|
Parrot 3.9.0 "Archaeopteryx" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC Set by moderator on 19 October 2011. |
|||
|
00:39
benabik joined
00:40
whiteknight joined
00:55
diakopter joined
|
|||
| diakopter | from nqp, I'm trying to print to stderr using a pir:: call; pir::printerr__vS isn't working; any tips? | 00:56 | |
| bacek_at_work | diakopter, try something like 'my $e := pir::getstderr__p(); pir::print__vPS($e, "message")' | 01:01 | |
| diakopter, printerr is dynop and you have to load io_ops beforehand | |||
| diakopter | ooo | ||
|
02:19
jsut joined
|
|||
| dukeleto | ~~ | 02:51 | |
|
02:54
Themeruta joined
02:59
woosley joined
03:05
nbrown joined
|
|||
| alvis | dukeleto: ping | 03:35 | |
| dukeleto | alvis: pong | 04:11 | |
| alvis | oh, hey! Just had a coupl'a questions. | ||
| dukeleto | alvis: go for it | 04:12 | |
| alvis | Sorry ... give me just a sec please. | ||
| ok ... back. | 04:13 | ||
| my question is, should a create a separate branch for the docs? Or just keep it all in '/docs'? | 04:14 | ||
| I'm not sure I understand precisely what benabik meant with, keeping the docs with the code, so I'm gonna email 'im, if he's not on channel now. | 04:15 | ||
| make that, "i create a ..." | 04:16 | ||
| dukeleto | alvis: there are a lot of docs. which are you talking about? | 04:20 | |
| alvis: are you writing new docs? changing old docs? | |||
| alvis | sorry to be unclear. primarily, something like the manuals. | 04:21 | |
| dukeleto | alvis: yes, make a branch :) | ||
| alvis: make a branch for everything | |||
| alvis: in your local repo: git checkout -b my_docs_branch | |||
| alvis | ahh, ok. that makes sense to me as well. | ||
| dukeleto | alvis: will create my_docs_branch | ||
| alvis | right. i just wasn't sure about the "keep the docs close to the code" statement. | 04:22 | |
| next, what is the policy on taking over an old, defunct project? Specifically, Kea-CL. | 04:23 | ||
| I've sent Bob Rogers a coupl'a emails to no avail, so .... | |||
| I'd really like to see it cleaned up and running on Parrot. It's a long-term kinda thing, but .... | 04:24 | ||
| Just to be clear: I'm NOT asking to make it a part of the parrot repository. I'm just asking as a general matter. | 04:43 | ||
| Don't want to violate any policies or rules of which I am unaware. | 04:44 | ||
|
04:47
janus joined
|
|||
| dukeleto | alvis: i would say, grab the kea-cl code and start hacking on it | 05:45 | |
| alvis: nobody has touched it in a long time | |||
| alvis | dukeleto: Ok, thanks. Just didn't want to break any policies or violate some unwritten rule of which I am unaware. | 05:47 | |
| dukeleto: Besides, he did a lot of very, very good work. I hate to see it lost. | |||
| dukeleto: one last thing (then i'm done for the evening): I assigned myself to migrate docs.parrot.org to Github pages. | 05:49 | ||
| dukeleto | alvis: awesome! good luck reviving kea-cl! document the process so others can learn from it :) | 06:13 | |
| dukeleto .sleep() | |||
| cotto | ~~ | 06:27 | |
|
07:45
diakopter left
08:19
mj41 joined
08:23
mj41 joined
09:33
lucian joined
|
|||
| dalek | kudo/nom: 11e836e | moritz++ | src/Perl6/SymbolTable.pm: use more nqp:: ops in SymbolTable |
11:17 | |
|
11:38
Psyche^ joined
12:09
Psyche^ joined
12:14
Psyche^_ joined,
bluescreen_ joined
12:18
whiteknight joined
|
|||
| whiteknight | good morning, #parrot | 12:20 | |
| moritz | good morning whiteknight | 12:27 | |
|
12:28
particle joined
|
|||
| Coke | msg alvis "keep the docs near the code" is shorthand for "don't create separate documentation files in a different location from the code you're documenting." for tutorials this doesn't make sense. for APIs it makes some sense. (so, the .ops files might want to keep their docs inline, as now, and have a build step that extracts it (as now)) | 12:32 | |
