Parrot 2.10.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Long live Git! github.com/parrot/parrot | git clone git://github.com/parrot/parrot.git
Set by moderator on 18 November 2010.
fbrito Is anyone here also in charge of the Google Code-In Parrot tasks? 00:34
00:41 fbrito left
dalek rrot/mappedbytearray: b46284d | NotFound++ | t/pmc/mappedbytearray.t:
test MappedByteArray 'supported' method
 00:42
00:43 bacek joined
dalek TT #1874 created by allison++: ports/ directory is listed in .gitignore, can't commit files to it 00:52
TT #1874: trac.parrot.org/parrot/ticket/1874
allison fbrito: you're probably looking for dukeleto, though I don'g know if he's around today 00:53
*don't
dalek rrot: 43e7a36 | allison++ | src/gc/system.c:
Fix call in ifdef'd code for ia64 arch, signature of trace_mem_block was changed, but call not updated.
 01:02
01:26 theory left 01:58 gbacon left 02:09 gbacon joined
dalek rrot: ad90ddc | allison++ | ports/debian/patches/ (2 files):
Add a patch to debian packaging, fixes ia64 build failure in 2.9.1 packages.
 02:14
rrot: b626c46 | allison++ | ports/debian/control:
Updated generated Debian package file for 2.9.1 release.
rrot: 1d23436 | allison++ | ports/debian/changelog:
Add ia64 patch to Debian changelog file for 2.9.1 release.
02:31 whiteknight joined
whiteknight good evening, parrot 02:43
bluescreen hi 02:58
whiteknight hi bluescreen 03:03
bluescreen whiteknight: feel free to destroy all I did in the last commit 03:06
whiteknight why?
bluescreen if you consider it is wrong
I've taken a free license to implement Parrot_api_get_result 03:08
whiteknight I'll look at it. I never like to just undo things other people's work
thanks! Thats something we do need
bluescreen most of the failing test cases are because of the fact we prefix error messages with "Parrot VM:" 03:09
i didn't know either to change the cases or change the wording
whiteknight at this point we should change the wording to fit the tests
we want the behavior of the branch to match the behavior of master as much as possible to avoid deprecation delays 03:10
Once the branch merges, we can talk about changing/improving behavior
bluescreen good, that makes sense 03:11
whiteknight yeah, makes things easier 03:12
03:26 whiteknight left 04:02 petdance joined
dalek rrot: 50c6137 | jkeenan++ | / (2 files):
Adapt tests to cover all execution paths in the current, git-oriented version
 04:05
04:39 gbacon left 04:48 cotto joined
cotto ~~ 04:48
05:19 rurban_ joined 05:22 rurban left, rurban_ is now known as rurban 05:46 plobsing_ joined 05:47 plobsing left
Coke recommend way to generate a patch from git? 06:02
(that preserves author info)
*recommended 06:04
sorear git format-patch 06:16
Coke ... hopefully I can figure out how to use the damn thing. ;) 06:19
I would rant about the docs sucking here, but then dukeleto will scold me. hurm. 06:22
ah, there we go. 06:23
trac.parrot.org/parrot/ticket/1716 could use some testing. 06:26
06:33 petdance left
Coke sorear++ 06:34
06:39 luben left
dalek TT #1111 closed by coke++: cannot build on OSX/intel with --optimize 07:01
TT #1111: trac.parrot.org/parrot/ticket/1111
TT #890 closed by coke++: build failed on intel mac osx 10.5
TT #890: trac.parrot.org/parrot/ticket/890
cotto msg plobsing Any progress getting pirc's tickets migrated to github? 07:12
aloha OK. I'll deliver the message.
