Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2011-08-30

| Channels | #parrotsketch index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:03 whiteknight joined #parrotsketch
08:24 lucian joined #parrotsketch
13:30 bluescreen joined #parrotsketch
17:18 benabik joined #parrotsketch
17:19 benabik DID: submitted code to GSoC repository  WILL DO: start school, write more about PACT
17:57 atrodo joined #parrotsketch
18:23 pmichaud left #parrotsketch
18:28 Util # Done
18:28 Util * Missed last week's #ps.
18:28 Util * Helped with 64-bit Cygwin issues via Amazon EC2. Progress made, but no resolution yet.
18:28 Util # Plan to do:
18:28 Util * Cygwin issue. Minor Perl 6 work.
18:28 Util # Blockers:
18:28 Util $WORK
18:28 Util # 7-day ticket report:
18:28 Util 4closed: fixed
18:28 Util 2  closed: invalid
18:28 Util 1  closed: wontfix
18:28 Util 2  new
18:28 Util .end
18:40 NotFound joined #parrotsketch
18:48 NotFound What I did:
18:48 NotFound -parrot
18:48 NotFound * Testing and minor fixes.
18:48 NotFound -winxed
18:48 NotFound * Predef constant __DEBUG__ and --debug command line option.
18:48 NotFound * Constant propagation in casts to int and string.
18:49 NotFound * Minor refactors and optimizations.
18:49 NotFound What I will do:
18:49 NotFound No plan
18:49 NotFound EOR
18:59 contingencyplan joined #parrotsketch
19:13 lucian joined #parrotsketch
19:19 kid51 joined #parrotsketch
19:28 cotto_work *did:
19:28 cotto_work - think about our deprecation policy and how to improve it
19:28 cotto_work - started all-hll-test branch to experiment with quick-and-dirty hll testing
19:28 cotto_work - attempted to help rip out threads, got lots of segfaults
19:28 cotto_work - read up on how real assembly languages work (need to do a bit more)
19:28 cotto_work *todo
19:28 cotto_work - finish profiling docs
19:28 cotto_work - get automated hll testing deployed to something
19:28 cotto_work *blockers
19:28 cotto_work - got eMaker Huxley 3d printer kit.  Putting it together is proving highly distracting.
19:28 cotto_work *eor
19:29 cotto_work any last-minute reports?
19:30 lucian joined #parrotsketch
19:31 Util Hello
19:31 cotto_work howdy
19:31 benabik joined #parrotsketch
19:32 dukeleto joined #parrotsketch
19:32 dukeleto hola
19:32 benabik greetings
19:33 lucian joined #parrotsketch
19:33 cotto_work How'd this week go?
19:33 NotFound Hola
19:34 whiteknight joined #parrotsketch
19:35 dukeleto did we have a weekly goal?
19:35 cotto_work I don't think so.
19:35 NotFound Survive hurricanes.
19:35 cotto_work aim high.  That's what I always say.
19:36 atrodo And earthquakes.  don't forget those
19:36 cotto_work Did any issues come up during the past week that need attention?
19:38 whiteknight nope
19:39 cotto_work any questions?
19:39 * dukeleto dips out. will be back soon.
19:39 whiteknight q1q
19:39 cotto_work whiteknight: go for it
19:40 whiteknight I'm thinking I would like to add new PIR ops as variants of get_attribute which can return other native types: N I and S
19:40 benabik Much like the new find/set_lex ops?
19:40 whiteknight for now, they would just unbox results internally, but with 6model and other features they can help improve generated code and eventually offer optimization potential
19:41 whiteknight yes, just like new find/set_lex
19:41 whiteknight OO-heavy PIR code inevitably does lots of unnecessary boxing/unboxing of attribute values. I want to start making that disappear from user code, even if we continue to do it internally for a while
19:42 bubaflub joined #parrotsketch
19:43 plobsing joined #parrotsketch
19:43 whiteknight any thoughts on that?
19:43 NotFound +1
19:43 * benabik likes it.
19:43 * kid51 is caught up in $job problems
19:43 * lucian votes for clearer separation between method retrieval and method calling
19:43 whiteknight lucian: what do you mean?
19:44 lucian Find_method is annoying to use fro languages where methods are just functions
19:44 * lucian is on a phone
19:45 whiteknight okay, that's a separate issue, but one that we do need to tackle
19:45 lucian It was just related to generated code and oop
19:46 NotFound It can impact language interoperation.
19:47 plobsing lucian: how could method retrieval and calling be more separated than having a separate find_method and invoke?
19:48 lucian plobsing: argument passing assumes things (like implicit self). this needs workarounds
19:48 lucian So, rather, argument passing at the call
19:48 whiteknight I definitely agree with that. PCC has a little bit too much magic
19:49 plobsing lucian: there's nothing implicit about self in the bytecode format.
19:49 plobsing if you dislike IMCC helping you out with that, don't mark your subs :method
19:50 atrodo doesn't :method imply other things as well?
