Camelia, the Perl 6 bug

IRC log for #parrot, 2009-03-23

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:01 mikehh joined #parrot
00:01 dan What does this mean when I try to use target=parse: "load_bytecode" couldn't find file 'library/Data/Dumper.pir'
00:09 AndyA joined #parrot
00:14 kid51 From: 'make help':
00:14 kid51 Installation:
00:14 kid51 install:           Install under '/usr/local' on Unix systems
00:14 kid51 install-dev:       Same as 'install' but also install support for development.
00:15 dan kid51: It worked, thanks.
00:37 Whiteknight joined #parrot
01:05 tetragon_ joined #parrot
01:20 NotFound joined #parrot
01:20 NotFound hi
01:34 simontwo hello
01:34 dukeleto joined #parrot
03:02 megazooch ighto
03:02 megazooch :O
03:08 dukeleto there does not seem to be a parrot built-in for getting the polar representation of a Complex, am I missing something or is on purpose?
03:12 Infinoid If you feel that's something useful, you can always submit a patch. :)  (I don't deal with Complexes much, so I wouldn't know)
03:35 janus joined #parrot
03:51 TiMBuS joined #parrot
04:05 Counterspell joined #parrot
04:34 dukeleto Infinoid: would I be adding a new method to the Complex pmc vtable?
04:54 cotto dukeleto, we're trying to refer to VTABLE entries as functions rather than methods.  If you see documentation referring to VTABLE methods, it's a bug.
04:55 dukeleto cotto: good to know
04:56 cotto For what you want, you'd add a METHOD to the Complex PMC, similar to Complex's ln.
04:57 cotto There should be plenty of similar code there to give you the general idea.
05:01 cotto Hmmm.  Apparently our documentation is quite buggy.
05:02 msmatsko joined #parrot
05:08 cotto Also, don't forget to add tests when you submit the patch.
05:13 cotto 'vtable entries' is fine too.
05:14 cotto though not preferred
05:44 japhb joined #parrot
07:15 uniejo joined #parrot
07:29 contingencyplan joined #parrot
08:20 cotto (make headerizer)++
08:46 dalek parrot: r37644 | cotto++ | trunk (7 files):
08:46 dalek parrot: [hash] change Parrot_new_INTVAL_hash to parrot_new_intval_hash, make it return a Hash*
08:46 dalek parrot: also, unrelated trailing space fix
08:46 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37644/
08:51 mikehh t/codingstd/trailing_space.t failure - I have a patch - should I post it
08:53 masak joined #parrot
08:53 cotto I think I just fixed it with that commit.
08:53 cotto irony++
08:54 cotto mikehh, if that commit didn't catch it, feel free to nopaste
08:55 dalek parrot: r37645 | cotto++ | trunk/src/hash.c:
08:55 dalek parrot: [hash] move similar functions together, reformat code (no functional changes)
08:55 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37645/
09:43 mikehh cotto: I am about to extract all the compiler warnings from make world - do you think that a worthwhile project?
09:59 cotto mikehh, I don't think it'd hurt.  We generally try to avoid compiler warnings.
10:00 cotto I don't think you'll find very many.
10:01 cotto either way, it's long past bed time.  night
10:01 mikehh I've got about a dozen or so with gcc 4.3.2 on Kubuntu Intrepid i386
10:02 mikehh night - it's monrning for me
10:03 mikehh s/monrning/morning/
10:10 dalek parrot: r37646 | moritz++ | trunk/src/ops/experimental.ops:
10:10 dalek parrot: [cage] try to clean up POD errors from experimental.ops
10:10 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37646/
10:23 bacek joined #parrot
10:40 Piper joined #parrot
10:40 Piper Hi there.  I am Piper.  I am now publicly logging this channel.  If you don't want to be logged, please leave now.
10:47 Piper joined #parrot
10:47 Piper Hi there.  I am Piper.  I am now publicly logging this channel.  If you don't want to be logged, please leave now.
10:50 masak Piper: for a bot, you're awfully talkative. >:-(
10:50 masak and a bit rude, too.
10:50 moritz especially for a bot that's supposed to listen, not to talk
10:50 masak oh well. it's early in bot history still.
