01:52 sena_kun joined 01:54 Altai-man_ left 02:28 Xliff joined
Xliff Can someone tell me how perl6's version number (for moar-blead) is constructe? 02:28
I know the 2020.06 part comes from the git tag, but where are the build number and git hash computed? 02:29
[Coke] 'git describe' ? 02:36
Xliff 2020.06-78-ga30014e1a 02:38
So why is the reported version still '2020.06-30-g389f0a81'
I'm on my own local branch that I'm trying to keep synce with master, so it's weird. 02:40
Hmmm.... OK, so I see it's the number of commits made since the tag. *sigh* 02:41
But that syncs up with my copy. 02:43
[Coke] Did you re-configure? 02:50
Xliff Yep
[Coke] might need that to get the change into your locally built copy
Xliff Hmmm....
[Coke] dumb question, you sure you're testing the right version?
Xliff I'm getting this when reconfiguring: "Note: switching to '2020.06-14-g5832b2628" 02:56
Which makes no sense. That puts me in detached head.
This is done by Configure.pl
I'm now compiling piece by piece by hand. 03:03
Files in install/bin are stale. 03:11
And not getting moved back when doing "make install" for some reason.
ln -s bin install/bin # Fixes issue. 03:12
[Coke] do you have a prefix specified in your config.status? 03:16
(wondering if you accidentally changed install dirs at some point) - anyway, glad you have something working 03:20
Xliff --prefix=. 03:22
I think that's a default.
It's odd. Last exec files in that dir were dated just 2 days ago. 03:23
03:52 Altai-man_ joined 03:54 sena_kun left 05:10 vrurg left 05:11 vrurg joined 05:16 vrurg left 05:44 vrurg joined, vrurg left, vrurg joined 05:53 sena_kun joined 05:54 Altai-man_ left 07:51 Altai-man_ joined 07:54 sena_kun left 08:23 oddp joined 08:53 sena_kun joined 08:56 Altai-man_ left
lizmat Files=1307, Tests=113021, 219 wallclock secs (29.09 usr 8.45 sys + 3056.77 cusr 288.95 csys = 3383.26 CPU) 09:10
timotimo i believe the default prefix is "./install"? 09:33
if you're using --gen-moar, you'll be getting your versions managed for you, which is the opposite of what you want, so building the pieces sparately manually is what you want to do
09:51 Altai-man_ joined 09:54 sena_kun left 09:56 softmoth left
Geth_ rakudo: niner++ created pull request #3792:
Give a hint about missing provides in a META6.json when module isn't …
jnthn nine++ # I was going to post on the problem-solving issue suggesting just that... :) 10:41
10:41 patrickb joined 11:08 robertle left
lizmat notable6: weekly 11:41
notable6 lizmat, 3 notes: gist.github.com/f9b84f690c160d2f55...304c61ccf9
lizmat notable6: weekly reset
notable6 lizmat, Moved existing notes to “weekly_2020-07-13T11:41:26Z”
11:52 sena_kun joined 11:54 Altai-man_ left
[Tux] Rakudo version 2020.06-72-g85320f466 - MoarVM version 2020.06-22-gaf5b89b64
csv-ip5xs0.824 - 0.824
csv-ip5xs-207.798 - 7.885
csv-parser24.321 - 24.513
csv-test-xs-200.389 - 0.402
test7.702 - 7.947
test-t1.857 - 1.876
test-t --race0.831 - 0.880
test-t-2031.017 - 32.109
test-t-20 --race8.844 - 8.855
12:03 vrurg left 12:04 vrurg joined 12:09 vrurg left 12:12 vrurg joined 13:24 MasterDuke left 13:27 Xliff left, Geth_ left, leont_ left, maggotbrain777 left, finsternis left, nebuchadnezzar left 13:28 Xliff joined, Geth_ joined, leont_ joined, maggotbrain777 joined, finsternis joined, nebuchadnezzar joined 13:41 patrickb25 joined 13:43 patrickb left 13:49 sena_kun left
Geth_ nqp: 9dd7a18fed | (Jonathan Worthington)++ | src/NQP/Actions.nqp
Remove override identical to overridden method
nqp: 10e87758b8 | (Jonathan Worthington)++ | src/QRegex/P6Regex/Actions.nqp
Tweaks to regex compiler API

  * Expose the capture name calculation procedure so that RakuAST can use
  * Remove an unused parameter to the alternation NFAs calculation
nqp: 32be685690 | (Jonathan Worthington)++ | src/QRegex/P6Regex/Actions.nqp
Remove reliance in a $*W object in alt NFA gen