Coke ah, that's why those aren't closed yet. 07:15
07:43 allison left 07:59 fperrad joined
dalek TT #906 closed by plobsing++: [PIR] Assignment syntactic sugar restricted to destination parameters 08:53
TT #906: trac.parrot.org/parrot/ticket/906
TT #556 closed by plobsing++: setref on derived pmc aggregate class loses data on return 09:09
TT #556: trac.parrot.org/parrot/ticket/556
09:20 bacek left
dalek TT #587 closed by plobsing++: pirc 09:25
TT #587: trac.parrot.org/parrot/ticket/587
tadzik Parrot is a register-based process virtual machine designed to run dynamic languages efficiently. It uses just-in-time compilation for speed to reduce the interpretation overhead. 09:28
ā†‘ Wikipedia. It's not entirely true, is it?
plobsing_ not entirely. designed is such a strong word.
but also the JIT thing is a lie 09:29
tadzik that's what I meant
to be edited? 09:30
plobsing_ does it matter? 09:31
tadzik well, it's a bit misleading IMHO 09:32
but the JIT is there, isn't it? It's just not used
sorear no, the JIT was ripped out a couple months aog
IMHO "register-based" is not important enough to get a slot in the first sentence 09:33
plobsing_ jit is overrated imnsho. parrot isn't dominated by op-dispatch for most tasks, which is most what anything short of a very smart jit would help with. 09:34
s/most what/most of what/ 09:35
sorear exactly 09:39
dalek TT #12 closed by plobsing++: Task list for completing PIRC 10:29
TT #12: trac.parrot.org/parrot/ticket/12
TT #14 closed by plobsing++: fix braced arguments + nested macro expansion
TT #14: trac.parrot.org/parrot/ticket/14
TT #16 closed by plobsing++: fix memory leaks in pirc
TT #16: trac.parrot.org/parrot/ticket/16
TT #17 closed by plobsing++: Test all grammar rules of PIRC
TT #17: trac.parrot.org/parrot/ticket/17
TT #42 closed by plobsing++: make PIRC load-on-demand component
TT #42: trac.parrot.org/parrot/ticket/42
TT #55 closed by plobsing++: decorate all function parameters with ARGIN etc.
TT #55: trac.parrot.org/parrot/ticket/55
TT #163 closed by plobsing++: implement keyed multi types
TT #163: trac.parrot.org/parrot/ticket/163
TT #173 closed by plobsing++: lexicals not found in PIRC 10:30
TT #173: trac.parrot.org/parrot/ticket/173
TT #172 closed by plobsing++: get rid of C-strings for STRING_TYPE constants in PIRC
TT #172: trac.parrot.org/parrot/ticket/172
TT #229 closed by plobsing++: invoke build process of PIRC from main makefile
TT #229: trac.parrot.org/parrot/ticket/229
TT #576 closed by plobsing++: Add a configure step, --pirc
TT #576: trac.parrot.org/parrot/ticket/576
TT #1146 closed by plobsing++: same code works on windows, not on mac
TT #1146: trac.parrot.org/parrot/ticket/1146
TT #1387 closed by plobsing++: need --maintainer option for pirc
TT #1387: trac.parrot.org/parrot/ticket/1387
TT #1388 closed by plobsing++: pirc does not build with C++
TT #1388: trac.parrot.org/parrot/ticket/1388
10:39 bacek joined
dalek TT #173 reopened by bacek++: lexicals not found in PIRC 11:01
TT #173: trac.parrot.org/parrot/ticket/173
rrot: 1e89f76 | NotFound++ | src/dynext.c:
fix a check for nullness that must be STING_IS_NULLness in dynamic loading that was causing some segfaults
 11:22
rrot: cd7949a | NotFound++ | src/dynext.c:
protect a bit more Parrot_dyn_load_lib against mistakes like the one fixed in 1e89f76
 11:40
rrot/mappedbytearray: 43e7a36 | allison++ | src/gc/system.c:
Fix call in ifdef'd code for ia64 arch, signature of trace_mem_block was changed, but call not updated.