10:51 cognominal joined #parrot
10:52 alvar joined #parrot
11:05 clunker3 joined #parrot
11:23 fperrad joined #parrot
11:41 ruoso joined #parrot
12:15 bacek joined #parrot
12:47 alvar joined #parrot
13:03 gryphon joined #parrot
13:16 PerlJam joined #parrot
13:26 Coke ~~
13:29 Coke warnings: https://trac.parrot.org/parrot/wiki/BuildWarnings
13:36 Coke RT down, presumably due to the perl.org server migration.
13:36 moritz but the RT pages now load faster than ever before :-)
13:43 riffraff joined #parrot
13:51 Coke "you funny, doctah jones."
13:54 Infinoid Coke: That's for unfixable warnings?  Or just a random place to stick the fixable stuff?
13:55 Coke all of the above?
13:55 Infinoid (I've got some pthreads and gdbm warnings in my system headers which have been around for a long time, and aren't parrot's fault.)
13:56 Coke I think certainly for the unfixable. I think the other ones are debatable, but having a central page for them makes sense to me.
13:56 Coke (especially since warnings aren't failures.)
13:56 Coke (except when they are elsewhere)
14:02 Infinoid and the yacc/lex warnings!  yay.
14:04 mikehh I get 27 warnings in 13 files - Kubuntu Intrepid i386 gcc version 4.3.2 (Ubuntu 4.3.2-1ubuntu12)
14:05 Infinoid Coke: I'm not really sure how you intend to format that file to keep it manageable, but my warnings are up, please move things around however you like.
14:05 Andy joined #parrot
14:06 dalek tracwiki: v2 | Infinoid++ | BuildWarnings
14:06 dalek tracwiki: Add the warnings I regularly see (and ignore).
14:06 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Bu​ildWarnings?version=2&action=diff
14:06 shorten dalek's url is at http://xrl.us/bek9rs
14:11 Infinoid Some of those items might already have tickets.
14:28 dalek parrot: r37647 | fperrad++ | trunk/lib/Parrot/Distribution.pm:
14:28 dalek parrot: [tools] fix RT #43691
14:28 dalek parrot: languages are gone, so no longer confusion between Lisp and Lex files (*.l).
14:28 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37647/
14:40 Coke Infinoid: added the ones I have on top of that.
14:40 Coke I'd put in the RT, but rt down.
14:42 dalek tracwiki: v3 | coke++ | BuildWarnings
14:42 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Bu​ildWarnings?version=3&action=diff
14:42 shorten dalek's url is at http://xrl.us/bek9ui
14:42 Coke ... I just had a terrible idea to rename TWIP.
14:46 pmichaud good morning #parrot
14:47 pmichaud 13:51 <Coke> "you funny, doctah jones."
14:47 pmichaud Odd, we watched that episode just two nights ago... :-|
14:51 dalek tracwiki: v4 | mikehh++ | BuildWarnings
14:51 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Bu​ildWarnings?version=4&amp;action=diff
14:51 shorten dalek's url is at http://xrl.us/bek978
14:58 Coke pmichaud: heh.
14:59 Coke timbunce++ (I can now run Devel::NYTProf against pmc2c.pl)
14:59 Coke sadly, no obvious "FIXME"s leap out.
15:01 ruoso joined #parrot
15:02 particle io bound, then?
15:03 Coke possibly.
15:05 davidfetter seen pmichaud
15:05 purl pmichaud was last seen on #parrot 17 minutes and 49 seconds ago, saying: Odd, we watched that episode just two nights ago... :-|
15:05 davidfetter pmichaud, is anybody working on getting rakudo working with a pre-installed parrot?
15:06 Coke I was under the impression that already worked.
15:06 Coke (and was used as the basis of a patch to partcl which I haven't applied)
15:13 PacoLinux joined #parrot
16:13 szabgab I've started to write this Inline::Rakudo module that would load perl6.pbc and then allow the user to execute perl 6 code from perl 5
16:13 szabgab the problem is that when I call    load_bytecode '$ENV{RAKUDO_DIR}/perl6.pbc'
16:14 szabgab it complains about not finding PCT.pbc
16:14 szabgab unless I happen to be in the parrot directory
16:15 szabgab can I tell in any way parrot to look for the PCT.pbc and other files in a specific directory?
16:27 Coke are you running against an installed parrot?