| aloha | OK. I'll deliver the message. | ||
| moritz | also I think it means "don't put it into a separate repository", so that the pained developer only needs to keep one repo up-to-date | 12:37 | |
|
12:43
bluescreen__ joined
|
|||
| dalek | Heuristic branch merge: pushed 20 commits to nqp/bigint by jnthn | 12:48 | |
| p/bigint: 60d70e2 | jonathan++ | src/ops/nqp_bigint.ops: Update bigint ops to work with new repr API. |
|||
| p/bigint: e388430 | jonathan++ | src/ (3 files): Split bigint repr struct into instance and body. Split instance_of into allocate and initialize to match new API. |
13:03 | ||
| p/bigint: f4f9287 | jonathan++ | src/6model/reprs/P6bigint.c: Toss defined, which is gone from the repr API. Fix up attribute related repr methods. |
|||
|
13:07
Coke joined
|
|||
| dalek | p/bigint: b0f876a | jonathan++ | src/6model/reprs/P6bigint.c: Implement copy_to, get rid of repr_clone. |
13:09 | |
|
13:11
pbaker joined
|
|||
| dalek | p/bigint: 1eea4e5 | jonathan++ | src/6model/reprs/P6bigint.c: Update get/set native repr functions to new API. |
13:15 | |
| p/bigint: 324df8e | jonathan++ | src/6model/reprs/P6bigint.c: Update GC related bit of the bigint repr. Builds again. |
13:23 | ||
| p/bigint: d387ce9 | moritz++ | src/6model/reprs/P6bigint.c: initialize return value in P6bigint -> get_int |
13:33 | ||
| p/bigint: f9262fc | jonathan++ | t/nqp/60-bigint.t: Avoid a curious, seemingly new issue by removing a needless .HOW (what get_knowhow gives is already a HOW). |
13:52 | ||
|
13:57
woosley left
|
|||
| dalek | rrot/mls/kill-events-in-ehqueue: 5e2019c | mls++ | / (4 files): create Parrot_cx_delete_upto_handler_local() function for better abstraction |
14:04 | |
| rrot/mls/kill-events-in-ehqueue: 25e580f | mls++ | / (3 files): Define init_pmc for ExceptionHandlers that takes a key to initialize the handled types. Currently not optimized, we still build the RPA. Also change PCT to create the exception handlers that way. |
|||
|
14:05
nbrown joined
|
|||
| dalek | p/bigint: 2163a93 | jonathan++ | src/6model/reprs/P6bigint.c: Fix get_int/set_int in P6bigint. |
14:22 | |
| p/bigint: ebffd93 | jonathan++ | tools/build/Makefile.in: Make the right thing depend on the bigint repr. |
14:36 | ||
|
14:40
lateau joined
|
|||
| dalek | rrot/mls/kill-events-in-ehqueue: 1c313a0 | mls++ | src/ (2 files): Directly use Key as handles_type when possible. |
15:04 | |
|
15:06
lateau1 joined
|
|||
| NotFound | mls: ping | 15:21 | |
| mls | you rang? | ||
| NotFound | mls: in 25e580faf7 you are making changes in ExceptionHandler unrelated to the main issue. | 15:22 | |
| mls | what's the "main issue"? | 15:23 | |
| NotFound | "Define init_pmc for ExceptionHandlers ..." | ||
| mls | oh, you're right, some optimization crept in | 15:24 | |
| or do you mean the pct change? | |||
| NotFound | The changes in can_handle | ||
| It will be better to have these in separated commits. Or even better, out of the branch. | 15:25 | ||
| mls | yes, That should probably have gone to master instead | ||
| NotFound | Not a big problem, but is better for testing and eventual regresions. | 15:26 | |
| moritz | fwiw yesterday all rakudo tests passed on top of mls/kill-events-in-ehqueue | ||
| mls | (or a different commit "opimize some keyed_str lookups away if the base_type matches") | ||
| moritz: NotFound++ has a point: those changes shouldn't be one commit | 15:27 | ||
| NotFound likes micro-commits | |||
| mls | I can split the commit and use 'push -f' | 15:28 | |
| Coke | I think on a branch that's probably acceptable. | ||
| but IANAGL. | 15:29 | ||
| mls | ok, will do | ||
| (I hope poor dalek can cope) | |||
| Coke | I'd get buyin from at least one other person before doing that. ;) | 15:30 | |
