|
weekly Rakudo status meetings with phase transitions: Tue 19:00 UTC | IR clogs at irclog.perlgeek.de/phasers/today Set by moderator on 26 April 2011. |
|||
|
10:04
[Coke] left
10:06
[Coke] joined
14:07
pmichaud joined
|
|||
| pmichaud | My #phasers pre-report: | 14:38 | |
| What I did this past week: | 14:39 | ||
| - more benchmarks of Rakudo and Parrot | |||
| - built a simple sub-level profiler for Rakudo | |||
| - patched rakudo to get ~15% speed improvements | |||
| - helped restore NCI functionality for zavolaj and others | |||
| - created NCI::Utils::ncifunc as a replacement wrapper for dlfunc | |||
| - drafted a "how Parrot is doing" message for PDS | |||
| - attended PDS | |||
| - drafted a new relationship manager policy for Parrot/Rakudo | |||
| Rakudo-specific improvements: | |||
| - added infix:<gcd> and infix:<lcm> operators | |||
| - discussed changes to .perl method | |||
| - added "config.status" to Configure.pl | |||
| - fixed NaN relational comparisons | |||
| - fixed bugs in --makefile-timing flag | |||
| - refactored Mu.item to be much faster | |||
| What I plan to do this next week: | |||
| - more improvements to rakbench and profiling tools | |||
| - more Rakudo performance fixes | |||
| - create an interim Star release (likely but not definite yet) | 14:40 | ||
| - reviews of reports, nom, nqp | |||
| EOR | |||
| We're still looking for comments and suggestions for pmichaud.com/sandbox/relman-draft.txt | 14:42 | ||
| tadzik | pmichaud++, on relman-draft too | 14:48 | |
| moritz | pre-report: | ||
| DID: | |||
| + some small rakudo/nom Configure fixes, identified a problem wrt case insensitive paths on macos | |||
| + made the setting tests in nqp pass | 14:49 | ||
| + general house cleaning in rakudo and nqp | |||
| + watched tadzik++ work on gsoc stuff. I can't claim to have to done any real "mentoring" yet :-) | 14:50 | ||
| + poked a few niecza things | |||
| - did not get around to the planned BackTrace speccing | |||
| PLANS FOR NEXT WEEK: | 14:51 | ||
| + actually do speck BackTrace | |||
| s/ck/c/ | |||
| + blog about the current state of X:: spec | |||
| EOR | |||
|
15:01
spinclad left
|
|||
| tadzik | moritz++ # don't be so modest about this mentoring thingy :) | 15:02 | |
| moritz | tadzik: if my mentoring efforts so far are enough, it's going to be an easy summer for me :-) | 15:48 | |
| tadzik | isn't it supposed to be easy for you and hard for me? :) | ||
| moritz | :-) | ||
| tadzik | plus I haven't started the work yet. Officially, I mean :) | ||
|
15:57
[particle] left
16:03
[particle]1 joined
16:16
[particle]1 left
16:27
[particle] joined
16:42
[particle] left
17:07
[particle]1 joined
17:08
[particle]1 is now known as [particle]
17:56
mberends joined
18:13
colomon joined
|
|||
| Util | Pre-report: slow progress on Darwin binary. Tuits minimal. EOR | 18:45 | |
| mberends | o/ | 18:59 | |
| tadzik | hello | 19:00 | |
| tadzik needs git-global-log | |||
| moritz | \\o | ||
| anybody got a report? | |||
| tadzik | yes | ||
| moritz | pmichaud and I pre-pasted | ||
| tadzik: then go ahead | |||
| tadzik | so, assisted YAPC::Russia hackathoners in Panda hacking, they submitted some nice patches | 19:01 | |
| started more concrete GSoC work, got some tests to pass with a basic parser, still more like week 0 than week 1 though | 19:02 | ||
| raised some RT tickets from the dead, get some of them to the should-be-closed state, assisted by moritz++ as usual | 19:03 | ||
| planes: GSoC week #1 before the week #1 | |||
| that'd be it | 19:04 | ||
| moritz | tadzik++ | ||
| any other reports? | |||
| tadzik | what happened to our ambassadors in Russia? | 19:05 | |
| moritz | any other reports? anything we need to discuss? | 19:10 | |
| colomon | quick report: I reverted my Rakudo a couple of months to get zavolaj working again, then added negative numbers to Math::BigInt. Also kibitzed a good bit on #perl6 about gcd, profiling, .Numeric, etc. Hope to do some more with Math::BigInt and profiling in the next week. .EOR | 19:11 | |
| mberends | ah, zavolaj. has anyone tried the new patch? it didn't work for me :-( | 19:12 | |
| tadzik | to discuss: goals for the book? | ||
