49b7389 Generalize mp_get_double() 8841c42 Mod post-Grisu3 Num renderer… 067c059 Stringify Num using Grisu3 algo
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....8-ga116cfb
roast: zoffixznet++ created pull request #406:
Fixes/Fudges for Grisu3 Num stringification
roast/master: 8 commits pushed by (Zoffix Znet)++
Kaiepi what would i need to do to implement support for wchar_t as a type for nativecall? 04:31
never mind, i think i worked out what i need to do 04:41
Geth nqp: 8b5d8e3d10 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 2 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...9-ga5ed7ea a5ed7ea Tweak naming of double mantissa size define 4d3fc28 Fix handling of denormals in nqp::div_In		 07:14
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...9-ga5ed7ea
rakudo: 63617a8d52 | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] 8b5d8e3 [MoarVM Bump] Brings 2 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....9-g8b5d8e3
MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...9-ga5ed7ea a5ed7ea Tweak naming of double mantissa size define 4d3fc28 Fix handling of denormals in nqp::div_In
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....9-g8b5d8e3
samcv can anyone tell my what i'm doing wrong with nqp::iterator and hash's? 07:40
nqp: my %hash := nqp::hash('what', 'here'); my $iter := nqp::iterator(%hash); say(nqp::iterkey_s($iter));
camelia You have not advanced to the first item of the hash iterator, or have gone past the end
at <tmp>:1 (<ephemeral file>:<mainline>)
from gen/moar/stage2/NQPHLL.nqp:1585 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:eval)
from gen/moar/stag…
samcv nqp: my %hash := nqp::hash('what', 'here'); for %hash { say(nqp::iterkey_s($_)) } 07:41
camelia what
samcv this seems to work though
ah it seems I need to use nqp::shift. I should add that to the nqp documentation since that isn't that clear 07:45
Geth nqp: e0be6c62af | (Samantha McVey)++ | docs/ops.markdown
[doc] Make it clear to use nqp::shift for hash iterator's
 07:54
nqp: 1684794ebc | (Samantha McVey)++ | docs/ops.markdown
[doc] Use perl6 fenced code blocks instead of just 'perl'

Should highlight better on github or compatible renderers.
rakudo: 8422d7b4e2 | pmurias++ (committed by Zoffix Znet) | src/Perl6/Actions.nqp
Parse scientific notation numbers using div_In
 09:24
rakudo: 17446fc131 | (Zoffix Znet)++ | src/core/allomorphs.pm6
Fix parsing of denormal Nums in &val

We can't just raise 10 into a power to a magnitude of a Num, as denormals' magnitude would not fit into a normal double.
Fix by using an Int instead and then producing the Num with nqp::div_In op. This aligns the parsing in &val with the parsing done in the grammar for literal numbers.
roast: ec20be04b6 | (Zoffix Znet)++ | S32-num/stringify.t
Remove trailing whitespace
 09:35
travis-ci Rakudo build failed. Zoffix Znet 'Fix parsing of denormal Nums in &val 10:05
travis-ci.org/rakudo/rakudo/builds/358003954 github.com/rakudo/rakudo/compare/6...446fc1318f
buggable [travis build above] ☠ All failures are due to: failed make test (1 failure). Across all jobs, only t/04-nativecall/09-nativecast.t test file failed.
[Tux] Rakudo version 2018.03-70-g17446fc13 - MoarVM version 2018.03-24-gd322741ce
csv-ip5xs0.818 - 0.888
csv-ip5xs-207.775 - 8.855
csv-parser39.683 - 40.060
csv-test-xs-200.440 - 0.486
test9.677 - 9.783
test-t2.498 - 2.799
test-t --race1.056 - 1.465
test-t-2044.439 - 48.628
test-t-20 --race15.591 - 19.062
 11:50
Geth nqp: be7fd36188 | (Stefan Seifert)++ | 4 files
Make NQP's source digest available through the compiler config

This gives NQP's user precise information about the NQP version used and makes it possible to detect any change in NQP's sources.		 11:58
rakudo: d91e930edf | (Stefan Seifert)++ | 4 files
Force a rebuild of compiler and modules on changes to NQP