| moritz | mls: I know, the only connection was the branch name, which made me remember that I ran that test yesterday | ||
| NotFound | People advocated the switch because is much easy to branch and merge and thus less need to make lots of unrelated changes in One Big Branch. We switched, so let's profit from that point. | 15:36 | |
| (the switch to git) | 15:37 | ||
| Coke | NotFound: I'm specifically asking about mls rewriting publicly pushed history. | ||
| I agree in general that more isolated commits are better for review, etc. | 15:38 | ||
| NotFound | Rewriting history in public repos is bad, according the few docs I've read on that matter. | 15:40 | |
| dalek | rrot/mls/kill-events-in-ehqueue: a38bbec | mls++ | src/ (2 files): Optimize some keyed_str lookups and a isa() call away when the base_type matches |
||
| rrot/mls/kill-events-in-ehqueue: da7c786 | mls++ | / (2 files): Define init_pmc for ExceptionHandlers that takes a key to initialize the handled types. Currently not optimized, we still build the RPA. Also change PCT to create the exception handlers the new way if possible. |
|||
| mls | oh oh, force -f does strange things | ||
| rrot/mls/kill-events-in-ehqueue: 0d94c86 | mls++ | src/ (2 files): Directly use Key as handles_type when possible. |
|||
| mls | It also updated master ;( | ||
| Coke | mls: what happened to "don't do that yet!?" | ||
| moritz | there are several approaches: 1) start a new branch 2) simply revert the commit in the branch, and add it back again in pieces | ||
| moritz can't think of more approaches actually | 15:42 | ||
| Coke | mls: I don't see any updates to master today on github. | 15:45 | |
| dalek | rrot: b1dbbad | nine++ | t/pmc/alarm.t: Remove faulty tests from t/pmc/alarm.t The removed tests tested for alarms firing in a certain order which cannot be guaranteed due to preemption introducing a random element into the mix. |
||
| rrot: e20b3be | nine++ | / (4 files): Merge branch 'master' of github.com:parrot/parrot |
|||
| mls | more karma for nine ;) | ||
| good thing that git is so forgiving | 15:50 | ||
| (nothing happened, please move on ;) ) | |||
| NotFound | And one more nitpick: that new init_pmc wants a test. | 15:51 | |
| mls | yes, true. I'll add one. | ||
| whiteknight | if it's good, we will get it a test for christmas | ||
| mls | so, how do we specify "handle_types_except" | 15:52 | |
| NotFound | ok(tree_has_bells() && santa_had_come()) | ||
| This is a good test for christmas? | |||
| mls | we could use negative values, but imcc doesn't evaluate constant expressions | 15:53 | |
| so we probably would need EXCEPT_EXCEPTION_... = -<some_value> macors | 15:54 | ||
| i.e. a "negative" one for every exception type | |||
| or we can define one EXCEPT_TYPES value | 15:55 | ||
| and put that in front of the exception types | |||
| Btw, what do you goys think of moving the CONTROL types to something like 0x1000 and defining an EXCEPTION_ALL and CONTROL_ALL type? | 15:56 | ||
| NotFound | You said you aren't afraid of imcc hacking? Add constant evaluation to it, then ;) | ||
| mls | that's a bigger task than the one we talked about yesterday ;) | 15:57 | |
|
15:57
not_gerd joined
|
|||
| not_gerd | hello, #parrot | 15:57 | |
| whiteknight | mls: +1 to the rearranging the exception type numbers. It's a relatively small change | 15:58 | |
| mls | what about the "ALL" types? | ||
| whiteknight | mls: Eventually, I would like to stop using exception type numbers and use actual exception classes, but that's a different story | ||
| mls: +1 to the ALL types too | |||
| NotFound | Note that the pir limitation does not make such feature unusable. Winxed can do constant evaluation with pir-included macro consts. | 15:59 | |
| mls | yes, and I can define one EXCEPT_CONTROL_ALL, that's actually all nqp/rakudo needs | 16:01 | |
