|
Post closed tickets in your report. | Note: This channel is for our weekly status meetings (Tuesdays at 19:30 UTC); you probably want #parrot instead. | irclog: irclog.perlgeek.de/ Set by moderator on 29 September 2011. |
|||
|
02:48
bluescreen joined
03:55
tewk__ joined
04:00
tewk joined
07:23
lucian joined
08:16
contingencyplan joined
18:47
contingencyplan joined
19:19
nine joined
19:20
benabik joined
19:22
benabik joined
|
|||
| benabik | I updated the random notes I had on "how I'd write PCT" into something resembling documentation - github.com/parrot/PACT Input very welcome. I may hack on this off and on for a while, but my tuits are all being spent on school. | 19:26 | |
| (EOR) | |||
| cotto | done: working on M0 research, slow due to poor tuits | ||
| hello | 19:33 | ||
| Util | Hello | 19:35 | |
| No report for me; $WORK persists | |||
| nine | Got green_threads into a working state. Will try to reduce overhead in the single task case to zero. I'm unsure about how to improve the voluntary task switching, but hopefully will come up with something. Upcoming exams may reduce my tuits somewhat in the next two weeks. | 19:37 | |
| dukeleto is here, but no report. Mostly updated our Changelog with news since the last release. | |||
| cotto | Is anyone interested in jumping in to help with M0 research? It's been progressing, but much more slowly than I'd like due to tuits. | 19:38 | |
| (not asking for a firm commitment, just interest) | |||
| dukeleto | cotto: where does the research live? what exactly needs to be researched? | ||
| dukeleto puts on devil's advocate hat | 19:39 | ||
| cotto | dukeleto: currently I'm looking at how common cpu architectures deal with various types | ||
| dukeleto | cotto: are we doing too much research on M0? Perhaps we should forge ahead with iterative improvements? | ||
| cotto | what can be manipulated directly, what the arch knows about, etc | ||
| dukeleto | cotto: we could research until the heat death of the universe, and still not have a new VM core. How much research are we going to do? | 19:40 | |
| cotto | dukeleto: I have a fairly well-defined stopping poitn (fill in a table for 4 cpus). | ||
| *point | |||
| dukeleto | cotto: i would really like to see this research in the m0-spec branch. Is that possible? | 19:41 | |
| cotto | dukeleto: quite possible. I'll do that before I do anything else. | ||
| actually, how about the github wiki? | 19:42 | ||
| dukeleto | cotto: since the github wiki is also a repo, yes. | 19:44 | |
| cotto: somewhere in some parrot git repo will satisfy my needs :) | 19:45 | ||
| Util | A partially completed table (or just the empty but well-defined table) would be a prod to contributors | ||
| cotto | sounds like a plan | ||
| any other thoughts (related or otherwise)? | 19:47 | ||
| Util | q1 unrelated | 19:50 | |
| (dequeuing) | |||
| A thought for those with more tuits than I: According to my (fixed) run of parrot_coverage.pl on Sep 7th, 347 of Parrot's 2198 opcodes are completely untested (0% coverage of a function in core_ops.c.gcov) by the test suite. | |||
| I assume that we will not try to test many of those 347, due to M0 plans. | |||
| However, if gcov was run against the test suite of a client language (like Rakudo), we *might* find that some untested opcodes are used by a client. | |||
| If so, those opcodes would be worth immediate attention in the test suite. | 19:51 | ||
| END_OF_THOUGHT | |||
| (and yes, I need to dust off my fix to parrot_coverage.pl and commit it) | |||
| cotto | Util: good idea | 19:54 | |
|
20:03
benabik_ joined
|
|||
| cotto | any other thoughts? | 20:04 | |
| let's call it a wrap | 20:10 | ||
|
21:29
benabik left
23:03
bubaflub joined
23:10
whiteknight joined
|
|||