Parrot 1.8.0 Zygodactyly released | cs_csr_merge branch need testing! | Priority: trac.parrot.org/parrot/wiki/ClassV...eOverrides | Latest modified TT's: icanhaz.com/parrotbugs | Parrot Languages: icanhaz.com/parrotlang
Set by moderator on 3 December 2009.
chromatic You send him to school, you buy him books, you read his thesis, and he still types 'make install'. 00:03
00:13 brrant joined, plobsing_ joined
dalek TT #1358 closed by coke++: 'make realclean' doesn't clean some nqp-rx files 00:15
rrot-plumage: 88655fa | japhb++ | TODO:
[META] Update TODO
 00:17
rrot-plumage: 3453ef9 | japhb++ | TODO:
[META] Add doc TODOs
rrot-plumage: dbe72a0 | japhb++ | src/ (2 files):
[plumage,LIB] More Plumage::Project refactoring

  * Move responsibility for mkpath($build_root) to Plumage::Project's fetch action
  * Excise %STAGES and supporting code from plumage.nqp
  * Improve API of perform_actions{,_on_projects}
  * Add stub of update stage to Plumage::Project, reusing (and falling back to) fetch
00:17 plobsing1 joined
japhb wheeee 00:18
Infinoid++ again
I CAN HAZ VERBOSITY
00:18 plobsing1 left
darbelo Infinoid++ # WE CAN HAZ LONG COMMIT MSGZ! 00:22
Oh, that reminds me who's LOLCODE's keeper nowadays?
t/01-vars.t ............ Failed 1/6 subtests
dalek tracwiki: v44 | bacek++ | ParrotQuotes 00:23
tracwiki: trac.parrot.org/parrot/wiki/ParrotQ...ction=diff
Whiteknight chromatic: need to make install to test things like pla and matrixy 00:29
what's surprising is that Parrot doesn't have a "make uninstall" target
...and you read my thesis?
00:29 tetragon joined
chromatic No one read mine. 00:30
japhb chromatic, you had a thesis? On what?
Whiteknight if I had to guess, I would say music theory 00:31
chromatic I was in the performance track, so no thesis there.
dalek TT #1336 closed by mikehh++: Convert t/pmc/bigint.t to PIR 00:32
chromatic For the history major... well, let's just put it this way, no one's interested unless they want to trace the obscure philosophical development of a line of thought from the second century to the 11th.
japhb OK, I'm at least curious enough to want to know what that line of thought actually was .... 00:34
chromatic If I'd been clever enough at the time, it would have been Augustinianism, Aristotelianism, and the Veneration of Mary. 00:35
mikehh All tests PASS (pre/post-config, smoke (#30591), fulltest) at r42897 - Ubuntu 9.10 amd64 (g++ with --optimize) 00:37
moderator Parrot 1.8.0 Zygodactyly released | Priority: trac.parrot.org/parrot/wiki/ClassV...eOverrides | Latest modified TT's: icanhaz.com/parrotbugs | Parrot Languages: icanhaz.com/parrotlang 00:37
japhb chromatic, but given that intro, you weren't, so ... what was it then? :-) 00:37
chromatic I missed Aristotle. 00:41
dalek rrot-plumage: 77385e8 | japhb++ | src/plumage.nqp:
[plumage] Minor cleanups
 00:44
rrot-plumage: a364c53 | japhb++ | t/ (2 files):
[TESTS] exit(1) after a failing test run; fix broken test missed because failing tests used to exit(0)
japhb Is it intended that parrot subs can be called with extra named arguments that the callee does not expect to receive, and no error occurs? 00:48
Whiteknight japhb: yes, I think there is a setting though 00:49
darbelo Intended or not, I think it's been that way for a while.
japhb Whiteknight, a setting? Where would I look for that?
Whiteknight I have no idea, to be honest 00:50
japhb darbelo, sure, I'm not surprised by that. (Well, actually, I'm curious if it was before or after the PCC merge). I was just curious if that was regarded as proper behavior, or an unfixed bug.
Coke I would imagine that the error is supposed to occur, actually, based on a reading of the CC pdd.
chromatic There's a warning to enable, but nothing does.