16:28 szabgab not, one that was built by Rakudo using --gen-parrot
16:28 szabgab BTW here is the code: http://github.com/szabgab/perl6-in-perl5
16:29 rg joined #parrot
16:53 magnachef_ joined #parrot
17:02 pmichaud davidfetter: there have been several folks who have wokred on getting Rakudo to work with an installed parrot, but I don't know the current status of those efforts.
17:02 pmichaud afaik it doesn't 'already work'
17:03 pmichaud szabgab: Parrot should already be smart enough to know where to locate PCT.pbc -- if not, that sounds like a bug in parrot.
17:03 davidfetter pmichaud, k. btw, is rakudo in a state where it's worth packaging it?
17:03 davidfetter (ports, debs, RPMs...)
17:04 pmichaud davidfetter: short answer: yes
17:04 pmichaud longer answer:  we definitely want to know what we need in order to package it.
17:04 pmichaud (i.e., if there are any major stumbling blocks to packaging it)
17:04 davidfetter well, getting it working with an installed parrot would be a big one
17:05 davidfetter so you can list parrot (or parrot-dev, or whatever) as a dependency
17:05 pmichaud I haven't seen (haven't looked for) documentation that describes how dynpmcs and dynops are to be generated now.  Rakudo is still using the old methods.
17:06 davidfetter hrm. that'd be an "upstream" issue, i suppose
17:06 davidfetter !dynpmc
17:06 davidfetter hrm
17:07 pmichaud allison submitted a patch towards the end of february that was supposed to address it, but that patch was before I did the big refactor of Rakudo's configure/build process
17:07 pmichaud I haven't had time to review the patch and figure out exactly what it was attempting to do.
17:08 pmichaud (and the patch is based on the old build environment, so it's now a fair bit out of date.)
17:08 davidfetter dynops?
17:08 purl dynops is, like, definitely borked on windows. i forgot about that
17:08 davidfetter hrm
17:08 szabgab pmichaud, or the way rakudo is building it with the --gen-parrot flag, or the way I stup things
17:08 davidfetter dynpmcs?
17:09 pmichaud szabgab: when rakudo builds parrot with the --gen-parrot flag, it's just doing a standard developer build.
17:09 pmichaud it's not doing any thing out of the ordinary as far as building parrot goes.
17:09 szabgab maybe it should :-)
17:10 szabgab though I guess the standard parrot build should work too
17:13 jonathan rakudobug?
17:13 purl it has been said that rakudobug is mailto:rakudobug@perl.org
17:14 sjn pmichaud: did you make a hotel reservation in Oslo?
17:15 sjn pmichaud: only hotel name I have registered here is "Best Western" which is kinda silly since there are probably 15-20 of them in Oslo
17:16 pmichaud sjn: yes, I made a reservation -- I'm at the BW Anker
17:16 pmichaud (also listed at http://www.perlworkshop.no/np​w2009/wiki?node=WhoStaysWhere )
17:17 pmichaud jonathan: where are you staying while in Oslo?
17:17 jonathan pmichaud: I *think* the same one...I was going to check that with batman
17:18 jonathan I know it's reserved.
17:25 n8willis joined #parrot
17:25 Khisanth joined #parrot
17:26 NordQ joined #parrot
17:28 jonathan pmichaud: Best Western Anker Hotel
17:33 sjn pmichaud: great
17:42 * Coke misses RT.
17:43 pmichaud jonathan: maybe add yourself to http://www.perlworkshop.no/np​w2009/wiki?node=WhoStaysWhere ?
17:44 jonathan pmichaud: Will do.
18:07 Woody4286 joined #parrot
18:26 dalek parrot: r37648 | allison++ | trunk/docs/pdds/pdd22_io.pod:
18:26 dalek parrot: [pdd] Documenting finalized network I/O decision.
18:26 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37648/
18:27 cotto seen wknight8111
18:27 purl wknight8111 was last seen on #parrot 5 days, 23 hours, 14 minutes and 38 seconds ago, saying: is too much of a wuss to quit his job  [Mar 17 19:11:41 2009]
18:28 cotto mikehh, ping
18:39 barney joined #parrot
18:41 protorom joined #parrot
18:41 cotto hi barney
18:48 barney hi cotto
18:48 mikehh cotto: hi
18:53 * barney is away for dinner
18:55 cotto mikehh, in the page on compiler warnings you mention one for some OS X functions that already has an rt.
