|
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. |
|||
| MasterDuke | vrurg: you should add yourself to the CREDITS | 00:02 | |
| vrurg | MasterDuke: lizmat already said that. I'm waiting until something really interesting would be pulled in. | 00:03 | |
| Thanks anyway! :) | 00:04 | ||
| MasterDuke | i think 4 fixed issues is 3 more than you need to be included | 00:06 | |
| but that's up to you of course | |||
| vrurg | :D too focused on getting things fixed, perhaps. Just every time I get into a bug it turns out there is more bug lurking behind it. I just can't stop chasing them... | 00:14 | |
| MasterDuke | you definitely didn't start with the easiest possible bugs... | 00:16 | |
|
00:17
ggoebel joined
|
|||
| vrurg | MasterDuke: Not really. What would be so difficult in making Subset matching only the where constraint? (It is currently matching itself). But behind it there is a bug with about cloning proto methods (and, probably, not only them). | 00:19 | |
| Without fixing the latter the former can't be fixed too. | |||
| Otherwise there no complications. ;) | 00:20 | ||
|
00:27
lucasb left
00:52
ggoebel left
01:59
leont left
|
|||
| ugexe | There was no global deoptimization triggered by the profiled code. | 01:59 | |
| There was one global deoptimization triggered by the profiled code. | |||
| There were 17 global deoptimization triggered by the profiled code. | |||
| why does that say 3 seemingly different things | |||
| results from a --profile ^ | 02:00 | ||
| e.g. how were there 17 global deoptimizations triggered by the profiled code when it just said there was 0 (and 1) | |||
|
03:20
ufobat_ joined
03:22
ufobat left
03:27
squashable6 left,
squashable6 joined,
ChanServ sets mode: +v squashable6
|
|||
| ugexe | .tell patrickb gist.github.com/ugexe/a88a63154eee...81390e6d5b | 03:54 | |
| yoleaux | ugexe: I'll pass your message to patrickb. | ||
|
04:11
epony left
04:24
epony joined
04:32
kawaii left
04:34
kawaii joined
06:26
robertle left
07:46
epony left
08:00
[Tux] left
08:28
[Tux] joined
09:35
samcv left
09:38
samcv joined
09:43
robertle joined
10:14
leont joined
|
|||
| lizmat | ugexe: maybe it is a junction ? :-) | 10:14 | |
|
11:07
epony joined
|
|||
| AlexDaniel | jnthn: maybe github.com/rakudo/rakudo/issues/15...-475580927 | 11:28 | |
| pretty sure it also affects Cro, and I'm suffering because of leaks too… | |||
|
12:39
AlexDaniel left,
robertle left
12:40
AlexDaniel joined,
lucasb joined
|
|||
| jnthn | AlexDaniel: Maybe next week; really hope to have more time then. | 12:58 | |
|
12:59
patrickb joined
|
|||
| vrurg | jnthn: thank for your advises! I found the fix seemingly. Just didn't have time to test it throughly. | 13:31 | |
|
14:08
robertle joined
14:18
TimToady left
14:27
TimToady joined
14:36
MasterDuke left
15:01
AlexDaniel left,
AlexDaniel joined
16:10
robertle left
|
|||
| vrurg | Could somebody merge PR 2780, please? | 16:12 | |
| jnthn | vrurg: Should let it pass CI first, but I took a look over it and it looks sensible | 16:16 | |
| vrurg | jnthn: Thanks! | 16:17 | |
| I wanna get back to subsets and would prefer to merge the fix from the master. | |||
| lizmat | oops, saw this too late | 16:24 | |
| looks like we lost Geth ? | |||
| moritz timotimo ^^ | 16:25 | ||
| jnthn vrurg builds clean for me... running tests now | |||
| moritz | seems we have lost hack | 16:29 | |
| reboot | 16:30 | ||
|
16:30
p6lert left,
dalek joined,
ChanServ sets mode: +v dalek,
synopsebot joined,
p6lert joined,
Geth joined,
ChanServ sets mode: +v synopsebot,
ChanServ sets mode: +v Geth
|
|||
| lizmat | moritz++ | 16:31 | |
| jnthn vrurg spectests ok for me | 16:34 | ||
| Geth | rakudo: bea5199342 | (Elizabeth Mattijsen)++ | src/core/Parameter.pm6 Added helper method Parameter.cool_constant Not intended to be used outside of the core. Returns the Cool value of a Parameter if it is a Cool constant. Otherwise returns Nil E.g.: sub a("foo") { }; dd &a.signature.params[0].cool_constant; # "foo" sub b($foo ) { }; dd &a.signature.params[0].cool_constant; # Nil |
16:43 | |
| ugexe | curious, why not make it a private method and call that private method through nqp so as note to expose it publically? | 16:45 | |
| im probably misremembering and its private attributes that can be accessed | 16:47 | ||