Coke note that the extra-args check is deliberately disabled on args that aren't declared as taking any params.
japhb Coke, ... because we expect them to look at the call sig instead? 00:51
Coke call sig didn't exist when that decision was made. 00:54
dalek tracwiki: v45 | cotto++ | ParrotQuotes 00:56
tracwiki: This is how it usually works around here.
tracwiki: trac.parrot.org/parrot/wiki/ParrotQ...ction=diff
japhb Coke, some sort of pasm primitive support then?
00:59 kid51 joined
cotto_work chromatic, there's an "=point" in your blog post that looks out-of-place. 01:03
also, C<0 but true> 01:05
chromatic Thanks, cotto_work. 01:09
dalek rrot-plumage: 4770abe | japhb++ | src/plumage.nqp:
[plumage] Tell command_ subs what command invoked them; collapse all project action commands into one sub
 01:17
mikehh messages 01:19
cotto_work You have no new messages waiting.
mikehh cotto_work: just clearing messages I have sorted out :-} 01:21
cotto_work That's probably why you don't have any new ones. ;}
kid51 msg bubaflub Can you comment on trac.parrot.org/parrot/ticket/1267#comment:6 ? Thanks. 01:34
purl Message for bubaflub stored.
dalek TT #1327 closed by jkeenan++: remove old languages files 01:38
rrot: r42898 | jkeenan++ | trunk/config/gen/makefiles/docs.in:
Applying patch submitted by Andy Dougherty in �trac.parrot.org/parrot/ticket/1216.
 01:56
TT #1359 created by plobsing++: [RFC] Merge freeze/thaw with PBC 02:28
02:39 davidfetter joined
dukeleto 'ello 02:44
japhb o/
dalek rrot: r42899 | jkeenan++ | trunk/config/auto/headers.pm:
Delete commented-out code.
 02:46
cotto ITS CATURDAY 02:54
or at least close enough that I'm no longer required to do anything in particular
Whiteknight CATURDAY! 02:58
Whiteknight slowly surfs to 4chan
dalek rrot-plumage: 3f6629b | japhb++ | (3 files):
[DOCS,META] Add add-action document; refer to it in add-command; update CONTRIBUTING in the README
 03:03
03:07 davidfetter joined
cotto finally gets his pafo cla ready to send off 03:17
04:00 Essobi joined
japhb Grrr. 'sudo make install' does not work in parrot HEAD, because of the redundant rebuild of docs/ops/ using perldoc, which drops privileges ... to user nobody, thus not having permissions to overwrite the files, which are mode 0644. 04:30
dukeleto japhb: patches welcome? 05:07
purl patches welcome is ponies welcome or Set Objectives, Achieve Results! or swahili for "Put up or shut up."
japhb dukeleto, I'm actually in the process of trying to do that.
dukeleto cotto: put it in the mail!
japhb: great to hear
japhb The first problem is that I'm mistified as to why the redundant rebuild is happening in the first place.
(trying something ....) Oh, that worked. Hmmmm. Now to figure out why. 05:08
05:09 cotto_w0rk joined
japhb Ah. Two revs ago, jkeenan applied a patch from doughera to the docs makefile. I'm going to take a wild guess this is the problem. 05:12
05:14 cotto_working joined
dalek pir: 0513ece | dukeleto++ | TODO:
Clarify the TODO
 05:17
pir: f1b25ca | dukeleto++ | t/harness.pir:
Properly print a newline when a test file fails with exit code 0
japhb Yup, r42898 is bad; updating trac 05:26
05:38 JimmyZ joined
JimmyZ good afternoon, #parrot 05:42
davidfetter good evening, JimmyZ 05:45
05:45 cotto_work joined
JimmyZ good evening, chromatic & davidfetter 05:45
dukeleto hola 05:47
05:50 cotto_w0rk joined
davidfetter dukeleto, long time no see :) 05:55
06:01 JimmyZ joined
JimmyZ ;) 06:02
06:02 JimmyZ left 06:04 JimmyZ joined
cotto It's nice to see Tapir using the exit_status method before I got around to using it for the purpose that I originally implemented it. 06:37
06:40 bacek joined
bacek o hai 06:40
cotto hio bacek
bacek cotto, glad to see you without "_work" suffix :) 06:41
cotto yeah. I can commit stuff now. 06:42
and the Internet connection is a low slower :(
s/low/lot/ 06:43
bacek cotto: if you'll commit new shiny generation GC we can wait little bit :) 06:44
cotto I can commit a generational GC that breaks the build and segfaults a lot. How about that? 06:45
also, it won't be a generational GC
more like a G 06:46
chromatic I thought Whiteknight was going to produce some documentation on the GC api. 06:48
cotto ISTR it's on his long todo list 06:54
moritz Coke: fix for TT#1358 confirmed 07:47
dukeleto davidfetter: i barely ever run into you ;) 07:50
davidfetter dukeleto, i should hope not. that'd hurt
dukeleto cotto: what about working on your G(C) on a branch? maybe a git branch? 07:52
cotto It wasn't a serious offer. 07:53
However, this weekend would be a good time to do some profiling work.
