Parrot 3.7.0 "Wanda" | parrot.org | Log: irclog.perlgeek.de/parrot/today | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 6 September 2011.
cotto ~~ 00:10
01:05 AzureStone joined, whiteknight joined
whiteknight good evening, #parrot 01:05
01:10 soh_cah_toa joined 01:20 soh_cah_toa joined
soh_cah_toa whiteknight: are there still any tests failures on whiteknight/kill_threads and whiteknight/frontend_parrot2? 01:36
whiteknight: can't do it myself yet b/c i'm having problems w/ winxed trying to build anything (reinstalling for the second time now)
whiteknight soh_cah_toa: whiteknight/frontend_parrot2 already merged to master. I need to delete that branch from github
whiteknight/kill_threads was passing all tests as far as I remember
soh_cah_toa oh ok, cool
01:42 soh_cah_toa_ joined
soh_cah_toa_ whiteknight: just to double check...you don't see 6model making it into this month's release, right? 01:44
whiteknight no
barring a miracle
soh_cah_toa_ ok
whiteknight: on the subject, i just got a copy of "the art of the metaobject protocol" so i wanna figure some things out about 6model :) 01:49
right now, we don't use a mop for pmc's which is why it's kinda inappropriate to call pmc's objects, right?
whiteknight I need to get a copy of that book for myself
soh_cah_toa_: we sort of use a sort of object model for them
soh_cah_toa_ hm, ok 01:50
whiteknight depends how you define "object"
soh_cah_toa_ so, being that it's called a "metaobject protocol" and not "metapmc protocol", doesn't this turn pmc's into fully-fledged objects?
i mean, i'm all for using 6model, but won't it make pmc's a little too high level? 01:52
unless maybe i'm misunderstanding 6model a bit :\\ 02:00
sorear pmcs are objects. 02:02
pmcs are not Objects.
core Parrot has two independant and mostly incompatible object models
this is too many
soh_cah_toa_ ??? 02:03
object vs Object?
sorear object is a technical term as discussed in TaotMOP
Object is one of the parrot core pmcs
soh_cah_toa_ oh i see
sorear Parrot has two object models - the PMC model and the Class/Object model
soh_cah_toa_ ew
sorear every system that has objects has a "mop" 02:04
soh_cah_toa_ so 6model will replace both so that we only have one mop, right?
sorear when you see something like Moose, it's not adding a mop from nothing - it's providing a cleanly factored, object-oriented wrapper for the Perl 5 MOP
exactly
soh_cah_toa_ right 02:05
ok cool
sorear sometimes people reserve the term "mop" for mops that are object-oriented
soh_cah_toa_ alright, i was always under the impression that even though you could get away w/ it, it was kinda "politically incorrect" to say that pmc's are objects 02:09
i guess just not in the "traditional" sense 02:10
sorear soh_cah_toa_: it's not politically incorrect, it's just confusing
soh_cah_toa_ yeah
sorear soh_cah_toa_: "pmcs are objects" is technically correct, but without context people will think you meant "pmcs are Objects", which is wrong
soh_cah_toa_ ok, that makes sense 02:11
whiteknight pmcs do have attributes, do have methods, do have metaobjects (PMCProxy), etc 02:14
they're just very warty, very low-level types of objects
and they're sort of backwards with respect to the metaobject 02:16
dalek sella/tap_stream: 611d6ad | Whiteknight++ | s (10 files):
Start moving the harness towards using an Stream for TAP input. Requires several small changes to FileSystem and Query libraries.