| colomon | log overnight indicated pmichaud++ has been working with the Parrot guys trying to get things resolved wrt zavolaj. | 19:13 | |
| moritz | goals for the book: get it to a stage where it can be printed | 19:14 | |
| Util | FYI, Parrot PDS is this Saturday | 19:15 | |
| Is there a hit-list of book issues anywhere? | 19:16 | ||
| tadzik | I think I created a list of ideas | ||
| github.com/perl6/book/blob/master/book-ideas | |||
| Util | Thanks; I have not looked at the book in too long. | 19:17 | |
|
19:18
mikehh joined
|
|||
| mberends | we don't seem to have enough people here now to discuss the parrot relationship managers, but we should remember to think of suggestions. | 19:20 | |
| colomon | one thing (not book) -- I don't know if it's been discussed much from this perspective, but letting Rakudo out with a broken zavolaj was arguably a pretty big fail on our part. what should we be doing differently to prevent that in the future? | ||
| tadzik | Zavolaj needs better tests | ||
| colomon | true. | 19:21 | |
| tadzik | I think the tests are ran before shipping the modules with the Star release | ||
| colomon | Were any Zavolaj tests run before the regular Rakudo release? | ||
| Util | colomon: Define some minimal test suite for the release managers to (have someone) run. | ||
| tadzik | the problem with zavolaj and tests is that they won't probably be too OS-agnostic, are t | 19:22 | |
| well, we can "link" to C stdlib | 19:23 | ||
| I can write some tests when zavolaj gets fixed, and then poke jnthn or someone to test them on windows | |||
| colomon | even that might not be OS-agnostic from zavolaj's point of view. | ||
| tadzik | I think stdlib is "is native('')" in zavolaj | 19:24 | |
| colomon | and quoting pmichaud, let's not let the perfect be the enemy of the good. Tests on ANY platform would have shown the zavolaj bug. | ||
| tadzik | true | ||
| colomon | seems like maybe in an ideal world, we'd test *all* the modules the day before release, with the new Parrot? | 19:27 | |
| they shouldn't be showstoppers by themselves, but they might well turn up showstoppers. | |||
| send out automated "your module has regressed" messages? | 19:28 | ||
| tadzik | emmentaller was supposed to do that too | ||
| (the cpantesters for perl6, a bare-bones version I wrote even worked for some time) | 19:29 | ||
| colomon | what are the current issues that need to be addressed with it? | 19:30 | |
| tadzik | I may put it on my ultimate "after exams" list | ||
| mikehh | If anyone can let me know what tests are needed I can include that in my parrot and rakudo tests | ||
| tadzik | if I was to resurrect it now, I'd rather rewrite it. Instead of analyzing neutro (brr) output, one could write a module which overrides the Panda::Tester class and make it note down the results | 19:31 | |
| I have no tuits for it now, but can help whoever wants to give it a try | |||
| mberends | mikehh: we don't have any proper tests that are platform agnostic. But I have just thought of a solution, to write a small platform agnostic "for-testing" local library in C. | 19:32 | |
| in fact an own library for testing purposes would achieve better coverage and diagnostics. | 19:33 | ||
| mberends resolves TODO that | 19:34 | ||
| colomon | mberends: you're thinking for zavolaj testing, yes? | ||
| mberends | colomon: yes | ||
| colomon | mberends++ | 19:35 | |
| mberends | that solves the problem within the zavolaj repo, and will involve adding only one more requirement to the Rakudo release guide. | ||
| colomon | tadzik: I don't know how many tuits I'll have, but I'll take a look at the Panda stuff and see if I can sort out what would be needed | 19:36 | |
|
19:36
[particle] left
|
|||
| tadzik | colomon: cool | 19:39 | |
| I see some pies implementation oriented for testing I wrote for popolnik, can send you the code | 19:44 | ||
| colomon | thanks | 19:46 | |
| tadzik | I'll git it once I make sure it works, and how :) | 19:47 | |
|
19:48
[particle] joined
19:55
[particle] left
|
|||
| mberends | it looks like #phasers this week is ending | 19:57 | |
| o/ | |||
|
19:57
mberends left
20:02
[particle] joined
|
|||
| colomon | hmmm, masak tweeted 5 hours ago... | 20:41 | |