dukeleto cotto: don't you know quips on #parrot are legally binding? i will expect a G(C) branch right away ;) 07:56
how do i get the proper CPPFLAGS that i need to use parrot header files in C? how do i programatically figure out that i need to do, say, -I/usr/local/include/parrot/1.8.0 ? fperrad or japhb might know 08:02
PConfig{ccflags}, looks like 08:07
parrot_config ccflags exists 08:08
but no cppflags
cotto I smell a patch. 08:11
davidfetter smell that? do you smell that? 08:12
nothing else in the world smells like that
cotto we don't appear to have any cppflags 08:17
davidfetter i love the smell of cppflags in the morning 08:18
chromatic Do you have pkg-config installed? See parrot.pc 08:32
08:37 iblechbot joined
dalek parrot: 415fb6b | dukeleto++ | (2 files):
First attempt to get pgTAP to work with PL/Parrot

ERROR: column c.relistemp does not exist at character 308 STATEMENT: CREATE OR REPLACE FUNCTION _temptypes( TEXT )
   RETURNS TEXT AS $$
   SELECT array_to_string(ARRAY(
   SELECT pg_catalog.format_type(a.atttypid, a.atttypmod)
   FROM pg_catalog.pg_attribute a
   JOIN pg_catalog.pg_class c ON a.attrelid = c.oid
   WHERE c.relname = $1
   AND c.relistemp
   AND attnum > 0
   AND CASE WHEN attisdropped THEN false ELSE pg_type_is_visible(a.atttypid) END
   ORDER BY attnum
   ), ',');
   $$ LANGUAGE sql;		 08:39
dukeleto oy vey
parrot: f3687f4 | dukeleto++ | (2 files):
Add a tweaked pgtap.sql that works on my setup
parrot: 7d60aac | dukeleto++ | HOWTO:
Thankfully, we can set USE_PGXS=1 and then we do not need the Postgres source to compile. theory++
parrot: 339eaaa | dukeleto++ | TODO:
Add a TODO file
parrot: cc93f25 | dukeleto++ | (2 files):
Merge branch 'pgtap'
 08:40
parrot: d01a03c | dukeleto++ | src/handler/Makefile:
Grab the postgres version number for later inspection, thanks pgTAP and theory++
parrot: 93c8025 | dukeleto++ | src/handler/ (2 files):
First attempt to get the proper ccflags and junk to embed Parrot in the language handler
dukeleto i would say *incoming*, but it is too late
ERROR: incompatible library "/Users/leto/git/postgresql/contrib/plparrot/libplparrot.so": missing magic block 08:42
who took my magic block?
ENOMAGICBLOCK 08:43
davidfetter dukeleto, nu? 08:49
PG_MODULE_MAGIC;
dukeleto davidfetter: it was a red herring 08:50
davidfetter: i was messing with CPPFLAGS and screwing the pooch, evidently
chromatic Your compiler flags are not in the pooch. 08:51
dukeleto chromatic: how do you know if you don't look? 08:52
davidfetter wonders if this is going to work
chromatic It's too dark inside a dog to read. 08:53
dukeleto chromatic: not if you bring a blowtorch 08:54
dalek parrot: a705c02 | dukeleto++ | src/handler/Makefile:
Fiddling with this actually breaks PL/Parrot. Let's not do that.

1..2 ERROR: incompatible library "/Users/leto/git/postgresql/contrib/plparrot/libplparrot.so": missing magic block HINT: Extension libraries are required to use the PG_MODULE_MAGIC macro.		 08:56
parrot: 2fe1598 | dukeleto++ | Makefile:
Makefile for the lazy bums like me who want to type less
08:57 fperrad joined
davidfetter just include it unconditionally. we're going to EOL the versions that didn't have it soon anyhow :) 08:57
dukeleto davidfetter: theory gave me lots of help at the pdx hackathon. he showed me pgTAP's Makefile and explained it 08:59
dalek tracwiki: v1 | cotto++ | CottoTasklist 09:00
tracwiki: initial version
tracwiki: trac.parrot.org/parrot/wiki/CottoTa...ction=diff
davidfetter dukeleto, yay!