| Geth | rakudo: dd1c805949 | (Elizabeth Mattijsen)++ | src/core/Main.pm6 Simplify candidate selection in MAIN usage A recent change tried to be smarter on which MAIN candidates to show. This was however too smart for its own good, omitting candidates when it shouldn't. Now limiting of candidates will only happen if the first given parameter matches a **constant** value as the first parameter. So if you have multiple MAIN candidates that have "foo" as the first parameter, and you give "foo" ... (6 more lines) |
16:48 | |
| lizmat | ugexe: I needed the method in Main.pm, not in Parameter.pm | ||
| also: "trusts" doesn't work in the setting | 16:49 | ||
| ugexe | right but i thought one could nqp::foobar('!methodname', $someobj) | ||
| but im pretty sure im thinking of attributes now | 16:50 | ||
| lizmat | nqp::getattr you mean ? | ||
| ugexe | probably | ||
| lizmat | yes, you can, but this is Parameter.pm6 logic. Felt weird to put that in Main | ||
|
16:52
patrickb left
|
|||
| Geth | rakudo: 2b991664d4 | (Elizabeth Mattijsen)++ | 2 files Move "cool_constant" logic to Main Thereby effectively hiding it from the outside world. ugexe++ |
17:10 | |
| vrurg | BTW, is there branch for the upcoming 6e? | 17:56 | |
|
18:36
leont left
19:04
robertle joined
19:21
travis-ci joined
|
|||
| travis-ci | Rakudo build passed. Elizabeth Mattijsen 'Move "cool_constant" logic to Main | 19:21 | |
| travis-ci.org/rakudo/rakudo/builds/510017981 github.com/rakudo/rakudo/compare/d...991664d415 | |||
|
19:21
travis-ci left
19:35
hankache joined
|
|||
| timotimo | anybody want to proofread my blergpost? | 19:58 | |
| wakelift.de/p/2dc5d8df-2b40-4706-8...d90d569b7/ | |||
| robertle | I'd start by calling it a blogpost, 'blerg' has negative connotations for some poeple ;) | 20:20 | |
| timotimo | but it's on my blagoblag | 20:25 | |
| hankache | timotimo++ | 20:32 | |
| lizmat | timotimo: "type name like MixBag" MixBag ?? BagHash perhaps?? /me dunno about MixBag | 20:46 | |
|
20:58
hankache left
|
|||
| lizmat | timotimo: that's my only nit | 20:58 | |
| timotimo++ | |||
| timotimo | ha, you're right, mixhash or baghash not mixbag | 21:05 | |
| i mean there could be a custom type named that of course | |||
| lizmat | could be :-) | 21:08 | |
|
22:09
lucasb left
|
|||
| timotimo | published that blog post: wakelift.de/2019/03/22/intermediat...snapshots/ | 22:41 | |
|
22:53
MasterDuke joined,
MasterDuke left,
MasterDuke joined
|
|||
| MasterDuke | timotimo: "The very first post on this blog already had a little bit about the heap snapshot profiler in the first post on this very blog." has a redundant beginning and ending | 22:59 | |
| timotimo | haha | 23:03 | |
| good catch | 23:04 | ||
| MasterDuke | timotimo: and i think "The individual attributes can be uncompressed and stored away in parallel and since there's one count shared between many attributes." has an extra 'and' between 'parallel' and 'since' | 23:06 | |
| AlexDaniel | m: use v6.e.PREVIEW | 23:07 | |
| camelia | ===SORRY!=== Cannot call method 'typed_panic' on a null object |
||
| AlexDaniel | mmm… | ||
| jnthn | m: use v6.e | 23:08 | |
| camelia | ===SORRY!=== Cannot call method 'typed_panic' on a null object |
||
| jnthn | Well, oops | ||
| I mean, it's better than silently accepting it, but still LTA | |||
| MasterDuke | timotimo: maybe missing some words at the end here? "relocatable perl6 binary in," | 23:10 | |
| timotimo | fixed, or tried at least | 23:12 | |
| MasterDuke | nice work, definitely looking forward to getting heap snapshot support in moarperf | ||
| timotimo | thanks | ||
| tell me about your favourite features? :) | |||
| what you'd like to see, i mean | |||
| MasterDuke | well, don't remember if it does it now, but being able to know what exactly is getting collected in each run | 23:15 | |
| timotimo | *gasp* | 23:17 | |
| impossible | |||
| MasterDuke | hm, maybe not each exact object, but counts of types | 23:18 | |
| timotimo | that's quite a bit easier :) | 23:25 | |
| there's loooooots of space in between snapshots for ... basically anything at all | 23:35 | ||
| it'll quickly turn into a horrifying hodgepodge of everything forever | 23:37 | ||
| Geth | roast: vrurg++ created pull request #521: Test for GH #2772 |
23:43 | |
| rakudo: vrurg++ created pull request #2784: New roast test |
23:47 | ||
| jnthn | Actually if you run both the normal profiler *and* the heap profiler than you'd have the data to calculate what was allocated but then collected, no? :) | 23:59 | |