 11:49
rrot/mappedbytearray: ad90ddc | allison++ | ports/debian/patches/ (2 files):
Add a patch to debian packaging, fixes ia64 build failure in 2.9.1 packages.
rrot/mappedbytearray: b626c46 | allison++ | ports/debian/control:
Updated generated Debian package file for 2.9.1 release.
rrot/mappedbytearray: 1d23436 | allison++ | ports/debian/changelog:
Add ia64 patch to Debian changelog file for 2.9.1 release.
rrot/mappedbytearray: 50c6137 | jkeenan++ | / (2 files):
Adapt tests to cover all execution paths in the current, git-oriented version
rrot/mappedbytearray: 1e89f76 | NotFound++ | src/dynext.c:
fix a check for nullness that must be STING_IS_NULLness in dynamic loading that was causing some segfaults
rrot/mappedbytearray: cd7949a | NotFound++ | src/dynext.c:
protect a bit more Parrot_dyn_load_lib against mistakes like the one fixed in 1e89f76
rrot/mappedbytearray: 5f1b7a3 | NotFound++ | / (10 files):
Merge branch 'master' into mappedbytearray
rrot: 0c0b7c7 | bacek++ | src/gc/gc_ms2.c:
Don't try to remove "constant" PMC from PoinerArray.

They are very-very special and handled separately. This commit prevent breakage of Pointer_Array invariant about storing only pointers belong to it.		 12:42
rrot: 2f60133 | bacek++ | / (9 files):
Merge branch 'master' of github.com:parrot/parrot
TT #1875 created by bacek++: Deprecate "constant" PMCs. 12:54
TT #1875: trac.parrot.org/parrot/ticket/1875
13:06 whiteknight joined
whiteknight good morning, #parrot 13:09
Tene ... o.O 13:11
I never knew about gdb -tui
tadzik good afternoon whiteknight 13:14
whiteknight Tene: gdb -tui? What does it do? 13:15
Tene whiteknight: curses-based interface to gdb
whiteknight oh, nice 13:18
Tene you can switch to it from a running gdb session with: C-x a
bacek aloha, humans 13:19
seen moritz
aloha moritz was last seen in #parrot 16 hours 19 mins ago saying "fbrito: in what did you not have success?".
13:20 rurban_ joined
whiteknight bacek: I agree with you. constant PMCs are not user-facing and don't need a deprecation cycle 13:20
bacek msg moritz Can you retest rakudo on bleeding-edge parrot? At least RELEASE_2_10_1-523-g2f60133
aloha OK. I'll deliver the message.
bacek whiteknight, yes. I personally lost quite few hours debugging failures related to "constant" pmcs. 13:21
whiteknight bacek: that sucks. Your time is too valuable to be doing that kind of crap
rip it out and don't look back
bacek I would like to. But it's not only my call. 13:22
13:22 rurban left, rurban_ is now known as rurban
bacek I don't know who designed them. And may be (unlikely) there is other reasons behind them. 13:23
dalek rrot/key_func_renames: cf5af19 | (Lļæ½o Grange)++ | / (18 files):
All public function in src/key.c renamed from key_* to Parrot_key_*.
 13:24
NotFound whiteknight: have you looked at the mappedbytearray branch? 13:26
whiteknight NotFound: a little bit 13:28
NotFound whiteknight: I'm thinking I can marge it to master now, declaring the PMC as experimental, to get more exposure. 13:29
whiteknight NotFound: do you have tests and everything? I wouldn't merge it to master without a few tests
bacek NotFound, what about changing FileHandle PMC to support mmap instead? 13:30
NotFound There are some tests, mainly to check that it doesn't break anything.
bacek: I think we already have too much multi-purpose PMC with too much conditional.