18:55 cotto The page should have a link to the rt.
18:56 mikehh I think that was Coke on the OS X stuff
18:57 mikehh I just added my gcc 4.3.2 bits
18:59 cotto Right you are.  I'll bug Coke about it.
18:59 cotto Coke, ping
19:06 Coke cotto: pong
19:06 Coke cotto: RT is DOWN, or I would have linked to the rt. =-)
19:07 Coke s/is/was/
19:08 Coke ... and in the text, it says, " (There's an RT for this one)" sheesh. =-)
19:08 cotto no excuses now ;)
19:11 Coke fixed
19:13 dalek tracwiki: v5 | coke++ | BuildWarnings
19:13 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Bu​ildWarnings?version=5&amp;action=diff
19:13 shorten dalek's url is at http://xrl.us/bembft
19:14 cotto thanks
19:17 Tene joined #parrot
19:23 mikehh one of the longstanding warnings in src/thread.c - suggest braces around empty body in do 'statement' is: CLEANUP_POP(1);
19:23 contingencyplan joined #parrot
19:23 mikehh so far I havewn't been able to find this macro
19:25 Coke use ack.
19:26 Coke ack -a CLEANUP_POP =>
19:26 Coke include/parrot/thr_pthread.h ; include/parrot/thr_windows.h
19:26 Coke and include/parrot/thread.h
19:27 mikehh ah - I was looking in the parrot includes not the c library
19:27 Andy better still, ack for cleanup_pop, which will be case-insensitive
19:27 particle or make tags-vi && vim -t CLEANUP_POP
19:29 desertmax joined #parrot
19:33 dalek eclectus: b482db0 | (Bernhard Schmalhofer)++ | t/pmc/boolean.t:
19:33 dalek eclectus: add tests for stringification and numification
19:33 dalek eclectus: review: http://github.com/bschmalhofer/eclectus/comm​it/b482db08841cadf890b8b98e0a9556383495ab2a
19:33 shorten dalek's url is at http://xrl.us/bembiy
19:33 dalek eclectus: 41012b4 | (Bernhard Schmalhofer)++ | t/pmc/boolean.t:
19:33 dalek eclectus: include/ is in the searchpath
19:33 dalek eclectus: review: http://github.com/bschmalhofer/eclectus/comm​it/41012b420f0e437c39b719e5cbcada526c8f3db5
19:33 shorten dalek's url is at http://xrl.us/bembi2
19:34 Coke particle: does that work when it has multiple definitions?
19:34 Coke Andy: wouldn't you need ack -i ?
19:34 Andy nope
19:34 Andy ack assumes -i if you have no caps
19:34 Andy see --smart-case
19:34 protorom left #parrot
19:34 Coke hurm. my initial reaction is 'ew".
19:34 Coke with mismatched quotes, even.
19:35 Coke doesn't seem to be present in my ancient 1.80
19:35 Andy so throw a --no-smart-case in your .ackrc
19:36 Coke defines are case sensitive anyway, aren't they?
19:36 Andy sure
19:36 Andy but who knows what generated it.
19:36 Andy Just in general -i is good I've found
19:37 particle it brings you to the first, then you navigate to the others with :ts 2 or whatever
19:37 Coke Andy,particle: nifty
19:37 * Coke updates his ack
19:45 mikehh #  define CLEANUP_POP(a)     pthread_cleanup_pop(a)
19:45 dalek pipp: c85f726 | (Bernhard Schmalhofer)++ |  (4 files):
19:45 dalek pipp: Merge branch 'master' of git@github.com:bschmalhofer/pipp
19:45 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/c85f726e2bbad68ca8d753a972b50fbd411ee2c3
19:45 shorten dalek's url is at http://xrl.us/bembkd
19:45 dalek pipp: 812a445 | (Bernhard Schmalhofer)++ | t/pmc/ (3 files):
19:45 dalek pipp: include is in the search path
19:45 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/812a445acdbfaaee9e7d0a41cda1fc72549efbc9
19:45 shorten dalek's url is at http://xrl.us/bembkj
19:45 Coke wierd. I have pod simple 3.05 installed. I install 3.07 ; in cpan, i type 'r' ... shows me I have 3.05 installed and that 3.07 is available.