We use the NQP compiler's new source-digest to detect any changes to NQP. Changes will trigger a rebuild of rakudo and also affect the Perl 6 compiler's ID. Thus precomp files will be outdated accordingly and can be recompiled for the new NQP dependencies.
With this the need for a manual "make clean" after NQP updates and "Missing or wrong version of dependency" errors in general should be greatly reduced.		 12:11
MasterDuke nine++ 12:13
dogbert11 make: *** No rule to make target 'tools/build/check-nqp-version.pl', needed by 'check_nqp_version'. Stop. 12:21
where do I get this tool? 12:22
timotimo cool 12:24
dogbert11: did you Configure.pl before this?
oh, of course 12:25
dogbert11 yes I did: perl Configure.pl --gen-moar --gen-nqp --backends=moar 12:26
MasterDuke nine: is there a tools/build/check-nqp-version.pl that you forgot to check in? 12:27
timotimo that must be it, yeah
travis-ci Rakudo build failed. Stefan Seifert 'Force a rebuild of compiler and modules on changes to NQP 12:29
travis-ci.org/rakudo/rakudo/builds/358035133 github.com/rakudo/rakudo/compare/1...1e930edf4c
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 12:30
Geth rakudo: 5298b7aaed | (Samantha McVey)++ | 2 files
Make sure .encode and .decode accept the same encoding aliases

Internals.pm6 (used for .decode method) was missing utf-8-c8 alias which was worked with .encode but not .decode.
Registry.pm6 was missing windows1252 and windows1251 although they worked with .decode they did not work with .encode ... (6 more lines)		 12:42
samcv night all o/ 12:44
FROGGS night samcv 12:51
Geth rakudo: 265180550c | (Stefan Seifert)++ | tools/build/check-nqp-version.pl
Add missing tools/build/check-nqp-version.pl

Thanks to MasterDuke++ for the reminder		 12:52
nine Always the same, whenever I create a new file...
timotimo :) 12:53
travis-ci Rakudo build failed. Samantha McVey 'Make sure .encode and .decode accept the same encoding aliases 13:00
travis-ci.org/rakudo/rakudo/builds/358041911 github.com/rakudo/rakudo/compare/d...98b7aaedc8
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
travis-ci Rakudo build failed. Stefan Seifert 'Add missing tools/build/check-nqp-version.pl 13:13
travis-ci.org/rakudo/rakudo/builds/358044132 github.com/rakudo/rakudo/compare/5...5180550c98
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 13:14
MasterDuke nine: make: *** No rule to make target 'gen/nqp-version', needed by 'blib/Perl6/ModuleLoader.moarvm'. Stop.
nine MasterDuke: did it run tools/build/check-nqp-version.pl? 13:15
MasterDuke yeah
nine Oooh...probably needs an NQP bump 13:16
I explicitly tested it without but of course did not remove the gen/nqp-version file that was already there
MasterDuke i get the same thing locally at HEAD/HEAD/HEAD
nine MasterDuke: did you run NQP's Configure.pl, too? 13:17
MasterDuke yeah
same error after a `make clean` in rakudo
Geth rakudo: 2365b558af | (Stefan Seifert)++ | tools/build/NQP_REVISION
Bump NQP version for source-digest compiler config
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....gbe7fd3618
MasterDuke17++ created pull request #1652: Rename tools/install-dist.pl to install-dist.p6
nine MasterDuke: what does nqp --show-config | grep source-digest give you? 13:18
MasterDuke nqp::source-digest=c4431924123b8e16a26e3acc95be5977fba2666b
nine And there's no gen/nqp-version file even after check-nqp-version.pl ran?
MasterDuke no gen/nqp-version 13:19
nine: i think i know what it is 13:21
check-nqp-version.pl does `my @config = qx(nqp --show-config);` 13:22
which calls my system installed nqp
which is 2017.06
it should call the one in whatever --prefix i set with Configure.pl 13:23
travis-ci Rakudo build failed. Stefan Seifert 'Bump NQP version for source-digest compiler config' 13:35
travis-ci.org/rakudo/rakudo/builds/358049674 github.com/rakudo/rakudo/compare/2...65b558af52
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
Geth rakudo: eb71edccfd | (Stefan Seifert)++ | 4 files
Use correct nqp binary instead of relying on $PATH