09:00 fperrad_ joined
dukeleto is currently stealing bits from plperl.c 09:05
davidfetter excellent!
purl plays air guitar
davidfetter the one tim bunce fixed up?
dukeleto davidfetter: i am looking at the version in 8.3.8 09:06
davidfetter: and only looking at the basics. when did bunce fix it up?
davidfetter initialization, safe containers, stuff like that
properly runs END blocks as of the patch
dukeleto davidfetter: i am down in the internals, so that stuff doesn't effect me much, but good to know about 09:08
09:10 fperrad_ joined 09:17 fperrad_ joined 09:23 fperrad_ joined 09:35 fperrad_ joined
dalek tracwiki: v2 | chromatic++ | CottoTasklist 09:50
tracwiki: trac.parrot.org/parrot/wiki/CottoTa...ction=diff
cotto Yay! Answers. 09:51
chromatic++ 09:52
dalek parrot: d3accdd | dukeleto++ | src/handler/plparrot.c:
Add a basic skeleton for plparrot_call_handler
 09:54
parrot: 4761a72 | dukeleto++ | Makefile:
Improve Makefile
parrot: e07cd49 | dukeleto++ | src/handler/plparrot.c:
Make a test pass by returning the first element of whatever the call_handler function was given
09:58 JimmyZ joined 10:04 fperrad joined 10:21 fperrad joined 10:25 szabgab joined 10:50 fperrad_ joined 10:53 fperrad joined 11:05 mikehh joined 11:16 viklund joined
dalek l: d552308 | fperrad++ | xml/sax/xml/grammar.pg:
remove POD
 11:17
11:38 iblechbot joined
dalek nxed: r239 | julian.notfound++ | trunk/Makefile:
use harness in stage 1 tests
 12:40
12:46 mikehh joined 13:12 iblechbot joined 13:34 joeri joined 13:46 kid51 joined
dalek rrot: r42900 | jkeenan++ | trunk/config/gen/makefiles/docs.in:
Reverting r42898 per report in �trac.parrot.org/parrot/ticket/1216.
 14:01
14:13 lucian joined 14:17 payload joined, naypalm joined 14:29 payload joined 14:45 payload joined 14:50 mikehh joined
mikehh All tests PASS (pre/post-config, smoke (#30606), fulltest) at r42900 - Ubuntu 9.10 i386 (g++ with --optimize) 15:09
15:29 he joined 15:30 he_ joined 15:50 Psyche^ joined
dalek pir: b6b04ec | Whiteknight++ | (2 files):
quick todo note
 16:17
pir: 20fe6a4 | Whiteknight++ | (3 files):
Merge branch 'master' of git@github.com:leto/tapir
p-rx: e6cb831 | pmichaud++ | src/Regex/Regex.pir:
Initial version of regex matching.
p-rx: fe021de | pmichaud++ | (4 files):
Add ability to smartmatch regex/token/rule (although anchoring is
p-rx: 01c9533 | pmichaud++ | (5 files):
Refactoring for Regex::Method.
p-rx: 6297c85 | pmichaud++ | src/Regex/Regex.pir:
Remove obsolete file.
p-rx: df1c1da | pmichaud++ | src/NQP/Actions.pm:
Small refactor of regex/token/rule action.
p-rx: 5a0677a | pmichaud++ | (3 files):
Add /.../ form of regexes and matching.
16:20 cconstantine joined 16:47 mikehh joined 16:53 payload joined
Coke japhb: I was AFK; I think you're looking for a super well thought out design decision that let subs with no params have no args checking; I think it was just so you could write one-offs with a :main that got passed an ignored argv. 16:57
*passed in 17:00
er, no, I had it right the first time.
17:00 flh joined
flh hi everyone 17:02
purl Howdy, flh, you fantastic person you.
Coke down, purl. 17:03
purl And down SAFE!