 02:17
sella/tap_stream: ddf30e0 | Whiteknight++ | src/query/ (2 files):
Add in a project stage to project a single value into a sequence of values
sella/tap_stream: 32c64ed | Whiteknight++ | src/ (8 files):
Rework the TestExecutor and the TapParser to take a stream and iterate over it for most parsing tasks. Various other cleanups and fixes. Remove some left-over debugging statements
sella/tap_stream: bccb694 | Whiteknight++ | src/ (6 files):
Several fixes to get inline compilation working again with streams
soh_cah_toa_ whiteknight: you saw my message the other day? how should you specify which dde to read? by address? offset? what? 02:20
for podds, that is
whiteknight soh_cah_toa_: yeah, I'm not really sure. pc seems like what we have most readily available
soh_cah_toa_ actually, does bytecode even have the concept of addresses? i'm not sure 02:21
i don't think so
whiteknight it's relocatable, so the pc value is an offset from the start of the segment 02:24
sorear dde? 02:25
soh_cah_toa_ i thought so
data description entity
the new debug segment
dalek sella/tap_stream: fa265df | Whiteknight++ | src/harness/TestExecutor.winxed:
fix
 02:29
sella: 32c64ed | Whiteknight++ | src/ (8 files):
Rework the TestExecutor and the TapParser to take a stream and iterate over it for most parsing tasks. Various other cleanups and fixes. Remove some left-over debugging statements
 02:33
sella: bccb694 | Whiteknight++ | src/ (6 files):
Several fixes to get inline compilation working again with streams
sella: fa265df | Whiteknight++ | src/harness/TestExecutor.winxed:
fix
sella: d8dfe10 | Whiteknight++ | s (18 files):
Merge branch 'tap_stream'
soh_cah_toa_ anybody ever seen "Cannot open output file runtime/parrot/languages/winxed/winxed.pbc" before while building? 02:37
i've reinstalled 3.7.0 twice but still can't build anything
somehow, a few directories in runtime/parrot/languages are now owned by root 02:41
i'll just pull a new repo, so nevermind
dalek rrot/soh-cah-toa/tt-2155: d7539e4 | soh_cah_toa++ | runtime/ (3 files):
Added README to 'runtime' directory.
 02:42
rrot/soh-cah-toa/tt-2155: f798be4 | soh_cah_toa++ | src/README:
Added README to 'src' directory.
rrot/soh-cah-toa/tt-2155: f0b754f | soh_cah_toa++ | / (2 files):
Renamed src/dynpmc/README.pod to dynpmc.pod and moved it into the 'docs' directory as this is a more appropriate place for documentation like this.
rrot/soh-cah-toa/tt-2155: 061c618 | soh_cah_toa++ | tools/README:
Added README to 'tools' directory.
rrot/soh-cah-toa/tt-2155: 2d781c5 | soh_cah_toa++ | / (12 files):
Corrected copyright notices in README files so that they are for 2011 only.
dukeleto ~~ 03:15
soh_cah_toa_: did you do something with sudo ?
soh_cah_toa_ dukeleto: for 'make install' but i always do a 'make realclean' 03:16
hm, maybe i need 'sudo make realclean'
actually, i never installed that branch so that makes no sense 03:18
whatever :\\ 03:19
benabik o/ 03:39
aloha: smolder? 03:45
aloha benabik: smolder is not automatic, but smolder clients can be
benabik aloha: no, smolder is a web service to collect test reports 03:46
aloha benabik: Okay.
benabik aloha: smolder is also smolder.parrot.org/app/projects/smoke_reports/1
aloha benabik: Okay.
benabik aloha: smolder is also unstable
aloha benabik: Okay.
benabik aloha: smolder?
aloha benabik: smolder is a web service to collect test reports or smolder.parrot.org/app/projects/smoke_reports/1 or unstable
04:36 nbrown joined
dukeleto aloha: smolder is also leaking memory 05:18
aloha dukeleto: Okay.