19:54 mikehh Coke: I have had this problem before - check your install directories
19:54 Coke "they exist!" =-)
19:54 Coke There are two copies of Pod::Simple; I presume one is the core, one is my upgrade.
19:55 Coke annoying that perl+cpan aren't DTRT.
19:55 mikehh you will probably find that the new version was installed but in a different directory - lower in the search heirarchy
19:57 barney is get_string() still the VTABLE for stringification ?
19:58 Coke barney: SFAIK.
20:00 barney In Eclectus I seem to get a different stringification than implemented in get_string()
20:00 dalek eclectus: 4a8ed6d | (Bernhard Schmalhofer)++ |  (2 files):
20:01 dalek eclectus: add implementation and test of name()
20:01 dalek eclectus: review: http://github.com/bschmalhofer/eclectus/comm​it/4a8ed6d4dd623c00e3a492b9918f8f333c9328a1
20:01 shorten dalek's url is at http://xrl.us/bembor
20:03 Coke barney: code sample?
20:04 dalek parrot: r37649 | NotFound++ | trunk/src/parrot_debugger.c:
20:04 dalek parrot: [debugger] allow running parrot_debugger without a file to debug
20:04 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37649/
20:04 barney http://github.com/bschmalhofer/pipp/blob/812a445a​cdbfaaee9e7d0a41cda1fc72549efbc9/t/pmc/boolean.t
20:04 shorten barney's url is at http://xrl.us/bembpm
20:06 NotFound barney: I fixed a thing like that in ecmascript some weeks ago.
20:08 dalek parrot: r37650 | cotto++ | trunk/src (5 files):
20:08 dalek parrot: [hash] use get/set_pointer instead of PMC_struct_val on Hash-like PMCs
20:08 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37650/
20:09 barney get_string() in ecmascript/src/classes/Boolean.pir ?
20:10 NotFound barney: Yes
20:12 barney (np preserved svn history)--
20:12 barney (non preserved svn history)--
20:13 barney found it. POD wasn't protected by C-comments
20:15 NotFound barney: https://trac.parrot.org/languages/browser​/ecmascript/trunk/src/classes/Boolean.pir
20:15 shorten NotFound's url is at http://xrl.us/bembrf
20:18 dalek eclectus: 79b14fd | (Bernhard Schmalhofer)++ |  (2 files):
20:18 dalek eclectus: fix parsing of *.pmc file, by adding a C startcomment
20:18 dalek eclectus: review: http://github.com/bschmalhofer/eclectus/comm​it/79b14fd48abf5a1787122f486f96a9924c3c2a14
20:18 shorten dalek's url is at http://xrl.us/bembrs
20:19 barney Eclectus is (mostly) back, http://smolder.plusthree.com/app/pu​blic_projects/report_details/19325
20:19 shorten barney's url is at http://xrl.us/bembru
20:21 dalek parrot: r37651 | allison++ | trunk/ports/cpan (2 files):
20:21 dalek parrot: [ports] Adding CPAN release instructions as a separate guide.
20:21 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37651/
20:23 allison joined #parrot
20:25 dalek parrot: r37652 | allison++ | trunk/ports/cpan/pause_guide.pod:
20:25 dalek parrot: [ports] Moving PAUSE documentation into CPAN ports directory.
20:25 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37652/
20:28 Coke svn:up::git: ?
20:28 Coke pull?
20:28 purl KABLAMM! or English slang for snogging/kissing/taking someone home
20:28 allison purl: forget pull
20:28 purl allison: I forgot pull
20:29 dalek parrot: r37653 | allison++ | trunk/MANIFEST:
20:29 dalek parrot: [ports] Moving PAUSE guide to CPAN ports directory.
20:29 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37653/
20:30 Tene Coke: pull is okay.  I usually prefer git fetch; git rebase origin;
20:30 allison Coke: I've been using 'git pull' (there seem to be several different equivalents, depending on how you have your local git set up)
20:30 Tene In the case that you don't have extra commits on the local tree, they're identical.
20:30 Coke origin?