| you you prefer negative numbers as "except" values? | |||
| so you... | |||
| NotFound | It was just a quick idea, not a particualar preference. | 16:02 | |
| mls | (hmm, imcc doesn't even understand a 0x prefix) | 16:04 | |
| dalek | kudo/nom: 79c1c9f | moritz++ | Configure.pl: [Configure.pl] fix import list, not_gerd++ |
||
| NotFound | Sometimes I forget pir limitations because I usually check pir-level functionality from winxed. | 16:07 | |
| not_gerd | some fixes for msys: github.com/parrot/parrot/pull/192 | 16:12 | |
| whiteknight | not_gerd: I saw those yesterday. I think they got pulled already | 16:14 | |
| or, something happened. I don't remember | |||
| not_gerd | whiteknight: the pull request yesterday only disabled select to make it build again | ||
| dukeleto closed the request, but didn't apply it (or mention why it was closed) | 16:15 | ||
| whiteknight | oh, another pull request. Sorry. I'll take a look at it later, if nobody beats me to it | 16:21 | |
| dalek | rrot/mls/kill-events-in-ehqueue: 461e486 | mls++ | / (3 files): move control exceptions to their own number block, add _ALL types that catch any exception of that type |
16:29 | |
| rrot/mls/kill-events-in-ehqueue: cfab58d | mls++ | compilers/pct/src/PAST/Compiler.pir: use a key instead of a string for the return exception handler, as the key is cheaper |
16:36 | ||
| mls | lets see if nqp/rakudo still work | ||
|
16:48
pbaker joined
16:59
contingencyplan joined
|
|||
| dalek | rrot/mls/kill-events-in-ehqueue: 4e91e6c | mls++ | src/pmc/exceptionhandler.pmc: it has to be pmc/exception.h, otherwise the Configure script will generate a wrong Makefile dependency |
17:00 | |
|
17:16
alvis joined
|
|||
| dalek | rrot/mls/kill-events-in-ehqueue: 74ba753 | mls++ | compilers/pct/src/PAST/Compiler.pir: Argh, 'CONTROL' should not include CONTROL_ERROR, used in die()! |
17:29 | |
| mls | aaah, it hurts! it hurts! | 17:30 | |
| whiteknight | ...wat? | 17:31 | |
| mls | I'm tempted to move CONTROL_ERROR to the exception block, despite the name | ||
| or maybe add an EXCEPTION_ERROR used by die() | 17:32 | ||
| whiteknight | I like EXCEPTION_ERROR | ||
| or even something better like EXCEPTION_DIE | |||
| mls | yes, that would make sense | ||
| the current situation is that nqp's and rakudos CONTROL {} catch every control exception but one, CONTROL_ERROR | 17:33 | ||
| oh, actually CONTROL_ERROR can be used by warn(), which is supposed to be caught by a CONTROL catch block | 17:34 | ||
| (I currently use CONTROL_OK for that purpose) | 17:35 | ||
| whiteknight | bleh | 17:38 | |
| our exceptions system needs to be taken out back and shot | 17:39 | ||
| tadzik | listening to what #parrot talks about, is there anything that shouldn't? | 17:40 | |
| gms is cool I suppose | |||
| mls | actually I like pct | ||
| parrot's ops are pretty fast | 17:41 | ||
| hmm, what else... ;) | 17:42 | ||
| Coke | rakudo's pretty cool. | 17:49 | |
| whiteknight | tadzik: I overemphasize some problems. Exceptions are a particularly bad system with clear performance repercussions for Rakudo and other users | 17:52 | |
| it doesn't help that, for all the performance penalties, the interface is lousy, the capabilities are limited and non-intuitive, the usability is low, etc | 17:53 | ||
| Coke | and that it was designed either not against perl 6 or another HLL, or against an old version of the perl6 spec | 17:54 | |
| mls | that "resume" functionality makes it so tricky | 17:55 | |
| it's different from other VMs | |||
| Coke | just be sure to get buyin from moritz on a redesign, is all. ; | 17:58 | |
|
18:01
fperrad joined
|
|||
| whiteknight | Coke: No, you're right. It was clearly not designed with Perl6 in mind | 18:06 | |