05:57 mj41 joined, zby_home joined 06:02 rfwazoo joined, perlite joined 06:10 nbrown joined 06:23 zby_home_ joined 06:58 jsut joined 07:04 SHODAN joined 07:10 JimmyZ joined 08:19 contingencyplan joined 08:54 preflex_ joined 10:25 whiteknight joined
dalek esop: 4343c8c | Whiteknight++ | / (6 files):
fix makefile so we build harnesslib.winxed when it changes. Update harnesslib to reflect rosella changes
 10:41
esop: 8478c3b | Whiteknight++ | / (4 files):
Merge branch 'master' of github.com:Whiteknight/jaesop
rrot-linear-algebra: 25b07b0 | Whiteknight++ | t/ (2 files):
Some fixes to get the harness working again after recent rosella changes. Also, clean up the harness to take advantage of the new, cleaner APIs
 11:00
11:06 Eclesia joined
Eclesia hi 11:06
whiteknight hello Eclesia 11:07
11:35 Psyche^ joined 11:48 jsut_ joined, ambs joined 12:00 Eclesia left
whiteknight hmm...I'm getting a weird segfault now in PLA 12:26
the pc looks like it's getting screwed up after calling IMCCompiler.compile_file 12:30
I wonder if that extra call context mls removed from Parrot_pcc_invoke_from_sig_object was actually unnecessary 12:33
benabik o/ 13:31
13:35 JimmyZ joined
whiteknight this is upsetting to me 14:02
blah, have to go. Will dig into it later 14:05
15:07 gbacon joined 15:11 jsut joined 15:49 allison joined 16:35 Kulag joined
cotto ~ ~ 16:41
17:02 wagle joined
dukeleto ~~ 17:27
18:00 Coke joined 18:11 nbrown joined 18:15 Coke joined
benabik o/ 18:17
18:34 SHODAN joined 19:15 HSlayer joined 19:59 contingencyplan joined 20:09 mj41 joined 20:16 whiteknight joined
whiteknight good afternoon, #parrot 20:21
dukeleto whiteknight: howdy 20:23
whiteknight hello duke
dukeleto whiteknight: jolly good day 20:25
whiteknight: what is the current best practice for including rosella in other projects so they are not an external dep? 20:26
whiteknight dukeleto: what do you mean?
dukeleto whiteknight: my parrot-libgit2 has pbc of rosella commited in the repo, which is broken and wrong
whiteknight: i want to use rosella, but I don't want to force people to also install rosella. is that unreasonable? 20:27
whiteknight: i guess a git submodule could solve this
whiteknight that's a good question. We've talked about it before 20:28
obviously we never came to a good conclusion
I've been assuming people will get it from Plumage
cotto what about rakudo's approach?
whiteknight but as a user, if you have an alternate idea I would love to hear it
cotto download/install if it's not detected, otherwise use the installed version 20:29
dukeleto cotto: that is another valid approach. But I think I have an itch to use a git submodule, so I don't have to maintain code to download/install/detect 20:30
whiteknight git submodule works fine for me. You're also welcome to snapshot source if that's easier 20:31
I haven't been too good about tagging particular revisions of the individual libraries as they hit good stable points
20:51 Coke joined 20:53 contingencyplan joined 21:03 contingencyplan joined
tadzik 'evening 21:03
21:22 jsut_ joined 21:29 kj joined
whiteknight cotto: ping 21:35
PerlJam whiteknight++ 21:49
whiteknight don't ++ me yet. That email should have been sent two days ago, at least 21:53
It's about time we start paying attention to what is important 21:54
PerlJam you get ++ anyway :) 21:55
dalek rrot-libgit2: 9a40a3b | dukeleto++ | rosella/ (22 files):
remove rosella pbc
 22:03
rrot-libgit2: 3f59f18 | dukeleto++ | / (2 files):
Add rosella as a git submodule, fixes #4
dukeleto whiteknight: "Around the 1.0 release Perl6 moved to a new separate repository" should probably be "Around the 1.0 release Rakudo moved to a new separate repository" ... 22:07
whiteknight: and while you are at it, if you can insert a "Duke" I would much appreciate it. On whiteknight.github.com/2011/09/10/d...ttles.html
whiteknight where? 22:08
dukeleto whiteknight: s/Jonathan Leto/Jonathan "Duke" Leto/ 22:13
cotto whiteknight, pong 22:14
a blog post appears
cotto investigates
sorear whiteknight: some time ago, when diakopter was trying to sell the CLR to me, he told me that the design phase of the CLR involved a survey of "7 major academic and commercial languages" and a study of what features were needed to support them in an at least adequate way 22:15
whiteknight: you might try asking him for more details
dukeleto whiteknight: “Make Parrot the best platform possible for Rakudo, but do so in a way that adequately supports and does not actively preclude implementations of JavaScript and Python” <-- i can get behind that
cotto sorear, interesting
dukeleto whiteknight: i would love to hear a detailed plan for getting 6model into parrot core. Are you talking about using it wholesale? Rewriting it? 22:16
cotto dukeleto, "Parrot 5"? 22:18
er, whiteknight 22:19
nm, dukeleto
22:22 soh_cah_toa joined
sorear cotto: as someone who uses the result, its adequacy is wanting in some areas 22:23
cotto sorear, do you have a blog where you gripe about that kind of thing? 22:24
sorear no
cotto ok
sorear .NET 1.0, to a first approximation, used a sandbox model based on a chain of delegation 22:27
file accesses are allowed iff file access is allowed for all modules that own methods in the call stack
this system is deprecated now, but the legacy of it is that CLR tail calls are slower than just using a trampoline 22:28
sorear wouldn't know what to write on a blog 22:29
sorear does better when ey has an idea what people want to hear
cotto whiteknight, are you around? 22:37
whiteknight yessir 22:45
dukeleto wow, libgit2 has sexy docs: libgit2.github.com/libgit2/#HEAD/gr...itory_open 22:46
whiteknight sorry, had to eat dinner
cotto whiteknight, I guess that's forgivable.