20:30 purl i heard origin was lib/Parrot/Config.pm
20:31 particle joined #parrot
20:31 Coke origin is also a made up religion from the Stargate universe.
20:31 purl okay, Coke.
20:32 ujwalic joined #parrot
20:34 allison Tene: how do I wipe out all local commits and reset to the foreign master? "git reset --hard"?
20:35 Infinoid git reset --hard origin
20:35 Infinoid or origin/master, if you want to be ultraspecific
20:36 Coke I for one am glad we didn't switch parrot over to git just yet. ^_^
20:36 Infinoid I think it would have done parrot more good than it did rakudo, to be honest
20:37 * allison now wonders what "HEAD is now at 1c263b0 implement Match.chunks" means...
20:37 Coke Infinoid: long term, no doubt.
20:37 allison short-term it would have foobar'd us for the 1.0 release
20:37 Infinoid we have branch merging problems *now*, that is my only reason for caring about git
20:37 Coke allison: let me pre-empty the git folks:
20:37 Coke now that we've released 1.0 ... can we switch to git? =-)
20:37 allison no :)
20:38 Coke ... pre-empty? there's an odd slip.
20:38 mikehh agreed
20:39 allison ("pre-empty": but gives it a nice round feel)
20:39 barney Shouldn't string_chr()  be Parrot_str_chr() ?
20:40 allison lots of reasons, but the bottom line is that we've taken too many hits from infrastructure changes in the past 6 months, and need to be focusing on other things
20:40 Infinoid ...or at least focusing on fixing the issues with the new infrastructure we already have
20:40 mikehh at least in svn you can refer to r37645 or whatever,  with git -- wow
20:40 allison most of our failed merges are user-error, not SVN's fault, and git provides even fewer safety nets
20:42 gravity mikehh: You only need the first part of the git hash to refer to a change, so it's not really any worse than an svn rev in practice.
20:42 allison (I do, though, completely think git was the right choice for Rakudo and Cardinal, and several other languages. It's all about appealing to a particular language community.)
20:42 mikehh don't agree - at least svn is sequential
20:42 barney For PHP I should have switched to CVS
20:42 Infinoid I usually don't bother with the hashes, we have named branches for that.
20:43 Tene mikehh: I always copy/paste revisions anyway.  the format of the revision id isn't really relevant.
20:43 Tene It's a double-click regardless
20:44 Infinoid the svn merging issues I'm referring to were when using the proper tools in the proper way, it still fails with an obscure error message.  I've seen this happen with two branches so far, repeatably, and I don't think it was user error.
20:44 allison barney: string_chr is included in the deprecation list of "functions not named "Parrot_<subsystem>", but it may be going away entirely (it doesn't make much sense for non-ASCII character sets)
20:44 Infinoid I think svn has some issues with files changed in branch and renamed in trunk
20:45 Coke yes, that's a known svn issue. :|
20:45 mikehh I mean I have rakudo 1c263b0 and before that it was something like 44e1496
20:45 allison Infinoid: yes, if you rename a file in a branch, and then keep the branch around a long time, any subsequent changes to that file in trunk won't be merged in when you merge the branch into trunk
20:46 allison Infinoid: the solution, do file renaming as a short-lived branch, merge it into trunk, then start the larger refactors in a second branch
20:46 Infinoid it's not that the changes won't be merged.  It's that svn aborts the merge and leves the checkout in a trashed state
20:46 Infinoid But overall, you're right
20:46 allison Infinoid: I've successfully done it several times
20:47 Coke Infinoid: IME, it just throws out the merges, and leaves it in a workable state that is missing the changes.
20:47 Coke very (*#&$ annoying.
20:47 allison Infinoid: and, an aborted merge just needs a recursive revert to be restored to a completely clean state
20:47 Infinoid Ok.  maybe your version of svn is spiffier than mine then
20:47 Coke Apparently truMerge works around that issue.
20:48 Coke latest rakudo fails to build for me.