Fixes build when using a different nqp than the one in $PATH		 13:50
nine MasterDuke: should work now 13:52
MasterDuke nine++ working now (had to do a make clean first though) 14:19
travis-ci Rakudo build passed. Stefan Seifert 'Use correct nqp binary instead of relying on $PATH 14:30
travis-ci.org/rakudo/rakudo/builds/358056625 github.com/rakudo/rakudo/compare/2...71edccfd94
Geth nqp/master: 9 commits pushed by pmurias++ 15:06
nqp: 4580760b56 | pmurias++ | src/vm/js/nqp-runtime/reprs.js
[js] Make nqp::rebless check that the new type is compatible
 15:07
nqp: a62cef7dfa | pmurias++ | t/nqp/093-oo-ops.t
Add some test for forbidden reblessing
rakudo/iterator_mixin_re-use: 9f8231c62d | (Timo Paulssen)++ | src/core/Rakudo/Iterator.pm6
re-use parameterizations for *OnlyDelegate

until now, every call to delegate-iterator-opt-methods would create a completely new mixin, as it was parameterized on a value rather than a type.
Using an attribute for this instead means we always hit the cache if the type we're mixing into matches.		 19:21
timotimo passes spectests, anyone want to test with user programs? 19:45
AlexDaniel timotimo: IMO just merge it, more time on master means higher probability that we'll notice issues before the release 20:11
timotimo OK 20:12
Geth rakudo: 9f8231c62d | (Timo Paulssen)++ | src/core/Rakudo/Iterator.pm6
re-use parameterizations for *OnlyDelegate

