Camelia, the Perl 6 bug

IRC log for #parrot, 2009-04-10

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 jonathan But anyone who cares about the length of the UUID can just go grab the string and then get it's length.
00:01 bacek indeed.
00:04 dalek parrot: r37999 | bacek++ | branches/packfile_revamp/do​cs/pdds/pdd13_bytecode.pod:
00:04 dalek parrot: Remove uuid_length from public Packfile attributes.
00:04 dalek parrot: "But anyone who cares about the length of the UUID can just go grab the
00:04 dalek parrot: string and then get it's length." (c) jonathan
00:04 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37999/
00:04 jonathan oh noes responsibility!
00:09 AndyA joined #parrot
00:09 * bacek laughing evilly
00:11 * Coke_afk hears how many slots perl gets for GSOC, preliminarily.
00:11 jonathan Coke: Hopefully > 0?
00:13 kid51 joined #parrot
00:14 Coke I odn't know if it's secret, but yes. =-)
00:15 Coke subject to change until monday.
00:15 Coke Hey! Chromium got slots. that's cheating!
00:19 Infinoid Android probably did too
00:19 Coke TPF got a preliminary allocation of 10 slots. subject to change. I don't see anything that says it's a secret.
00:19 Coke no android.
00:19 Infinoid aaw
00:20 Infinoid is the parrot foundation applying separately, or are we begging for some slots from TPF?
00:21 Infinoid (the fact that they both have the same acronym is a bit unwieldy)
00:21 Coke damint.
00:22 Coke "I was just asked an excellent question on IRC - is it OK to share
00:22 Coke these numbers with our community?
00:22 Coke The answer is that is up to you. If it were me deciding, I would not
00:22 Coke " -LH
00:22 Coke ah well. =-)
00:22 Coke so, I stress again this is PRELIMINARY.
00:22 Coke Infinoid: pafo applied under the TPF umbrella.
00:24 Coke "nothing will be final until monday"
00:31 tetragon joined #parrot
00:35 jonathan bacek: Taking sleep now, catch me tomorrow if you have any more packfile questions :-)
00:35 jonathan night all
00:35 bacek good night jonathan
00:35 bacek I'll probably will have more packfile answers :)
00:35 bacek ETOOMANYWILLS
00:48 s1n joined #parrot
01:20 Tene joined #parrot
01:59 eternaleye joined #parrot
01:59 acajou joined #parrot
02:06 dalek parrot: r38000 | cotto++ | trunk/src/pmc/integer.pmc:
02:06 dalek parrot: [PMC] change Integer's i_add to a VTABLE function where possible
02:06 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38000/
02:07 jdv79 left #parrot
02:09 Infinoid joined #parrot
02:48 janus joined #parrot
02:52 petdance joined #parrot
03:26 joubert joined #parrot
03:51 amoc joined #parrot
03:54 dalek parrot: r38001 | coke++ | trunk/t/codingstd/c_function_docs.t:
03:54 dalek parrot: [t/docs] un TODO passing test
03:54 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38001/
04:04 * Coke ponders 'make depend'
04:11 TiMBuS joined #parrot
04:12 GeJ what would that do?
04:15 Coke automatimcally updated dependency information extractable from pmcs, c, ops, perl, pir, and pasm.
04:15 Coke so we don't have to actively manage it anymore.
04:54 dalek parrot: r38002 | coke++ | trunk/src/oo.c:
04:54 dalek parrot: [cage] remove unreachble code
04:54 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38002/
05:00 dalek joined #parrot
05:01 japhb joined #parrot
05:05 Infinoid bacek: ping
05:05 bacek Infinoid: pong
05:06 bacek erm. Good morning?
05:06 Infinoid Any objections to me making Socket.get_bool() return openness rather than createdness?
05:06 Infinoid it's 10pm here :)
05:07 bacek Infinoid: no objections.
05:07 Infinoid I want to rip the guts out of it, reverse the logic and use it to implement Parrot_io_socket_is_closed() (TT #547)
05:07 Infinoid so get_bool will just become a negating wrapper around that
05:07 bacek +1
05:07 purl 1
05:07 Coke isn't that a functional change?
05:08 Coke (and therefore requires a dep-cycle?)
05:08 bacek Anyway, Socket.socket returns PMCNULL in case of error
05:08 Infinoid Coke: No, all this stuff was added post-1.0.0.