19:51 kid51 q1q
19:51 NotFound If you use object.method you are being explicit about object being self.
19:51 NotFound If not, use: method(object_or_whatever)
19:52 lucian plobsing: I didn't know that. Pir seemed to want that
19:52 plobsing atrodo: it makes the PBC loader do magic stuff that we should eliminate at some point
19:52 plobsing but nothing that cannot be achieved by other means
19:52 lucian Also, I need to do methods for lang interop
19:53 plobsing a method is just something an object gives back when querried with find_method
19:53 plobsing there is no need for it to be a sub marked with :method
19:55 plobsing lucian: perhaps we should move this discussion to #parrot to unblock here
19:55 lucian I see
19:55 * lucian nods
19:55 cotto_work sounds like a good idea
19:56 cotto_work kid51, go ahead
19:56 kid51 let me report
19:56 * kid51 cagecleaned, i.e., reviewed about 20 Trac tickets and posted requests for review; nwellnhof++ for responding on 3 items; I've taken over TT #1044 and am stumbling through it.
19:56 * kid51 also did testing of other people's (i.e., whiteknight's) branches; have build failures on frontend_parrot2
19:57 kid51 * We have 512 open tickets (view 10), which is a handful fewer than last week.
19:57 kid51 * PDS was 4 weekends ago; can someone change the topic in this room?
19:57 kid51 * We're nearing halfway point in quarterly supported release cycle -- which means I will have to start bugging people about meeting roadmap goals very soon.
19:57 Topic for #parrotsketch is now 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: http://irclog.perlgeek.de/
19:57 kid51 * Best way not to be annoyed by that:  Work on the roadmap goals you signed up for :-)
19:57 kid51 EOQ
19:58 NotFound kid51: I fixed some problems in frontend_parrot2
19:58 kid51 Re: TT #1044: Eliminating our configuration dependence on Perl 5's $Config{ccflags} setting is going to be much more difficult than one might think.
19:59 kid51 NotFound: I will have to test when I get home
20:01 kid51 Please review http://trac.parrot.org/parrot/ticket/2184 and let Coke know if you have any objection to his merging it.
20:06 NotFound +1 to TT #1044
20:06 NotFound I mean, 2184
20:07 cotto_work kid51: I'm happy leaving 1044 on the back burner.
20:08 kid51 Well, I'll poke away at it as it's one of the few tickets I'm competent to handle.
20:08 kid51 but I won't rush it
20:08 * kid51 has to re-focus on $job; will backscroll later
20:10 cotto_work Does anything else need attention?
20:11 benabik Just a note: $class will probably keep me away from #ps for the next few months.
20:12 plobsing cotto_work: it has come up (again) that the "constant" table is misleadingly named
20:13 plobsing I'd like to consider a rename
20:13 whiteknight to what?
20:13 plobsing the best I can do is "object
20:14 plobsing which admitedly isn't much better
20:14 NotFound "thing" ;)
20:14 whiteknight it's a lot of work to replace something that isn't right with something else that isn't any better
20:14 plobsing but does not conote "objects declared as constants in the sourcecode"
20:14 plobsing I'm open to suggestions
20:14 cotto_work a good name is a prerequisite
20:15 whiteknight I like the idea of changing it, but I don't know what it would be
20:15 cotto_work same here
20:15 NotFound Is not so misleading, anyway. The table can point to non constant objects, but the values in the table are, aren't they?
20:15 plobsing true it can be considered a table of references (PMCs have ref semantics) which cannot be changed
20:16 plobsing but there are too many subtly different senses of the word "constant"
20:17 cotto_work If anyone thinks of a better name, we can use it.
20:17 NotFound It has non constant sense ;)
20:18 plobsing how about something along the lines of "the packfile pmc/string/num table"
20:21 NotFound Packfile store?
20:22 plobsing I'd like to buy a PMC
20:23 cotto_work that might work
20:23 plobsing anything that is *explicit* about packfile is likely much better
20:24 cotto_work plobsing: can you start a quick ticket?
20:24 plobsing cotto_work: will do
20:26 cotto_work anything else before we call it a wrap?
20:26 cotto_work plobsing: thanks
20:34 cotto_work This meeting seems to be over.  Let's call it a wrap.
20:35 NotFound left #parrotsketch
20:35 benabik left #parrotsketch
20:39 PacoLinux left #parrotsketch
20:52 bluescreen joined #parrotsketch
23:02 bubaflub joined #parrotsketch
23:18 Coke joined #parrotsketch
23:25 Coke joined #parrotsketch
23:32 Coke joined #parrotsketch
23:34 whiteknight joined #parrotsketch
23:38 Coke_ joined #parrotsketch
23:43 Coke joined #parrotsketch
23:49 Coke joined #parrotsketch
23:54 Coke_ joined #parrotsketch

| Channels | #parrotsketch index | Today | | Search | Google Search | Plain-Text | summary