17:05 patspam joined, bubaflub joined
flh I'm a bit annoyed by the recent CallSignature(Returns) refactor: it seems like I can't build by hand a full (i.e., with returns) CallSignature from C 17:09
I didn't find any API to play with the returns_* attributes of CallSignature: have I missed it? 17:12
17:14 mikehh joined
japhb Coke: Ah, I understand, thank you. 17:20
17:44 Zak joined 17:48 mikehh joined 17:54 mikehh joined 18:01 mikehh joined 18:06 Zak joined 18:08 mikehh joined
flh parrot has named args, but can we also have "named returns"? I couldn't find any example in Parrot sources 18:09
cotto f1h, yes. It works both ways iirc though named returns are much less common. 18:18
dalek nxed: r240 | julian.notfound++ | trunk/ (2 files):
add --stage option to t/harness and fix mistake in make test1, PacoLinux++
 18:27
cotto flh, the only example seems to be in t/op/calling.t starting on line 1921. 18:31
Consider yourself encouraged to write more tests.
flh I'll remember it :) yet for the moment, I rather consider myself encouraged to patch CallSignature 18:32
18:36 davidfetter joined 18:43 theory joined 19:00 preflex joined 19:06 theory joined
dalek rtcl-nqp: f634404 | (Will Coleda)++ | (16 files):
Add rudimentary [package], about the level of partcl's

  * library/ directory (copied from partcl copied from tcl)
  * [package require (tcltest|opt)]		 19:16
Coke pmichaud: ping 19:18
Tene: ping 19:21
19:29 chromatic joined 19:47 payload joined
dukeleto should we update parrot.org to point people at @parrotvm on twitter and @parrot on identi.ca ? 20:09
i see a steady stream of people following both accounts, but i have no idea how they find out about them 20:10
chromatic +1
purl 1
dukeleto chromatic: are the parrot.org htdocs stored in a repo somewhere? how do they get updated?
cotto dukeleto, Maybe the like birds and got confused. 20:11
;) 20:12
chromatic I think it's Drupal.
dukeleto Oh. My. $deity. 20:13
do we have an in-house web-design person? i try not to touch that stuff with a 10-foot pole 20:14
20:14 lucian joined
dukeleto all we actually need to do is decide how we want to link to twitter/identi.ca . two small text links are a good start 20:15
we can get all fancy and make them custom images if we want, or use one of the standard twitter images
so if I have a login to parrot.org, in theory i can update the templates associated with the index page? 20:16
i have never used drupal
20:18 Zak joined
chromatic I believe you can. 20:18
20:19 PacoLinux joined
plobsing_ gc question: roughly how much improvement would we see by not marking 500..600 leaf PMCs per GC run? 20:20
chromatic That depends on how many other PMCs we have to mark and sweep and the ratio of live/recyclable PMCs. 20:21
plobsing_ I'm just wondering if that is a small or large number of PMCs and whether it is worth the effort trying to eliminate their marking. 20:25
chromatic Which PMCs are those? 20:26
plobsing_ cached NCI call PMCs
they are constant and live the life of the interpreter. 20:27
chromatic Where do you see them?
dukeleto shouln't caches not be marked unless invalidated? 20:28
chromatic That depends what kind of cache.
plobsing_ src/nci.c, build_call_func 20:29
what I am thinking is that these should somehow be put in some non-collectible segment of memory (is the constant pool collectable?) 20:30
chromatic It isn't.
plobsing_ and then have the hash caching these contain pointers that the GC won't follow
chromatic It doesn't get swept; that's how it remains constant. 20:31
plobsing_ like a hash of ints
this is a slight hack, which is why I ask whether its worth it
chromatic You can try it. 20:32
Make them all constant PMCs (and make sure they don't contain any pointers to non-constant STRINGs or PMCs), make the HashPointer PMC constant, and after the first GC run, they won't get marked.
... unless they're reachable from somewhere other than IGLOBALS_NCI_FUNCS. 20:33
plobsing_ wait a sec, the HashPointer PMC might not be constant. It will always point to constants, but its contents might grow if we have an actual frame builder. 20:34
is that OK for the constant pool?
chromatic As long as it never points to anything that isn't constant, it's fine.