05:08 bacek Coke: it's between releases :)
05:09 Infinoid Anyway, get_bool will still return exactly the same thing... the "createdness" thing I mentioned above was what the documentation said it did, which isn't actually what it did.
05:09 * Infinoid is updating the doc too
05:09 Coke docs win.
05:09 Coke (then you're fixing a bug. whee.)
05:09 bacek How I can run single step from Configure?
05:09 Coke bacek: with great pain.
05:10 bacek Coke: thank you :)
05:10 Coke holy crud is 'make splint' noisy.
05:11 bacek Looks like &gen_from_def can't handle non-int defines...
05:14 dalek parrot: r38003 | Infinoid++ | trunk/src (2 files):
05:14 dalek parrot: [io] Implement Parrot_io_socket_is_closed().
05:14 dalek parrot: The Socket "get_bool" function (which is where I ripped this functionality from) had documentation that wasn't quite accurate; fix that too.
05:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38003/
05:15 Infinoid Damn, didn't quite get the ticket closed within half an hour of its creation. :)
05:18 Infinoid *sigh* Shoulda stuck to my rule, never commit anything after 10pm
05:22 Coke-ZZZ Infinoid+++
05:24 dalek parrot: r38004 | Infinoid++ | trunk (2 files):
05:24 dalek parrot: [io] Build fixes from previous botched patch.  Infinoid--
05:24 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38004/
05:29 bacek *incoming*
05:29 * bacek hides under table
05:30 dalek parrot: r38005 | bacek++ | branches/packfile_revamp/src/pmc/packfile.pmc:
05:31 dalek parrot: Start reimplementing Packfile PMCs to generate underlying PackFile instead of using it directly.
05:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38005/
05:31 dalek parrot: r38006 | bacek++ | branches/packfile_revamp/t/pmc/packfile.t:
05:31 dalek parrot: Restructure t/pmc/packfile.t.
05:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38006/
05:31 dalek parrot: r38007 | bacek++ | branches/packfile_revamp/t/pmc/packfile.t:
05:31 dalek parrot: Add tests for Packfile unpacking.
05:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38007/
05:31 dalek parrot: r38008 | bacek++ | branches/packfile_revamp/src/pmc/packfile.pmc:
05:31 dalek parrot: Really unpack PackFile and set internal attributes.
05:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38008/
05:31 dalek parrot: r38009 | bacek++ | branches/packfile_revamp/src​/pmc/packfiledirectory.pmc:
05:31 dalek parrot: Refactor PackfileDirectory PMC to use Hash as main storage.
05:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38009/
05:31 dalek parrot: r38010 | bacek++ | branches/packfile_revamp/t​/pmc/packfiledirectory.t:
05:31 dalek parrot: Update packfiledirectory.t to current behaviour.
05:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38010/
05:33 * bacek moves deeper under table
05:33 Infinoid go git-svn go
05:33 Infinoid (small, easily reviewable patches)++
05:34 bacek Infinoid: It's not actual history of commits :)
05:34 dalek parrot: r38011 | bacek++ | branches/packfile_revamp/t​/pmc/packfiledirectory.t:
05:34 Infinoid are you using stgit and constantly revisioning things, like I do?
05:34 dalek parrot: Remove unused test
05:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38011/
05:34 dalek parrot: r38012 | bacek++ | branches/packfile_revamp/do​cs/pdds/pdd13_bytecode.pod:
05:34 dalek parrot: Remove int-based accessors from PackfileDirectory. Add get_iter method.
05:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38012/
05:34 dalek parrot: r38013 | bacek++ | branches/packfile_revamp/t/p​mc/packfileconstanttable.t:
05:34 dalek parrot: Update packfileconstanttable.t to properly search for ConstantTable.
05:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38013/
05:34 dalek parrot: r38014 | bacek++ | branches/packfile_revamp/src/p​mc/packfileconstanttable.pmc:
05:34 dalek parrot: Refactor PackfileConstantTable to use RPA and RIA for store data.
05:34 bacek No, just git rebase --interactive
05:35 dalek parrot: Filling from real PackFile_ConstTable coming soon.
05:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38014/
05:35 dalek parrot: r38015 | bacek++ | branches/packfile_revamp/src/p​mc/packfileconstanttable.pmc:
05:35 dalek parrot: Implement loading of PackFile_ConstTable into PackfileConstantTable PMC.
05:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38015/
05:35 dalek parrot: r38016 | bacek++ | branches/packfile_revamp/src/pmc/packfile.pmc:
05:35 dalek parrot: Invoke PackfileDirectory.set_pointer to instantiate all other PMCs.
05:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38016/
05:35 dalek parrot: r38017 | bacek++ | branches/packfile_revamp/src/p​mc/packfileconstanttable.pmc:
05:35 dalek parrot: Remove unused code.
05:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38017/
05:35 dalek parrot: r38018 | bacek++ | branches/packfile_revamp (2 files):
05:35 dalek parrot: Generate PASM constants for PackFile_Constant types.
05:35 dalek parrot: Because of limitations of gen::parrot_include switch to use
05:35 dalek parrot: integers for PFC_* constant types.
05:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38018/
05:35 dalek parrot: r38019 | bacek++ | branches/packfile_revamp/src​/pmc/packfilerawsegment.pmc:
05:35 dalek parrot: Reimplement PackfileRawSegment to use own stored data.
05:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38019/
05:35 bacek One or two commits more...
05:35 Infinoid nice, I've never tried that.
05:35 * Infinoid tends to split patches the hard way
05:36 bacek Ah. It was last commit.
05:37 bacek Infinoid: I commit every finished bit. And than just squash commits into single logical one.
05:37 bacek (Not always unfortunately...)
05:38 Infinoid Me too, I just keep it all in stgit
05:38 Infinoid For some reason it fits my brain patterns better than raw git ever did.
05:40 bacek I never tried stgit. But TMWTDI works perfectly in git :)
05:41 * Infinoid notes for the logs that he's now maintaining git://squawk.glines.org/parrot-all/ in addition to git://squawk.glines.org/parrot-trunk/
05:41 Infinoid Yes.  But third party tools don't always work perfectly together... for instance git-svn confuses stgit and vice-versa
05:42 Infinoid good thing I don't run both of them within the same checkout :)
05:42 bacek :)
05:48 bacek git-svn still hangs with "Resetting to the latest refs/remotes/packfile_revamp"...
05:48 Theory joined #parrot
05:51 cotto bacek
05:51 bacek cotto:
05:51 cotto bacek++ #packfile hackery
05:51 bacek :)
05:51 cotto my fingers don't look fat, but the evidence says otherwise
05:52 bacek I think it was cat which jumped on keyboard
05:52 cotto ENOCAT
05:53 bacek don't tell anyone :)
05:58 dalek parrot: r38020 | bacek++ | branches/packfile_revamp/t/p​mc/packfileconstanttable.t:
05:58 dalek parrot: Use packfile_constants.pasm instead of "magik constants" in packfileconstanttable.t.
05:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38020/
05:58 dalek parrot: r38021 | bacek++ | branches/packfile_revamp/sr​c/pmc/packfilesegment.pmc:
05:58 dalek parrot: Add PackfileSegment.set_pointer.
05:58 dalek parrot: Do nothing for now except throwing exception.
05:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38021/
06:06 dalek rakudo: 3280d41 | pmichaud++ | build/gen_whatever_pir.pl:
06:06 dalek rakudo: Whatever arguments appear to have been flipped.
06:06 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/3​280d410372138e8f9e7a3f21b06fd9e78ed1848
06:06 shorten dalek's url is at http://xrl.us/beog4y
06:08 cotto The use of "whatever" as a technical term never ceases to amuse me.
06:11 dalek parrot: r38022 | bacek++ | branches/packfile_revamp/sr​c/pmc/packfilesegment.pmc:
06:11 dalek parrot: Remove throwing exception from PackfileSegment.set_pointer. bacek--
06:11 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38022/
06:11 dalek parrot: r38023 | bacek++ | branches/packfile_revamp/t/pmc (4 files):
06:11 dalek parrot: Move common test functions into packfile_common.pir
06:11 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38023/
06:12 bacek ok, little brake and start packing :)
06:17 dalek rakudo: 712982a | pmichaud++ | docs/spectest-progress.csv:
06:17 dalek rakudo: spectest-progress.csv update: 356 files, 10284 passing, 0 failing
06:17 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​12982aa848ded75be689dbd1d86106a88d2384d
06:17 shorten dalek's url is at http://xrl.us/beog5g
06:47 silug joined #parrot
08:08 barney joined #parrot
08:12 iblechbot joined #parrot
08:23 mikehh joined #parrot
09:09 rob joined #parrot
09:18 bacek Infinoid: around?