| dalek | rrot/mls/kill-events-in-ehqueue: a3ec104 | mls++ | src/pmc/ (3 files): do not use CONTROL_ERROR for internal exceptions, now only die() uses it |
||
| mls | enough for this week, you guys have a nice weekend! | 18:08 | |
| whiteknight | you too, mls! | 18:10 | |
| cotto | ~~ | 18:11 | |
| dukeleto | ~~ | 18:32 | |
|
18:32
mj41 joined
|
|||
| dalek | rdinal: 27f7162 | (Daehyub Kim)++ | setup.winxed: issue #18 * clean * test * smoke: need to modify * help * install: do not commit in this changes |
18:34 | |
| rdinal: 602e40f | (Daehyub Kim)++ | setup.winxed: Merge branch 'setup_winxed' of github.com:lateau/cardinal into setup_nqp |
|||
| rdinal: 4841027 | (Daehyub Kim)++ | setup.winxed: target install added |
|||
| rdinal: e393cd3 | (Daehyub Kim)++ | setup.winxed: remove useless codes |
|||
| rdinal: 32929b6 | (Daehyub Kim)++ | setup.winxed: setup.winxed all targets of setup.pir are supported and works now: * build * test * install * uninstall * clean * update * plumage * sdist * bdist * smoke * help changed: * clean: remove rpm directory and it's subdirectories also * build: build target will be runed before sdist and bdist Thanks NotFound++ |
|||
| rdinal: e8116ba | (Daehyub Kim)++ | setup.winxed: ignore pod with /* */ |
|||
| rdinal: 97588e0 | dukeleto++ | setup.winxed: Merge pull request #19 from lateau/setup_winxed Setup winxed |
|||
| rrot: 543e698 | (Gerhard R)++ | / (2 files): disable select dynpmc on MSYS |
18:38 | ||
| rrot: c0a28a9 | (Gerhard R)++ | / (2 files): fix make install on MSYS |
|||
| rrot: df69781 | (Gerhard R)++ | / (4 files): cleanup compiler flags, file generation, ignored files |
|||
| rrot: 8fde623 | dukeleto++ | / (5 files): Merge pull request #192 from gerdr/gerdr/msys-install fix make install on MSYS |
|||
| alvis | Coke, mortiz: Thank you for the clarification. | 18:41 | |
| dukeleto | if anybody on msys can test that the stuff that just got merged works, that would be awesome | 18:42 | |
| whiteknight | not_gerd is our only user/tester on that platform, as far as I know | 18:43 | |
| not_gerd | dukeleto: make install on MSYS was broken-- I'm the only one who uses it on that platform and I just used parrot directly from the build directory | 18:50 | |
| the new patch makes rakudo's --gen-parrot work - I don't guarantee anything beyond that... | 18:51 | ||
| btw, t/pmc/alarm.t hangs here - does it work on other windows configurations? | 18:59 | ||
| whiteknight | not_gerd: it doesn't hang on other systems, but alarms might be disabled there. We should probably disable that test file on msys for now | 19:14 | |
|
19:23
soh_cah_toa joined
19:45
fperrad joined
20:16
mj41 joined
20:17
perlite_ joined
20:32
davidfetter joined
20:46
ambs joined
|
|||
| dukeleto drools over www.genesi-usa.com/store/details/11 | 21:56 | ||
|
22:01
AzureStone joined
22:03
perlite joined,
dod1 joined
23:01
arnsholt_ joined,
AzureSto_ joined
23:03
wagle_ joined,
snarkyboojum_ joined,
dod joined
23:05
betterworld joined
23:07
cxreg joined,
Felipe joined,
athomason joined,
TiMBuS joined,
janus joined,
nnunley joined,
plobsing_ joined,
nopaste joined,
TonyC joined,
dngor joined,
simcop2387 joined,
PerlJam joined,
awwaiid joined,
Infinoid joined,
nine joined,
moritz joined,
sorear joined,
nbrown joined,
perlite joined,
pbaker1 joined,
Coke joined,
Patterner joined,
NotFound joined,
atrodo joined,
jlaire joined,
dukeleto joined,
TimToady joined,
pjcj joined,
ttbot joined,
Tene joined,
kthakore joined,
sri joined,
cotto joined,
preflex joined,
mj41 joined,
soh_cah_toa joined,
japhb joined,
ascent joined,
autark joined,
mls joined,
szbalint joined,
ingy joined,
allison joined,
man_in_shack joined,
aloha joined,
Khisanth joined,
tadzik joined,
p6eval joined,
rblackwe joined,
slavorg joined
23:20
benabik joined
23:26
nopaste joined
23:44
rfw joined
|
|||