Which does not exist under the Rakudo RakuAST frontend.
rakudo/rakuast: 0848ee7bad | (Jonathan Worthington)++ | src/Raku/ast/code.rakumod
Mark regex decls and quoted regexes as terms
rakudo/rakuast: 6b0249b143 | (Jonathan Worthington)++ | 3 files
Attach alternation NFAs to compiled regexes
rakudo/rakuast: 5ef940ffea | (Jonathan Worthington)++ | t/12-rakuast/regex.t
Some basic tests for building regexes via RakuAST
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2020/07/13/2020-28-bridges-7/ 14:09
14:13 sena_kun joined
vrurg lizmat++ 14:13
lizmat afk for a few hours& 14:45
Geth_ rakudo/rakuast: 210a5ee23b | (Jonathan Worthington)++ | src/Raku/ast/regex.rakumod
Add missing visit-children methods
rakudo/rakuast: d270fb8c04 | (Jonathan Worthington)++ | src/Raku/ast/regex.rakumod
Fix thinkos in backtrack modifier kinds
rakudo/rakuast: 59596c12cf | (Jonathan Worthington)++ | t/12-rakuast/regex.t
Cover a bunch more RakuAST regex construction
rakudo/rakuast: a6c91cca58 | (Jonathan Worthington)++ | 3 files
RakuAST support for regex quantifier separators
15:27 patrickb25 left 15:31 squashable6 left 15:32 patrickb joined 15:33 squashable6 joined 15:36 softmoth joined
Geth_ rakudo/rakuast: 9cab48da3a | (Jonathan Worthington)++ | 3 files
RakuAST for regex groups ([...])
rakudo/rakuast: dfe2d02cbb | (Jonathan Worthington)++ | src/Raku/Actions.nqp
Compile \, in regex into RakuAST::Regex::Literal

Don't think we need to model that it came from a backslash sequence by default (maybe make that available through the source handling, when we add that).
15:48 softmoth_ joined 15:49 softmoth left, sena_kun left
Geth_ rakudo/rakuast: d853ec3cd8 | (Jonathan Worthington)++ | 2 files
Store capture names in RakuAST regex compilation

Also, test cover the various forms of capturing by name via assertions syntax.
16:35 sena_kun joined
jnthn urgh, feeling a bit tired... Positional captures can be finished up tomorrow 16:57
timotimo I wonder how you know whether it is a day for dispatcher work 17:08
17:10 softmoth_ left 17:11 softmoth joined
Geth_ nqp: patrickbkr++ created pull request #645:
Fix retrieving template.html location
17:53 sena_kun left 18:06 sena_kun joined 18:35 sena_kun left 18:50 sena_kun joined
releasable6 Next release in ≈4 days and ≈23 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
19:01 softmoth left 19:04 sena_kun left 19:05 sena_kun joined 19:11 sena_kun left 19:28 MasterDuke joined 19:32 sena_kun joined
Geth_ rakudo: patrickbkr++ created pull request #3793:
Also set the $NQP_HOME HLL symbol in the nqp compiler object
¦ problem-solving: lizmat assigned to rba Issue pl6anet.org is a name that does not reflect Raku github.com/Raku/problem-solving/issues/212 19:55
patrickb timotimo: Some time ago when I implemented relocatability in rakudo you advised me to use set/gethllsymbol to save the raku and nqp home directories. Now I extended this to also have these symbols available in NQP. I'm not sure if this is a good way to do it, as NQP code relies on rakudo to set that symbol. Can you have a look at my two PRs above and give an opinion whether this is OK or the wrong way to do things? 19:56
20:00 ShimmerFairy left
Geth_ roast: c98d1c1114 | (Christian Bartolomäus)++ | 3 files
[JVM] Unfudge some tests
20:13 ShimmerFairy joined 20:30 sena_kun left 20:49 sena_kun joined 21:04 sena_kun left 21:19 sena_kun joined 21:31 patrickz joined
lizmat sleep& 22:01
22:18 patrickb left 22:19 patrickz left
gfldex raku: sub s(&c:(Any)) {}; s(-> $any {}); 22:39
evalable6 (exit code 1) Constraint type check failed in b…
gfldex, Full output: gist.github.com/b2c5e87b1899efcad3...fcc37b8d1f
gfldex raku: sub s(&c:(Any)) {}; s(-> Any $any {});
gfldex raku: sub s(&c:(Any(Mu))) {}; s(-> $any {});
gfldex The default type tends to be Any. Why does it check for Mu here? 22:40
timotimo that is the default constraint for aero functions 22:42
Geth_ rakudo/rakuast: 8854373295 | (Jonathan Worthington)++ | 7 files
RakuAST for the `do` statement prefix

Which is the simplest of the statement prefixes. This entails doing the basics that all statement prefixes need too.
jnthn I've been thinking about what an aero function is for the last 20 minutes and I just realized it's meant to be arrow function :) 23:07
timotimo I'm sorry I'm trying to save my wrists some pain and therefore using dictation 23:09
23:09 sena_kun left
timotimo usually it's easy to correct my mistakes but this IRC clients in the terminal where I can just click on words to correct them 23:09
jnthn Your wrists are certainly more important than my tired brain that's unable to guess stuff. :) 23:12
I guess `do` is the simplest of the statement prefixes. And now it's done.
I think it may produce better code in the statement case than the current compiler (it doesn't bother to thunk the statement, and it looks like we do today) 23:13
timotimo to do or done there is no try yet 23:14
I will react with joy whenever you are done 23:15
jnthn :)
Ah yes, I need to do the exception-y stuff too 23:16
timotimo yes because there is always a catch isn't there
jnthn Yes, but soon everything will be under control...
timotimo well I'm certainly glad you can return from insight control 23:17
jnthn goes quietly to sleep o/
timotimo good night and sleep whatever
I lazily await your return value 23:18
timotimo does tap dance 23:19
23:35 sivoais left 23:36 sivoais joined 23:38 oddp left 23:42 leont_ left