09:25 uniejo joined #parrot
10:00 schmalbe joined #parrot
10:23 cotto anyone here conscious who knows mmd?
10:28 bacek mind melting dispatch?
10:28 cotto sounds accurate
10:30 cotto bah.  sleep
10:30 cotto night
10:32 bacek g'night
11:02 Whiteknight joined #parrot
12:00 acajou Is Rakudo supposed to pass a spectest cleanly right now?
12:01 * acajou gets 105 failed tests...
12:04 jonathan acajou: What are you seeing failing?
12:04 jonathan I thought it passed cleanly at the moment.
12:04 acajou http://perl.pastebin.com/d4d917b6f
12:05 acajou Tried several times yesterday and this morning.  Never clean, but a lot of source updates too...
12:06 acajou Using git fetch / perl Configure.pl --gen-parrot / make / make spectest
12:07 * acajou afk
12:15 jonathan Odd. :-|
12:15 jonathan I guess try a make realclean first too...
12:15 jonathan But I'm not seeing any failings like that...
12:17 rafl joined #parrot
12:26 acajou OK, I'll try that today, thanks.
12:26 * acajou waves.
12:26 acajou left #parrot
12:44 silug joined #parrot
13:08 Infinoid bacek: I am now
13:13 mikehh I am getting failures in codetest at r38023 - t/codingstd/copyright.t - Duplicate copyright statement found in 12 files:
13:14 mikehh I had a look at the files and I am not sure if we can remove the duplicates
13:14 Infinoid That's a work in progress
13:14 joubert joined #parrot
13:15 mikehh The imcc stuff has Copyright (C) 2002 Melvin Smith <melvin.smith@mindspring.com> then Copyright (C) 2003-2009, Parrot Foundation
13:16 mikehh others are found in prunt statements and POD
13:17 mikehh s/prunt/print/
13:18 mikehh some are relayed to GNU copyright and things like that
13:18 mikehh s/relayed/related/
13:20 mikehh for examp;e in Configure.pl at line 31 - # Copyright (C) 2003 Free Software Foundation, Inc.
13:21 Infinoid Yeah.  From TT #543, it looks like the test was untodoed accidentally... but it's still a work in progress
13:23 mikehh That I can see - cotto at r37939 - maybe the test should only test the first couple of lines of the file - not further?
13:24 Infinoid that sounds reasonable
13:46 gryphon joined #parrot
13:52 Tene joined #parrot
14:13 Coke-ZZZ mikehh: no.
14:14 Coke right now there's a bug in the test that it acts differently in 5.8 vs. 5.10
14:32 Infinoid Coke, are there any cases where the test should be looking for copyright tags somewhere other than, say, the first 5 lines of the file?
14:32 Coke the cases where the copyright is at the end of the file? =-)
14:32 Infinoid Fair enough. :)
14:32 Coke I'm honestly curious why it behaves diffeently.
14:33 Coke I think a passing todo is better than a failing, though.
14:33 Infinoid Yeah, the portability issue should be fixed too.  I just thought mikehh's idea was an interesting one
14:59 riffraff joined #parrot
15:05 Theory joined #parrot
15:06 tetragon joined #parrot
15:16 joubert joined #parrot
15:36 bacek oh... 1:30AM here...
15:36 * bacek hates pure C even more...
15:37 bacek Infinoid: good... erm... evening?
15:37 Coke Infinoid: Parrot_str_append - should *a be ARGIN ?
15:38 Infinoid bacek: good morning :)
15:39 bacek I'm not sure is it already morning or still night...
15:39 dalek parrot: r38024 | bacek++ | branches/packfile_revamp/sr​c/pmc/packfilesegment.pmc:
15:39 bacek here we go
15:39 dalek parrot: Strip PackfileSegment to bare minimum.
15:39 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38024/
15:39 dalek parrot: r38025 | bacek++ | branches/packfile_revamp/src/p​mc/packfileconstanttable.pmc:
15:39 dalek parrot: Reindent code in packfileconstanttable.pmc
15:39 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38025/
15:39 dalek parrot: r38026 | bacek++ | branches/packfile_revamp (9 files):
15:39 dalek parrot: Initial implementation of writable Packfile PMCs.