until now, every call to delegate-iterator-opt-methods would create a completely new mixin, as it was parameterized on a value rather than a type.
Using an attribute for this instead means we always hit the cache if the type we're mixing into matches.
timotimo i'll also look if this same thing can be done in other places as well 20:19
nothing at first quick glance 20:20
Geth nqp: 27b252516f | pmurias++ | src/vm/js/nqp-runtime/sixmodel.js
[js] Fix nqp::istype
 20:22
timotimo AlexDaniel: do you think it's a good idea to put "ed" in the list of editors? after all, it's *the* editor! 20:27
AlexDaniel timotimo: probably not :) 20:28
timotimo: but feel free to add tools and other stuff
like, does github provide .gitignore for perl6 scripts?
can ack filter perl6 files? 20:29
etc.
timotimo: maybe this is the right upstream ticket for notepad++? sourceforge.net/p/scintilla/featur...ests/1207/ 20:33
timotimo oh, perhaps it is 20:34
AlexDaniel “Notepad++ on Windows and Geany on Linux are two widely used lightweight editors that use Scintilla for syntax highlighting.” 20:37
so I guess the same ticket applies to Geany also? 20:38
timotimo could be! 20:41
AlexDaniel “gedit includes syntax highlighting via GtkSourceView” 20:42
and pluma is based on gedit 2 I think
wiki.gnome.org/Projects/GtkSourceView 20:43
OK I wrote that down 20:51
for anyone wondering, we're talking about github.com/perl6/marketing/issues/10
Geth rakudo: 72dfddfa21 | (Timo Paulssen)++ | src/core/Rakudo/Iterator.pm6
fix accidental private methods
 21:18
lizmat nine: thanks for fixing the build issue, I can non build again 21:24
I nuked by install directory, but that didn't fix it :-
and now I'm trying to install Inline::Perl5 from scratch with zef
and now it appears to hang in t/modify_array.t for 5 minutes already :- 21:25
(
Zoffix nine: I can't build: make: *** No rule to make target 'gen/nqp-version', needed by 'blib/Perl6/ModuleLoader.moarvm'. Stop. 22:41
nine: that's with fresh checkouts of master/master/master
MasterDuke Zoffix: i would have expected a fresh checkout to be ok, but i had to run a `make clean` first in the rakudo dir 22:44
Zoffix MasterDuke: that just bypasses the step, doesn't it? 22:45
Just nuked everything again and tried to build again and still fails: gist.github.com/zoffixznet/d2aef2e...6d06a5d287 22:49
Yes, `make clean` after a failed build gets it going, but it breaks again if I do a full build, like when doing bumps.
scratch the last part. Looks like bumps are now working, so only the fresh build is busted 22:52
:/ 22:55
and only if you build the stuff separately it seems. Just running this seems to not have the issue: cd $(mktemp -d) && git clone github.com/rakudo/rakudo/ . && perl Configure.pl --gen-moar --gen-nqp --backends=moar && make && make test && make install
MasterDuke huh, weird 22:56
Geth nqp: ce0263e3f4 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 7 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...6-gb1f64db b1f64db Add missing include for Grisu3 dtoa function 67e5093 only suspend on actually must-suspend breakpoints d322741 Merge pull request #820 from samcv/shiftjis edfde24 Use windows-932 instead of shiftjis for enc name 62cfb2f Make sure to throw in shiftjis decode if byte after lead d44d7d1 Ensure a failed JIS result is distinct from 0 e8fdb35 Add shiftjis decode and decodestream support
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...6-gb1f64db
rakudo: 97677a792c | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 13 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....5-gce0263e ce0263e [MoarVM Bump] Brings 7 commits 27b2525 [js] Fix nqp::istype a62cef7 Add some test for forbidden reblessing 4580760 [js] Make nqp::rebless check that the new type is compatible ... (18 more lines)
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....5-gce0263e
Zoffix ZOFVM: Files=1277, Tests=153160, 148 wallclock secs (20.39 usr 3.40 sys + 3130.95 cusr 166.26 csys = 3321.00 CPU)
timotimo: BTW, here's another place where we do the parametarize-on-args thing: github.com/rakudo/rakudo/blob/72df...#L788-L793 22:58
timotimo that looks like we can have that cached, though 22:59
Zoffix Well, now I'm stumped. I essentially copied the commands Z-Script runs and ran them manually in a temp dir and didn't have any install issues :S gist.github.com/zoffixznet/6a869b2...1887889ce9 23:07
Zoffix hopes the bump fixed something and tries zscript again 23:08
nope 23:09
Ohhh 23:12
it's 'cause zscript stops as soon as it gets a failed exit code. I still get "make: *** No rule to make target 'gen/nqp-version', needed by 'blib/Perl6/ModuleLoader.moarvm'. Stop." but I ignore the exit code 23:13
*I still get that message in manual runs I mean
MasterDuke i assume you're doing a make clean or something like that? cause i don't see that anymore just doing a git pull and then Configure + make install 23:14
dogbert11 Zoffix: just a heads up, your latest work might have fixed RT's 128912-4
Zoffix MasterDuke: it's from fresh checkouts of everything. Nothing to clean 23:16
Even if I nuke rakudo dir only but leave install dir untouched, I no longer get the error
MasterDuke ah
timotimo what about ~/.perl6 and friends? 23:17
i wonder if they are involved at all?
strace -e open could surely tell you
Zoffix dogbert11: thanks. yeah, there should be a bunch. I'm yet to go through them and write tests and close 23:19
"Segmentation fault" on build nice
oh "fatal: sha1 file '/tmp/tmp.8B5iw2f661/nqp/.git/objects/pack/tmp_idx_uUckES' write error: No space left on device" :) 23:20
timotimo oops, ouch :) :)
dogbert11 Zoffix: RT #128817 might be another 23:22
synopsebot RT#128817 [open]: rt.perl.org/Ticket/Display.html?id=128817 [BUG] Num.perl doesn't round-trip numeric value
Zoffix yeah
dogbert11 RT #126826 23:23
synopsebot RT#126826 [new]: rt.perl.org/Ticket/Display.html?id=126826 [BUG] Numeric precision not arbitrary
dogbert11 perhaps
Zoffix No, that'll be addressed by CaR Grant, if it's approved 23:24
well, maybe the output of .perl changed, but that code is producing a Rat, which shouldn't have denominators over 64 bit and the user is assuming it should have arbitrary precision 23:25
uhh 23:27
c: 2018.03 say 0.99999999999999999999999 < 1
committable6 Zoffix, ¦2018.03: «True␤»
Zoffix c: 2017.03 say 0.99999999999999999999999 < 1 23:30
lag
committable6 Zoffix, ¦2017.03: «False␤»
Zoffix right, the bug itself was fixed by rt.perl.org/Ticket/Display.html?id=126826 23:31
.tell nine any build jobs that build MoarVM/nqp/rakudo separately from scratch and check exit code fail now after this message " No rule to make target 'gen/nqp-version', needed by 'blib/Perl6/ModuleLoader.moarvm'". You can repro with this set of commands: gist.github.com/zoffixznet/03fac7a...3525db42a8 23:33
yoleaux Zoffix: I'll pass your message to nine.

Powered by Cro and the Raku® Programming Language.

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

Thank you!