| Geth | rakudo/nom: 34626e3e6d | (Elizabeth Mattijsen)++ | 9 files Remove | from (|) sig in protos where possible If all candidates take no parameters, there is no point in the proto accommodating any parameters. This appears to make each of these methods about 4% faster. |
00:41 | |
| lizmat | good night, #prl6-dev! | ||
| yoleaux2 | 4 Feb 2017 19:54Z <MasterDuke> lizmat: given your recent work with .tail(), are there any circumstances where it isn't faster than [*-1]? if not, should we replace all [*-1] in the settings with .tail()? | ||
| lizmat | MasterDuke: for .tail, I don't think you can get any faster. For .tail(N) things *may* be slower in some cases | 00:42 | |
| MasterDuke | .tell lizmat i only see two places where tail(N) would be applicable anyway (src/core/Block.pm:253 and lib/Pod/To/Text.pm6:55), so i'll leave those for you to decide if it's worth it | 00:51 | |
| yoleaux2 | MasterDuke: I'll pass your message to lizmat. | ||
| Geth | rakudo/nom: 1b99196141 | (Zoffix Znet)++ | src/core/REPL.pm Fix `Mu` line return value causing issues in REPL Fixes RT#130719: rt.perl.org/Ticket/Display.html?id=130719 |
00:55 | |
| synopsebot6 | Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130719 | ||
| Geth | roast: 04fa01361b | (Zoffix Znet)++ | S19-command-line/repl.t Test `Mu` as line return value isn't an issue in REPL RT#130719: rt.perl.org/Ticket/Display.html?id=130719 Rakudo fix: github.com/rakudo/rakudo/commit/1b99196141 |
00:56 | |
| synopsebot6 | Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130719 | ||
| lizmat | . | 06:42 | |
| yoleaux2 | 00:51Z <MasterDuke> lizmat: i only see two places where tail(N) would be applicable anyway (src/core/Block.pm:253 and lib/Pod/To/Text.pm6:55), so i'll leave those for you to decide if it's worth it | ||
| lizmat | .tell MasterDuke the Block case is for .perl, for clarity I would keep that. The Pod/To/Text case seems more like a fit for a .skip ? | 06:51 | |
| yoleaux2 | lizmat: I'll pass your message to MasterDuke. | ||
| Geth | nqp: ef25c99bad | (Samantha McVey)++ | docs/ops.markdown Add documentation on unicmp_s op on MVM |
||
| rakudo/nom: 5af02a1e93 | (Elizabeth Mattijsen)++ | 9 files Revert "Remove | from (|) sig in protos where possible" This reverts commit 34626e3e6d8476305d305fc602ff6494e6e35817. In the end, I wasn't able to recreate the 4% savings that the initial change to .Bool seemed to indicate. Since there was no noticeable change in the time the spectest needs either (which I hope we would see), I'm going to revert this for now. |
08:37 | ||
| nine | lizmat: one good night's sleep and suddenly it becomes this simple: my $ctx := nqp::ctxcaller(nqp::ctx()); $ctx := nqp::ctxcaller($ctx) while $ctx and not nqp::existskey($ctx, '$=finish'); note nqp::atkey($ctx, '$=finish'); | 09:00 | |
| lizmat | :-) | 09:04 | |
| I knew it was something simple :-) | |||
| nine | But....it will not find it when I put the code in Inline::Perl5 because at the time the "use" runs, we haven't even parsed the POD yet :/ | 09:08 | |
| lizmat | mmmm | 09:26 | |
| have you tried putting in a $=finish at the time of the "use" and see if it gets filled ? | 09:27 | ||
| nine | lizmat: how could I? There cannot be any code after a =finish :) | 09:33 | |
| lizmat | yeah, hmmm | 09:45 | |
| RabidGravy | does someone want to take a quick look at github.com/rakudo/rakudo/pull/1012 ? It bit me in the behind yesterday | 09:49 | |
| timotimo | next up is brian duggan: Informal Domain Specific Languages with Perl 6 | 10:46 | |
| live.fosdem.org/watch/k4201 | |||
| Geth | rakudo/nom: 7c39bbf4de | (Jonathan Stowe)++ | src/core/CompUnit/Repository/Installation.pm Sort the binaries by version Assuming we always want to have the latest version of a binary unless it has been otherwise specified then this seems to be the simplest way to achieve that. |
11:10 | |
| nine | RabidGravy: I took the liberty to sqash those commits and push it | ||
| Geth | rakudo/nom: 6f6f0cf19e | (Samantha McVey)++ | 3 files Add unicmp op which uses unicmp_s nqp op to do string compare |
11:19 | |
| rakudo/nom: 5870ef97d1 | (Samantha McVey)++ | 2 files Hide the unicmp operator behind experimental |
|||
| RabidGravy | nine :) | 11:21 | |
| nine, while I was in the shower I was thinking that maybe the bulk of the "find the right binary" code could be moved from the wrapper into CompUnit::Repository | 11:25 | ||
| thus making the whole retroactive fixing thing easier in the future | 11:26 | ||
| on a somewhat related note, is there a way of making an installation use a specific (possibly non-core,) CUR implementation however it might be installed? | 11:28 | ||