15:39 dalek parrot: Create PackFile on fly, pack it, destroy it.
15:39 dalek parrot: Every Packfile*.pmc responsible for creating "native" PackFile segment
15:39 bacek Incoming :)
15:39 dalek parrot: in VTABLE get_pointer. Those pointers gathered by PackfileDirectory and
15:39 dalek parrot: added into PackFile_Directory for serialization and memory management.
15:39 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38026/
15:39 bacek Infinoid: r38026
15:40 bacek 10 hours of fight with packfiles...
15:40 Infinoid Coke: doesn't it append to *a?
15:40 Infinoid bacek: I'm in TZ=PST8PDT :)
15:41 bacek Infinoid: ah. Yesterday morning was beautiful.  I still remember :)
15:42 Coke Infinoid: the docs say "don't use the one you pass in, we return a value, use that.
15:42 Coke splint complains.
15:42 Coke "Fresh storage result (type STRING *) not released before assignment:"
15:43 Infinoid The docs are wrong, then.
15:43 Infinoid a->bufused += b->bufused;
15:43 Infinoid a->strlen  += b_len;
15:43 Infinoid a->hashval  = 0;
15:43 Infinoid return a;
15:43 Coke hurm. but we do potentially modify it, I suppose. just wondering how to make make splint happy
15:44 Infinoid hmm.  where was that splint warning, exactly?
15:45 Infinoid is that in the utf16 upgrade stuff?
15:46 nopaste "coke" at 72.228.52.192 pasted "splint output" (139725 lines) at http://nopaste.snit.ch/16221
15:46 Coke it's a little big.
15:47 Infinoid oh, the warning is in the caller
15:47 Infinoid Splint doesn't understand that the old value will be GC'd, I guess
15:53 Psyche^ joined #parrot
15:53 Coke hurm. my nopaste still hasn't completed locally. =-)
15:54 Infinoid still "transferring data" here too
15:54 Coke you have access to feather?
15:54 Coke ackshully...
15:54 Infinoid yeah
15:54 Infinoid my browser is reading thousands of "    src/nci.c:11750:9: Fresh storage temp_pmc created" lines at the moment
15:55 shucho joined #parrot
15:56 Coke feather now has ~coke/splint.out.gz
15:56 Coke er, make that ~coke/public_html/splint.out.gz
15:56 Coke (which is presumably available via the web)
16:01 Infinoid bacek: (r38026) Nice patch, I think it's a really interesting approach, and you're going to shortcut  whole bunch of effort this way.  How well does it work? :)
16:01 Infinoid ^a
16:01 joubert joined #parrot
16:02 bacek Infinoid: t/pmc/packfile.t It can produce simple PBC. And even parse it after :)
16:02 Infinoid bacek++
16:03 bacek I actually spend whole day trying to do it in different way - properly reuse available PackFile_* functions... EPIC FAIL.
16:03 bacek spent
16:03 jrockway joined #parrot
16:04 Infinoid Yeah, that was the route I was going down.  It was time-consuming.
16:06 bacek And it's almost impossible... Even serialized PackFile_Directory depends on PackFile.
16:07 Infinoid Coke: That page finally finished loading
16:07 Whiteknight cotto: ping
16:10 bacek need sleep
16:10 bacek afk # zzzz....
16:18 Infinoid Coke: I think quite a lot of that could be cleared with a few annotations to teach splint about the gc
16:18 * Infinoid is reading http://www.splint.org/manual/html/sec5.html, section 5.2.5 looks promising
16:21 Infinoid unfortunately I'm out of parrot tuits until probably monday night
16:23 particle- sailing?