;)
whiteknight cotto: I wanted to point you towards that new email, and get your thoughts on it
cotto Thank you for being unambiguous that Rakudo is not the source of our problems. 22:47
whiteknight it needs to be said. And re-said.
cotto whiteknight, where's that gist you had of possible future projects for Parrot?
er, projects, cleanups, refactors, etc 22:48
whiteknight gist.github.com/1198521
cotto that's the one
whiteknight that's just my personal sort of todo list. I'd be glad to add projects from other hackers as well
soh_cah_toa oh yay, debugging is there :) 22:49
cotto I'm thinking about what's a good next step.
whiteknight soh_cah_toa: and pretty high up
cotto whiteknight, you might move that the parrot org as a repo or wiki page.
soh_cah_toa that totally made my day 22:50
whiteknight we can move it someplace. I've been thinking we've needed something like a parrot-meta repo for a while
cotto whiteknight, start one then 22:51
and +1
soh_cah_toa yeah, the wiki is a good place for that
though the wiki needs some serious love :\\
cotto or moved to github 22:52
and then loved
dukeleto just coredumped parrot from winxed, for the win 22:53
cotto dukeleto++
whiteknight dukeleto: I did the same thing today in PLA tests. Methinks there is a bug afoot 22:54
cotto One thing I've been thinking about is the relationship between 6model and M0. Originally we planned to get 6model into Parrot, then move to M0, but part of the reason was because we thought we were going to preserve PIR compatibility.
If we're free to throw out the bad parts, it's less imperative to have an intermediate non-M0 6model. 22:55
whiteknight if we do that, it means longer time until we have any 6model 22:56
cotto I haven't convinced myself that this is the best way forward. At this point it's just an idea.
whiteknight I think it's critical enough that we shouldn't delay it any more than possible
cotto yes
whiteknight Considering the immediate beneficial impacts it could have on the development of Puffin and Jaesop, and the optimization potential we would gain throughout 22:58
because it's not just the 6model code, it's the 6model way of doing things: native typed attributes, flexible storage mechanisms, hinting and optimizations, etc
cotto There's also the consideration that it's always easier to write something the second time.
dukeleto whiteknight: on the other hand, a test in parrot-libgit2 started to inexplicabely pass while it used to throw odd errors from winxed 22:59
whiteknight: i think the coredumping from winxed is my fault, though. I was lying about StructView and calling C functions 23:00
dalek rrot-libgit2: 0776962 | dukeleto++ | / (2 files):
Add tests for Git2::Index and calling git_repository_index
 23:01
23:02 whiteknight joined
bubaflub dukeleto: that's great / bizzare 23:03
(re: the parrot-libgit2 fix)
dukeleto bubaflub: indeed
whiteknight dukeleto: I don't know what your problem was. In my case, IMCCompiler.compile_file is definitely causing segfaults 23:08
.compile() does it too 23:10
I'm going to try reverting that one commit from mls, see if that's the cause 23:14
...nope 23:19
23:30 rfw joined, patspam joined
patspam t 23:30
whiteknight u 23:31
Okay, I think I've narrowed down a pretty slim test case. I'm going to open a ticket 23:41
23:51 wagle joined
dalek TT #2191 created by whiteknight++: Problems with IMCCompiler PMC 23:59
TT #2191: trac.parrot.org/parrot/ticket/2191

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!