20:48 Coke Unable to open filehandle from path 'src/gen_actions.pir'
20:48 Coke (guessxing that's the PCT.pbc issue)
20:48 allison Coke: interesting, hadn't looked at truMerge before, will look at it more
20:48 dalek eclectus: 98f7075 | (Bernhard Schmalhofer)++ | src/pmc/eclectuscharacter.pmc:
20:48 dalek eclectus: fix get_string() in EclectusCharacter
20:48 dalek eclectus: review: http://github.com/bschmalhofer/eclectus/comm​it/98f707585954c7c0ff5d4e6ecf2d63a9570ae9d5
20:48 shorten dalek's url is at http://xrl.us/bembv2
20:49 Coke allison: was bitten by this bug at work, found that along with the bug report at svn.
20:49 barney allison: In Scheme there is a character class that is basically in Int, for printing it I need to convert the Int to a String
20:49 Coke (nearly every file in our repository is named wrong.)
20:50 Coke (and I can't fix it on a branch. or really while we have branches)
20:50 allison Coke: at the end of the day, we just have to bite the bullet and do it
20:50 Coke convert to git? ok. =-)
20:51 allison Coke: when I do file renaming in a branch, I just check the commit log of that file since the branch was made, and make sure I merge in any trunk changes before merging the branch back to trunk
20:52 Infinoid Uh, no.
20:52 Infinoid I have personally resolved two broken branch merges where the branch author had merged trunk changes into the branch and then everything was looking great... and then svn croaked merging the branch back into trunk.
20:52 Infinoid (And that has nothing to do with renaming, as far as I know)
20:53 allison I've resolved some pretty hairy ones too
20:53 Coke Infinoid: yup, that sounds different.
20:53 Infinoid I rather like pmichaud's approach of *never* merging trunk changes into the branch, which keeps the history linear
20:53 allison the worst are when the branch developer merges in trunk changes a piece at a time
20:53 Infinoid But this is all just working around a half-broken tool, which I still don't like.
20:53 allison Infinoid: pmichaud's approach has problems too
20:54 allison Infinoid: git has problems too
20:54 Coke (a piece at a time) but how else would you do that if you're going to keep the branch for longer than a <too short amount of time> ?
20:54 allison Coke: you do a complete merge from trunk a particular revision
20:54 Infinoid allison: git matches the branch workflow better.
20:54 barney Eclectus is back, http://smolder.plusthree.com/app/pu​blic_projects/report_details/19326
20:54 shorten barney's url is at http://xrl.us/bembw5
20:55 mikehh Ubuntu is using bzr
20:55 Infinoid Thing is, if you merge the trunk changes into the branch, that's the only necessary merge.  Pushing that back into trunk doesn't merge again
20:55 Infinoid So it's really frustrating that the branch author had already *done* the merge, but svn refused to pull the results of that back into trunk
20:55 allison Coke: that's clean, and keeps it up to date in documented chunks (that is, the next time you do a merge from trunk, you do it from the revision of the last merge to the current HEAD, and don't try to repeat earlier merges)
20:56 Infinoid So from my perspective, git doesn't have that problem.  It may have different problems, I suppose
20:56 Coke allison: ok. I thought that's what you meeant by a piece at a time.
20:56 allison mikehh: I like bzr, and would put it on the list of considerations in a couple of years when we think about infrastructure changes again
20:57 mikehh :-}
20:57 allison Coke: ah, ok, I meant "taking changes from one file here and there, and applying them to trunk"
20:57 allison I mean "applying them to the branch"
20:59 Coke barney: ping.
20:59 dalek parrot: r37654 | barney++ | trunk (3 files):
20:59 dalek parrot: [dynpmc] Removing the unused Pair PMC
20:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37654/
20:59 mikehh I think git is a lot like C - iy assumes you know what you are doing - which is fine for kernel hackers but I am not sure if it is that safe
21:00 Coke barney: there is no long a deprecation notice for the Pair PMC - so we can't rip it out until 1.5
21:00 Coke "no longer"
21:00 mikehh s/iy/it/
21:00 Coke msg barney there is no long a deprecation notice for the Pair PMC - so we can't rip it out until 1.5
21:00 purl Message for barney stored.
21:01 barney That's also valid for undocumented dynpmc's ?
21:01 Coke msg barney (also, removing a PMC invalidated PBC_COMPAT, which you didn't update there. for future reference)
21:01 purl Message for barney stored.
21:01 Coke oh, right, PBC_COMPAT prolly doesn't care about dynpmcs.
21:01 Coke but yes, it needs a deprecation notice.
21:02 Coke unless it's covered by another notice in there...