bacek NotFound, yes. But in this case it's improvements for FileHandle guts. Which is hidden behind public API. 13:31
NotFound A MappedHandle and/or a MappedIterator may be better, IMO. 13:32
bacek actually, split FileHandle to File and FileIterator will be better
NotFound, +1
(For Iterator)
NotFound Also, add a method to MappedByteArray to use an already opened FileHandle. 13:33
The current open is for using less resources, closing the handles that are not needed after the maping is done. 13:34
But I think is better to have more feedback before adding much more things. That's why I want to merge. 13:35
13:41 Psyche^ joined 13:42 Patterner left, Psyche^ is now known as Patterner
moderator Parrot 2.10.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Long live Git! github.com/parrot/parrot | git clone git://github.com/parrot/parrot.git | Please test rakudo with bleeding edge parrot! 13:52
bacek ok, I;ve got first clean run of "make spectest" in rakudo with latest parrot. 13:53
If anyone can also test it it will be very helpful. 13:54
Good night, folks
13:57 Themeruta joined 13:58 NotFound left
whiteknight FileIterator is too narrow. We need HandleIterator because we have many handle types 14:05
we should be able to iterate a socket as easily as a FileHandle, as easily as a StringHandle
dalek rrot: 8aa40c9 | Whiteknight++ | / (18 files):
Merge branch 'key_func_renames'
 14:10
rrot/embed_api: 7e25808 | (Mariano Wahlmann)++ | / (5 files):
modified pbc_* to make them compile with the embed_api
 14:23
rrot/embed_api: 9f7a94c | Whiteknight++ | src/embed/ (2 files):
make sure to set :main() args in the iglobals array. This makes ops2c run again. The build appears to complete. Also, fix a warning in EMBED_API_CALLIN where we were doing an assignment in the if() condition without parenthesis
rrot/embed_api: 7bda53a | (Mariano Wahlmann)++ | src/embed/embed_private.h:
Clean up api_jmp_buf after an API call
rrot/embed_api: 7c837f4 | (Mariano Wahlmann)++ | src/ (2 files):
no more core dumps
rrot/embed_api: 9addea1 | Whiteknight++ | include/parrot/interpreter.h:
small cleanups in interpreter.h. No functional changes
rrot/embed_api: 51acd98 | Whiteknight++ | src/ (2 files):
I don't know why it's ever a Null PMC, but if we have a Null config hash don't attempt to do anything with it
rrot/embed_api: be50d63 | bluescreen++ | / (9 files):
Implemented first version of Parrot_get_result, now a lot of failing test are because we're printing 'PARROT VM:' at the beggining of the error message. Also modified embed_private.h to remove the interp->final_error = NULL because that caused the error to be truncated each api call and get_result was truncating it.
14:54 nwellnhof joined 14:59 kid51 joined 15:08 jan left 15:15 lucian joined 15:34 kid51 left 15:54 Themeruta is now known as NotFound
Coke aloha, msg plobsing 556 was closed saying "this segfaults" - we need a ticket to track that segfault, yes? 16:07
aloha Coke: OK. I'll deliver the message.
Coke I am also curious why the description was changed on those tickets seventy times, but am merely curious. 16:10
16:11 jan joined
Coke aloha msg bacek - why doesn't aloha count "aloha, " as an unidle for the purposes of letting me know I have messages? 16:11
aloha Coke: OK. I'll deliver the message.
16:22 lucian_ joined 16:27 lucian left, PacoLinux left 16:30 PacoLinux joined
nopaste "bluescreen" at 192.168.1.3 pasted "Why is IMCC_warning returning when imcc_warn flag is set ?" (8 lines) at nopaste.snit.ch/26221 17:11
17:19 lidden joined 17:21 kid51 joined 17:30 TypeNameHere_____ left 17:41 theory joined
dalek a: 03a2fb5 | fperrad++ | dynext/pmc/lua (2 files):
s/key_append/Parrot_key_append/
 17:48
fperrad bacek, now, lua segfaults, for example : 17:51
$ parrot lua.pbc -e "print 'ok'"
dalek rrot: 0e34170 | nwellnhof++ | / (28 files):
Merge branch 'master' into string_checks
 17:56
rrot: 8e00918 | nwellnhof++ | / (42 files):
Merge branch 'string_checks'
18:01 hercynium left
dalek rrot/tt532_headerizer_refactor: ab904c8 | jkeenan++ | t/tools/dev/headerizer/ (2 files):
Add tests for case where sourcefile is a PMC. Add sample file used in illustrating that case.
 18:10
rrot: 19993a9 | nwellnhof++ | t/pmc/bytebuffer.t:
Refix typo reverted by previous merge
 18:12
TT #1862 closed by jkeenan++: Configure.pl should actually test for gcc -fvisibility=hidden support 18:15
TT #1862: trac.parrot.org/parrot/ticket/1862
18:54 TypeNameHere_____ joined 19:06 kid51 left
plobsing_ ~~ 19:06
19:06 plobsing_ is now known as plobsing
plobsing ~~ 19:06
silly bots 19:07
msg cotto sorry for the holdup on pirc tickets. I had a tuits shortage. it is done now. 19:08
aloha OK. I'll deliver the message.