Complications set in at src/pmc/nci.pmc lines 127, 131-133.
plobsing_ its not NCI pmcs 20:35
its the unmanaged struct PMCs that manage the C function pointers
chromatic Oh, right. I'm misreading this. 20:36
As long as they don't have non-constant initializers, they should be fine.
dalek nxed: r241 | julian.notfound++ | trunk/ (2 files):
Improve assignment in stage 1
 20:37
chromatic src/nci.c looks like the only place to use IGLOBALS_NCI_FUNCS, so you should be free and clear to make that a constant PMC. 20:38
Of course, that only applies to code that uses NCI. We delay initialization of the thunks until something needs one. 20:42
plobsing_ calling a non-vtable method on a core PMC fits that. I would assume thats a fairly common operation. 20:44
20:45 bubaflub joined
chromatic True. 20:45
pmichaud Coke: pong 20:46
dukeleto whoa, i just found my PL/Parrot talk on youtube: www.youtube.com/watch?v=WFbTAVleQ7s 20:49
i insulted everyone equally :)
the rest of the talks: www.pythian.com/news/5807/opensqlca...lk-videos/
Tene Coke: pong 20:51
cotto win 20:59
plobsing_ hmmm... the constant hash doesn't like being cloned (required to spawn threads apparently) 21:04
what should I use for benchmarking? 21:13
dukeleto plobsing_: euler_bench 21:14
purl it has been said that euler_bench is github.com/notbenh/euler_bench or that which dukeleto uses to benchmark parrot
plobsing_ dukeleto: are there any there that are particularily gc-heavy? 21:17
dukeleto plobsing_: www.nntp.perl.org/group/perl.perl5....52921.html
plobsing_: no, that is a benchmarking framework
21:18 mikehh joined
dukeleto plobsing_: there are gc benchmarks here: examples/benchmarks/gc_* 21:19
plobsing_: what are you trying to benchmark? a branch? against what?
chromatic oofib.pir can be GC heavy 21:20
the NQP-rx Actions.pm benchmark runs the GC a lot.
mikehh All tests PASS (pre/post-config, smoke (#30623), fulltest) at r42900 - Ubuntu 9.10 i386 (gcc with --optimize) 21:27
cotto What's the difference between CONTEXT() and CURRENT_CONTEXT()? 21:30
It's not entirely clear from tracing through the code. 21:31
dukeleto cotto: find the definition of each macro? 21:50
cotto That wasn't very helpful. 21:52
#define CURRENT_CONTEXT(interp) ((interp)->ctx) 21:53
#define CONTEXT(interp) Parrot_pcc_get_context_struct((interp), (interp)->ctx)
That function basically does the same thing. 21:54
pmichaud one of them uses the API, the other doesn't.
(that's my guess.) 21:55
cotto I'll poke bacek when he's on. He seems to know something about that part of the code. 21:56
dukeleto cotto: CONTEXT() looks newer 22:01
cotto: CONTEXT uses the pcc api, CURRENT_CONTEXT does not
cotto: so yes, they seem to do the same thing, but one is probably the newer, preferred API. 22:02
dalek nxed: r242 | julian.notfound++ | trunk/winxedst1.winxed:
operator - in stage 1
 22:05
22:06 plobsing_ joined
cotto silly bird 22:07
22:10 kid51 joined
Coke tene - partcl-nqp is failing a test with a for loop that contains an if that contains a break. the break isn't breaking. 22:11
pmichaud: how can I make, say, 1 == 2 invoke a function instead of a pirop? 22:12
nopaste "pmichaud" at 66.25.4.52 pasted "...what am I forgetting here?" (25 lines) at nopaste.snit.ch/18974 22:13
pmichaud Coke: just eliminate the :pirop from the token definition. It will then call a function called &infix:<==>
I'm planning to also allow a :name option that will let you specify the function to be called... just haven't done that yet. 22:14
cotto dukeleto, both CONTEXT and CURRENT_CONTEXT were added by bacek in the same (apparent) commit. 22:16
dukeleto cotto: was it a merge commit? 22:17
Coke pmichaud: you need to register the hll map in an :init block.
see partcl-nqp src/class/tclstring.pmc
s/pmc/pir
pmichaud ...it has to be in an init block? weird.
cotto r40958
checking...
dalek n: 0c0d119 | fperrad++ | (8 files):
update infrastructure with setup.pir (distutils)
n: dd934e1 | fperrad++ | (4 files):
refactor with opcode 'load_language'
n: c246251 | fperrad++ | plumage/fun.json:
add a Plumage description
n: 3606e26 | Jarrod++ | src/ (2 files):
Added changes to the language setup, provided by fperrad++
dukeleto Merge context_pmc3 branch into trunk.