16:23 purl sailing is pretty careful about wind too.  ;)
16:24 Infinoid among other things, yes :)
16:41 flh joined #parrot
16:45 joubert joined #parrot
17:27 contingencyplan joined #parrot
17:35 dalek rakudo: 029aa51 | (Moritz Lenz)++ |  (2 files):
17:35 dalek rakudo: bring Hash.reverse up to date; test S32-hash/reverse.t
17:35 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/0​29aa517ebe4c44dcd34276c384511cf49914161
17:35 shorten dalek's url is at http://xrl.us/beoij2
17:35 dalek rakudo: 7b2c47c | (Moritz Lenz)++ |  (2 files):
17:35 dalek rakudo: Implement Hash.push, and add S32-hash/push.t to spectest.data
17:35 dalek rakudo: All of these new tests pass except one, which trips over a Null PMC access in
17:35 dalek rakudo: type() from is_deeply in Test.pm
17:35 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​b2c47c49476916e5d4e44519fd11cbd8f1b0b2a
17:35 shorten dalek's url is at http://xrl.us/beoij4
17:35 dalek rakudo: c44a05b | (Moritz Lenz)++ | t/spectest.data:
17:35 dalek rakudo: we pass the new S09-typed-arrays/hashes.t
17:35 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c​44a05b7d93fac9b0f2f6d6106215d7ddb3e0a14
17:35 shorten dalek's url is at http://xrl.us/beoij6
17:35 dalek rakudo: 79aba97 | (Moritz Lenz)++ | README:
17:35 dalek rakudo: [README] reworked paragraph on rakudo.org
17:35 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​9aba97eeb02d906ca7d162db0c490e032a87a25
17:35 shorten dalek's url is at http://xrl.us/beoij8
17:45 Tene joined #parrot
17:49 dalek parrot: r38027 | whiteknight++ | trunk (2 files):
17:49 dalek parrot: Finally resolve RT#46719. One of the tests was using ge instead of isge, which miraculously produced a passing test using the old initialized register values. Fixing the test enables us to avoid initialization
17:49 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38027/
18:34 cotto Whiteknight, pong
18:44 tetragon_ joined #parrot
18:49 Whiteknight hey cotto, I wanted to ask what the status of the UnionVal stuff was?
18:51 nopaste "pmichaud" at 72.181.176.220 pasted "fun with comparisons" (22 lines) at http://nopaste.snit.ch/16228
18:53 cotto Whiteknight, most PMCs have been converted.  There are a couple of questionable status that I haven't done (Pointer, Ref) that are deprecated but still in use.
18:54 cotto I don't think there are any blockers to finishing the task.
18:55 cotto The remaining tasks are:
18:56 cotto * fix the remaining PMCs which use the macros (Pointer, Ref, SharedRef, Namespace, maybe others)
18:56 cotto * ack for any other instances of the macros and try to DTRT with them
18:56 cotto * remove the macros from pobj.h
18:57 cotto * done!
18:59 cotto For some reason, PMC_pmc_val appears a bunch in sharedref.c and ref.c.  I haven't looked into why, but it's probably not a difficult fix.
19:00 cotto If you'd like to help, it should parallelize well.
19:05 tetragon joined #parrot
19:06 Coke cotto: ref is up for removal. don't fix it, kill it.
19:08 cotto with fire?
19:08 Coke nuke it from orbit.
19:08 Coke nuke it from orbit is <reply>It's the only way to be sure.
19:11 acajou joined #parrot
19:13 acajou Hi all.  So, since yesterday, I can't get a clean spectest...  My latest attempt (with a source tree from this AM): http://perl.pastebin.com/d8b9e694
19:13 acajou Is no one else getting errors?
19:19 altious joined #parrot
19:34 japhb joined #parrot
19:36 dalek parrot: r38028 | whiteknight++ | trunk/src (2 files):
19:36 dalek parrot: Fix Parrot_str_to_int to detect overflow and throw an EXCEPTION_ERR_OVERFLOW.
19:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38028/
19:36 Whiteknight acajou: what platform are you on?
19:36 acajou gcc version 4.2.4 (Ubuntu 4.2.4-1ubuntu3)
19:37 acajou (Kubuntu 8.04.2)
19:37 acajou DId git fetch / make realclean / make / make spectest
19:37 acajou (and perl Configure.pl --gen-parrot  b4 make)
19:38 acajou I've tried about 6 times since yesterday, and it's never been clean.
19:39 jonathan acajou: Any chance you can run one of the tests directly and nopaste the output here or (perhaps better) on #perl6?
19:44 acajou 10-4
19:44 purl 6
19:47 acajou eg http://perl.pastebin.com/d341958bd
19:48 jonathan acajou: Ah, but if there is a .rakudo fiel you should run that, not the .t...
19:49 jonathan Otherwise can't tell which tests are failing that are not expected to be.