21:02 Coke (doesn't seem to be.)
21:02 Infinoid mikehh: How isn't it safe?  If you screwed up your tree, just don't commit :)
21:03 mikehh Infinoid:  but what happens if you do - commit O mean
21:03 mikehh this keyboard is driving me nuts
21:04 Infinoid mikehh: there's always reset for the local case and revert for the remote case, I suppose
21:04 Infinoid The kernel folks don't really have that problem.  Almost all their contributors don't have commit access, they have a great email review process
21:05 gravity Speaking of which, is there a mailing list where commit diffs go to, or just irc?
21:05 Infinoid http://lists.parrot.org/pipermail/parrot-​commits/Week-of-Mon-20090323/thread.html
21:06 shorten Infinoid's url is at http://xrl.us/bembzm
21:06 gravity Ah, thank you!
21:08 mikehh Infinoid: I was involved in a project using git and there was this clown - (who was a senior exec) - well never mind but I spent more time cleaning up after him than being productibe
21:08 Infinoid That sounds nasty regardless of which vcs was in use :(
21:09 mikehh You better believe it
21:09 Whiteknight joined #parrot
21:11 mikehh ome of my grandsons spilt a coke over my keyboard and I am having difficulty getting used to this new one
21:11 Infinoid :)  different key springiness always trips me up
21:11 mikehh you see s/ome/one/
21:12 Coke (split a coke) my daughter kicked over the cola I had near the imac keyboard. Grumble.
21:13 mikehh this is a slightly different size and I keep hitting the wrong key (next to it)
21:13 Coke SPILLED!
21:13 mikehh the cola version  not Will
21:14 Infinoid I tend to work heavily off of a heirarchy of local checkouts, so in a case like that, I probably just wouldn't pull upstream changes until the end of a ticket/task
21:14 Infinoid effectively forking the project
21:17 Infinoid tools like monotone let you control whose changes to trust / merge, maybe that situation is why
21:17 dalek parrot: r37655 | barney++ | trunk (3 files):
21:17 dalek parrot: [dynpmc] readding the Pair PMC, removal needs a deprecation period
21:17 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37655/
21:30 mj41 joined #parrot
21:38 bsdz joined #parrot
21:42 dalek markdown: 1fbf6a8 | (Francois Perrad)++ |  (5 files):
21:42 dalek markdown: implement code block
21:42 dalek markdown: review: http://github.com/fperrad/markdown/commit​/1fbf6a89191625b749c65cdf8b2b4c73a9f597bf
21:42 dalek markdown: 13ac70e | (Francois Perrad)++ |  (3 files):
21:42 dalek markdown: fix a TODO test
21:42 dalek markdown: but Smart is an extension, not strict Markdown
21:42 dalek markdown: review: http://github.com/fperrad/markdown/commit​/13ac70e043be6e2b70f53e1c26afff1ba005ccbf
21:42 shorten dalek's url is at http://xrl.us/bemb7d
21:42 shorten dalek's url is at http://xrl.us/bemb7f
21:53 dalek parrot: r37656 | cotto++ | trunk/src/pmc/hash.pmc:
21:53 dalek parrot: [PMC] replace most instances of PMC_struct_val from the Hash PMC, plus simplify code thanks to a saner hash API
21:53 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37656/
22:05 Whiteknight joined #parrot
22:29 GeJ Good mornin g everyone
22:30 cotto hi
22:30 purl hola, cotto.
22:39 dalek parrot: r37657 | cotto++ | trunk/src/pmc (4 files):
22:39 dalek parrot: [PMC] replace most remaining PMC_struct_val instances in Hash-like PMCs, remove some redundant code
22:39 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37657/
22:43 dalek rakudo: 7487710 | (Moritz Lenz)++ | t/spectest.data:
22:43 dalek rakudo: Add S05-metasyntax/repeat.t to t/spectest.data
22:43 dalek rakudo: Patch courtesy of Martin Kjeldsen, baest++
22:43 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​48771092a53e8910df4c633567c9a57b090a12e
22:43 shorten dalek's url is at http://xrl.us/bemcff
23:14 tetragon joined #parrot
23:20 Whiteknight joined #parrot
23:32 eternaleye joined #parrot

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

Parrot | source cross referenced