DRAGONS ENTER THE BUILDING.
Coke double checking, but... I think so.
dukeleto is r40958
pmichaud nopaste coming
dukeleto cotto: here is your gold watch ;)
cotto yup
dukeleto i though i recognized that commit number :) 22:19
i have received many gold watches as well
nopaste "pmichaud" at 66.25.4.52 pasted "still no luck with hll_map" (27 lines) at nopaste.snit.ch/18976
pmichaud if I do an explicit "box" in main, it gets the correct hll_mapped type, though.
Coke ah. 22:20
nopaste "pmichaud" at 66.25.4.52 pasted "still no luck with hll_map, #2" (32 lines) at nopaste.snit.ch/18977
Coke pmichaud: perhaps a missing use of the hll ctx function. 22:21
plobsing_ dukeleto: how do I get euler_bench to work against 2 differing parrot interpreters?
Coke (the fact that it works with box means you're registering it properly.)
pmichaud I wonder if that's a PCC refactor bug.
pmichaud tries an older parrot. 22:22
cotto msg bacek What's the difference between CONTEXT() and CURRENT_CONTEXT() and when should they be used?
purl Message for bacek stored.
cotto yay for the magical messaging robot 22:23
< purl-- >
dukeleto plobsing_: did you read the post i made to p5p? you list the paths to those interpreters in the yaml file 22:24
plobsing_: then you can run it like this: BENCH_CONFIG=bench.yaml ./bin/bench ~/svn/parrot/examples/benchmarks/include_pir.pir parrot --count=100 --order=avg 22:25
listing parrot in the args means it looks for all interpreters assigned to the parrot language in the yaml file
plobsing_: there should be an example yaml file that shows how to do it 22:26
Coke which namespace is the &infix:<==> in ? 22:28
the grammar's?
pmichaud it's a normal function call, like any other (so 'commands' for now) 22:29
sub &infix:<==>($a, $b) { ... } 22:30
looks like we lost hll mapping of arguments 22:31
chromatic I thought CURRENT_CONTEXT() was a source-level backwards compatibility macro.
Coke pmichaud: hurm. I would expect that to have broken partcl. 22:32
chromatic HLL mapping of auto-promoted arguments?
Coke runs a make test over there.
pmichaud chromatic: yes.
nopaste coming
nopaste "pmichaud" at 66.25.4.52 pasted "no mapping of native arguments in trunk" (35 lines) at nopaste.snit.ch/18978 22:34
pmichaud interestingly, return values *are* mapped.
chromatic That sounds easy to fix.
pmichaud so the hll_map test in t/pmc/parrotinterpreter.t passes
shall I file the ticket or wait for the imminent fix? 22:35
chromatic Let me read the code and think about it.
pmichaud (sounds like "file the ticket" then :-)
chromatic If I don't have it fixed in an hour, file away. 22:36
pmichaud okay.
chromatic That was easy. I thought there was a TODO about it. 22:40
That should do it. 22:48
dalek rrot: r42901 | cotto++ | trunk/src/runcore/profiling.c:
[profiling] fix filename reporting in profiling output
rrot: r42902 | chromatic++ | trunk (2 files):
[PCC] Made boxed incoming parameters respect HLL map types (reported by Patrick
pmichaud that does appear to fix things... thanks. 22:49
chromatic You're welcome. 22:56
pmichaud ugh, I thought TT #562 had been fixed by now. 22:57
plumage? 23:11
purl somebody said plumage was the future Parrot module ecosystem. It will include tools to search metadata, handle dependencies, install modules, and so forth. The repository is at gitorious.org/parrot-plumage/parrot-plumage and the design docs are at trac.parrot.org/parrot/wiki/ModuleEcosystem
japhb *pop*
yes ...?
purl Yada yada yada hasn't been implemented yet! (unless you run bleadperl)
pmichaud needing to look up some information from Glue 23:12
Coke ugh. I can implement ==, but I need a working "toInteger" first. bah. 23:13
23:13 mikehh joined 23:18 lucian joined, brrant joined 23:24 lucian joined 23:35 Zak joined 23:53 cconstantine joined
mikehh All tests PASS (pre/post-config, smoke (#30624), fulltest) at r42902 - Ubuntu 9.10 amd64 (g++ with --optimize) 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!