| I'm thinking about this because, for example, for a particular plugin schema I might want to store some additional metadata in the distribution object | 11:31 | ||
| for the loading part it is easy to make something that consumes that additional data as required | 11:32 | ||
| timotimo | hello, fosdem! \o | 12:01 | |
| tadzik | \o/ | 12:05 | |
| Geth | nqp: 439b51e89a | (Samantha McVey)++ | tools/build/MOAR_REVISION Bump MoarVM to get improvements to unicmp_s |
13:03 | |
| Ā¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...-g20dfa6bb | |||
| rakudo/nom: eb3356c6e0 | (Samantha McVey)++ | tools/build/NQP_REVISION Bump MoarVM/NQP to get improvements to unicmp_s |
13:04 | ||
| rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....-g439b51e8 ecb7edede6 | (Samantha McVey)++ | README.md |
|||
| travis-ci | NQP build failed. Samantha McVey 'Bump MoarVM to get improvements to unicmp_s' | ||
| travis-ci.org/perl6/nqp/builds/198548899 github.com/perl6/nqp/compare/ef25c...9b51e89a70 | |||
| lizmat | gist.github.com/lizmat/a98b5115f62...b12c9469d4 # Ovid made me do this | 13:11 | |
| conceptual "is write-once" trait for Attributes ^^^ | 13:12 | ||
| travis-ci | Rakudo build failed. Samantha McVey 'Bump MoarVM/NQP to get improvements to unicmp_s' | 13:31 | |
| travis-ci.org/rakudo/rakudo/builds/198549075 github.com/rakudo/rakudo/compare/5...3356c6e092 | |||
| buggable | [travis build above] ā Did not recognize some failures. Check results manually. | ||
| lizmat | looks like makefile issues ? | 13:34 | |
| samcv | nah it should be fixed now, it didn't get pushed to MVM | 14:12 | |
| yoleaux2 | 4 Feb 2017 22:04Z <raiph> samcv: Have you read github.com/apple/swift/blob/master...ifesto.md? | ||
| samcv | yoleaux2, and yoleaux being different bots strikes again | 14:13 | |
| jdv79 | lizmat: do you only write in nqp now?: i did something similar with jnthn's help a while ago: github.com/jdv/p6-attribute-setonce in real p6 | 14:22 | |
| Geth | rakudo/nom: 411782e959 | (Samantha McVey)++ | src/core/Cool.pm Make uniprop/unival/unimatch 2.1-0.1x as fast Use nqp::unipropcode instead of Rakudo::Internals.PROPCODE I may eventually remove Rakudo::Internals.PROPCODE entirely since because it is much slower, there is not much reason for it to continue to exist. uniprop with default arguments is 2.1x, as is uniprop-bool and uniprop-str is about 1.9x as fast. uniprop with arguments is about 0.1-0.15x faster |
16:02 | |
| rakudo/nom: 25fd1ca3e3 | (Samantha McVey)++ | src/core/IO/Path.pm Allow you to call .s on a directory for IO::Path On Linux at least, directories can have a size by themself. Don't throw when trying to check the size of a directory. The user should have access to whatever the OS says the directory size says. |
16:28 | ||
| rakudo/nom: c0a907fb5a | (Samantha McVey)++ | src/core/Rakudo/Internals.pm remove Rakudo::Internals.PROPCODE and Rakudo::Internals.PVALCODE These are no longer used since they were 5x slower than the nqp ops |
16:33 | ||
| roast: 3133826e99 | (Samantha McVey)++ | S32-io/file-tests.t Add a test to make sure checking the size of a dir doesn't throw |
16:44 | ||
| roast: d91275f886 | (Samantha McVey)++ | S05-mass/properties-general.t Add a test to check that <alpha> matches alphabetics that are not Letters Currently MVM only checks that things have general category 'L', but there are many other types which have Unicode property Letter. The JVM passes this test. |
16:45 | ||
| pmurias | samcv: is your grant proposal up for community review? | 17:58 | |
| timotimo | i think it barely missed the deadline? | 18:15 | |
| pmurias | timotimo: so it would end up moved to the next round? | 18:39 | |
| MasterDuke | pmurias: i think someone said it would happen in march | 18:51 | |
| yoleaux2 | 06:51Z <lizmat> MasterDuke: the Block case is for .perl, for clarity I would keep that. The Pod/To/Text case seems more like a fit for a .skip ? | ||
| Geth | roast: 93e8dc533b | (Daniel Green)++ | S05-mass/properties-general.t Increment missed plan |
20:11 | |
| japhb | .ask jnthn Did you get my response to RT #130716? It isn't appearing for me on the RT site, but I don't know if my response got somehow routed to you and not included in the bug log, or if it just got lost entirely. | 20:24 | |
| synopsebot6 | Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130716 | ||
| yoleaux2 | japhb: I'll pass your message to jnthn. | ||
| Geth | roast: 874acb12cf | smls++ | S32-list/sort.t Add test for RT #126859 |
21:08 | |
| synopsebot6 | Link: rt.perl.org/rt3//Public/Bug/Displa...?id=126859 | ||
| Geth | nqp: MasterDuke17++ created pull request #345: Intify things that cause the most mvm coerce smart numify |
23:06 |