19:49 dalek parrot: r38029 | cotto++ | trunk/src/pmc (2 files):
19:49 dalek parrot: [PMC] replace some code that uses Ref
19:49 dalek parrot: Neither of the diffs has test coverage, so this is my best guess at DTRT.
19:49 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38029/
19:49 acajou OK, sorry.
19:50 acajou Coming right up.
19:52 acajou http://perl.pastebin.com/d29e19c3b
19:54 jonathan acajou: You certainly have no local patches?
19:55 jonathan This is...odd.
19:56 acajou None whatsoever.  git clone and then only git fetch.  I'm just the peanut gallery here, excited that Perl6 looks almost here...
19:56 acajou Maybe I should nuke my source three and re-clone...
19:59 * acajou AFK.
20:03 jonathan acajou: That would eliminate the possibility of contamination, for sure.
20:06 particle- unless there's an installed parrot in the way, still
20:09 dalek parrot: r38030 | cotto++ | trunk (9 files):
20:09 dalek parrot: [PMC] remove Ref and SharedRef, bump PBC_COMPAT
20:09 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38030/
20:21 rob joined #parrot
20:36 jonathan particle-: I hadn't considered that one, good point.
20:45 contingencyplan joined #parrot
20:48 darbelo joined #parrot
20:50 cotto Infinoid, ping
20:57 cotto Infinoid, unping
20:57 cotto jonathan, ping
20:59 mikehh I keep my parrot and rakudo dirs separate - I generally start with parrot and do a amke realclean, svn up, perl Configure.pl+, make world, make html, make smoke and then-
20:59 jonathan cotto: pong
21:00 cotto jonathan, t/pmc/sockaddr.t is confusing
21:01 cotto It looks like it should be testing the SockAddr PMC, but it creates a Socket.
21:01 mikehh in rakudo: make realclean, git pull, perl Configure.pl --parrot-config=../parrot/parrot_config, maje, makr test, make spectest
21:01 cotto Could you add a quick note explaining why it does that?
21:01 jonathan cotto: No, because (a) I didn't write it and (b) I don't understand sockets even if I did write it. ;-)
21:02 jonathan cotto: bacek will probably have much more helpful answer
21:02 jonathan cotto: My best guess though, is the design changed.
21:02 cotto ok
21:02 jonathan When the ops went away, SockAddr was created by some method instead.
21:02 cotto svn blamed you for most of it, but I didn't dig very deeply.
21:02 jonathan Yes, I applied the patch that added it.
21:03 jonathan I more reviewed the C code and ensured the tests passed. And I think it looked sane at the time...
21:03 cotto gotcha
21:03 jonathan Anyway, bacek wrote the patch - it was before he got the commit bit though. :-)
21:03 cotto I'll ask him about it then.
21:04 cotto seen bacek
21:04 purl bacek was last seen on #parrot 4 hours, 53 minutes and 10 seconds ago, saying: afk # zzzz....
21:04 jonathan Cool. Sorry to not be much more helpful.
21:04 Whiteknight joined #parrot
21:07 msmatsko joined #parrot
21:11 iblechbot joined #parrot
21:15 dalek parrot: r38031 | cotto++ | trunk (11 files):
21:15 dalek parrot: [PMC] more cleanup from the Ref/SharedRef removal
21:15 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38031/
21:25 dalek parrot: r38032 | whiteknight++ | trunk/src/pmc/hash.pmc:
21:25 dalek parrot: remove a PMC_struct_val from src/pmc/hash.c
21:25 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38032/
21:36 * acajou nuked his repo and re-cloned, running spectest now.
21:38 acajou Looks like it'll be clean(er), already apssed S02 without error.  Not sure what had happened...
21:38 dalek parrot: r38033 | cotto++ | trunk/src/pmc/namespace.pmc:
21:38 dalek parrot: [PMC] remove PMC_pmc_val from NameSpace PMC
21:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38033/
21:40 jonathan acajou: Maybe some weird makefile not cleaning something issue...
21:43 * acajou nods.
21:45 * acajou would like to take a moment to express his fondness for the "-i" command line argument from Perl5, as S19 has its future as uncertain...
21:46 * acajou wonders why Kopete doesn't wrap long lines...
21:48 dalek parrot: r38034 | cotto++ | trunk/src/pmc/lexinfo.pmc:
21:48 dalek parrot: [PMC] removed unneeded PMC_pmc_val usage
21:48 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38034/
21:55 dalek parrot: r38035 | cotto++ | trunk/src/dynpmc/dynlexpad.pmc:
21:55 dalek parrot: [dynpmc] switch DynLexPad to ATTRs
21:55 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38035/
21:56 acajou All tests successful
21:56 purl ship that sucker
21:56 particle- we ship every month, purl.
21:56 purl particle-: what?
22:00 jonathan acajou: Ah, phew. :-)
22:01 GeJ Good morning everyone
22:02 dalek parrot: r38036 | cotto++ | trunk (5 files):
22:02 dalek parrot: [PMC] PMC_pmc_val is history
22:02 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38036/
22:02 acajou :)
22:02 Limbic_Region joined #parrot
22:08 dalek parrot: r38037 | whiteknight++ | trunk (2 files):
22:08 dalek parrot: update src/pmc/pointer.pmc to use ATTRs instead of PMC_struct_val. Requires a change to the nci thunk generator.
22:08 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38037/
22:11 darbelo left #parrot
22:13 cotto Whiteknight++ #I was hoping to avoid doing that.
22:14 cotto Although I'd have used a VTABLE in the generated NCI code.
22:15 dalek parrot: r38038 | whiteknight++ | trunk/src/pmc/pointer.pmc:
22:15 dalek parrot: add a custom destroy method to the Pointer PMC, so it can free the new memory that it allocates
22:15 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38038/
22:17 clunker9_ joined #parrot
22:18 dalek parrot: r38039 | cotto++ | trunk (2 files):
22:18 dalek parrot: [PMC] PMC_str_val is history
22:18 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38039/
22:21 Whiteknight cotto: yeah, a vtable might be a little better
22:22 Whiteknight probably should have just used get_pointer I guess. want to rework it?
22:22 cotto Sure.  It's an easy fix.  I'll just have to remember to build without jitcapable.
22:23 Whiteknight oh, I didn't do that. I guess I haven't tested it properly then
22:23 cotto We'll find out how it works soon enough. ;)
22:24 cotto While that's going, do you want to tackle PMC_struct_val?
22:24 cotto There aren't many instances of it left, so I don't think it'll take very long.
22:25 cotto PMC_int_val should be easy too.  You could probably take them both out with a fairly small commit.
22:26 Whiteknight I was working on PMC_pmc_val in the Namespace PMC, but am running into problems
22:26 cotto I got that already.
22:26 Whiteknight oh okay, that makes it easy then
22:26 Whiteknight I'll take PMC_struct_val?
22:26 cotto go for it
22:27 cotto and svn up frequently ;)
22:27 Whiteknight Yeah, lesson learned
22:30 cotto make svn up yours
22:41 dalek parrot: r38040 | cotto++ | trunk (2 files):
22:41 dalek parrot: [NCI] use VTABLE functions instead of twiddling with PMC internals directly
22:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38040/
22:42 bacek good morning
22:43 bacek cotto: (sockaddr.t) Previously it tests "sockaddr" op. But nowdays "sockaddr" is just method on Socket
22:50 Whiteknight good morning bacek
22:50 bacek good morning Whiteknight
22:54 eternaleye joined #parrot
23:01 * acajou waves.
23:01 acajou left #parrot
23:06 cotto Whiteknight, I'm working on PMC_int_val.
23:07 Whiteknight okay, I'm about to commit PMC_struct_val, just running a quick coretest to make sure I'm not acting retarded at the keyboard
23:08 cotto sweet
23:08 cotto Tests are nice like that.
23:10 Whiteknight okay, committed. What's next?
23:11 cotto Just PMC_int_val, although it's used in some jit code.
23:11 dalek parrot: r38041 | whiteknight++ | trunk (6 files):
23:11 dalek parrot: kill PMC_struct_val. kill it dead
23:11 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38041/
23:12 kid51 joined #parrot
23:12 cotto nm.  It's used in a switch *after* a break.  It's unreachable.
23:12 Whiteknight yay! dead code
23:22 eternaleye joined #parrot
23:29 eternaleye joined #parrot
23:32 tetragon_ joined #parrot
23:46 tetragon joined #parrot
23:56 cotto Whiteknight, it looks like 38037 also requires some jit code to be fixed.
23:58 joubert joined #parrot

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

Parrot | source cross referenced