bluescreen pblosing: do you know why IMCC_warnings is returning when imcc_warn flag is on 19:15
looks like twisted logic to me... Even more if I run parrot -w some.pir it removes warnings
plobsing 1) it is imcc, it doesn't have to make sense. 2) look at IMCC_warning()'s uses. they are all errors. IMCC_warnings() is akin to perror(). 19:17
3) warnings are underused in general, so this went undetected for a long time. 19:18
bluescreen hmmm... and then errors are being printed twice
plobsing but they aren't somehow. yeah. IMCC is weird.
bluescreen so we live with that? is there any TT#?
plobsing bluescreen: is it hurting anyone really badly? 19:19
bluescreen my mental sanity....
plobsing only because you know about it. don't look into compilers/imcc if you still value your sanity. 19:20
bluescreen lol... I came to that debugging some failures in make test for the embed_api 19:21
plobsing I'm in favour of leaving it. it mostly works. 19:23
bluescreen and pray for lorito?
plobsing just don't use warnings on pir you don't even know compiles
19:28 fbrito joined
fbrito whiteknight: Hello. Are you there? 19:28
19:28 hercynium joined
plobsing msg Coke TT #556 has a user doing something they shouldn't trying to get behaviour that doesn't exist. if we need a TT, we also need a pebkac flag. I'm also looking into simply ripping out the functionality that segfaults (it doesn't seem terribly useful). 19:33
aloha OK. I'll deliver the message.
dukeleto 'ello 19:36
fbrito whiteknight: When you come back please tell me :) I am Fernando, from Google Code In 19:38
or if there is anyone here in charge of Google Code In, please tell me :D 19:42
Coke plobsing: parrot shouldn't segfault, even if the user is an idiot. 19:46
plobsing which is why I'm looking to take away the functionality 19:48
Coke that works. 19:49
dukeleto fbrito: welcome! How is Code-In? 19:52
fbrito awesome! :D
dukeleto is the org admin for Parrot Foundation and The Perl Foundation for Google Code-In
fbrito: what have you been working on? I have not been keeping track of GCI stuff over the Thanksgiving holiday 19:53
fbrito I wrote NumMatrix2D examples to Parrot-Linear-Algebra
And now I am working writing unit tests for the same library 19:54
dukeleto fbrito: very cool! That is very advanced stuff! Have you studied Linear Algebra in school already? 19:55
fbrito Yes and No. I am from Brazil and math here (on schools) is not that advanced 19:56
But I spent 1 year in Germany (as an exchange student)
And the classes there were pretty advanced 19:57
dukeleto fbrito: What kind of math classes did you take there? Have you studied "Gaussian Elimination" in school? 19:58
fbrito Yes :D 19:59
But the classes were in German and when I arrived there I couldn't understand a word, ahhaha
But after some months things became better and I could learn lots of things
dukeleto fbrito: I studied for a summer in Germany and took classes in German, but just the basics of the language, along with some history and geography 20:00
fbrito: I am very impressed to see GCI students doing awesome stuff on Parrot-Linear-Algebra
fbrito: how did you hear about GCI? 20:01
fbrito I have some friends that worked on Google Summer of Code
And they told me about GCI 20:04
bacek Good morning, humans. 20:17
fperrad, ping
aloha, nopaste?
aloha bacek: nopaste is is nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl)
nopaste "bacek" at 192.168.1.3 pasted "fperrad, this is the reason for lua segaults. Mixture of "constant" and normal PMCs." (15 lines) at nopaste.snit.ch/26222 20:18
dukeleto fbrito: do you know how to send people messages with the aloha bot? 20:20
aloha msg fbrito this is how you send a message 20:21
aloha dukeleto: OK. I'll deliver the message.
dukeleto: Okay.
fbrito oh, interesting 20:23
dukeleto fbrito: when you send a message, the bot waits for that person to say something in the channel, then send that use a private message with your message 20:33
20:34 nwellnhof left
dukeleto s/that use/that user/ 20:34
fbrito aloha msg fbrito test
aloha fbrito: OK. I'll deliver the message.
fbrito Really?
Nice :D
dalek rrot: 04ab174 | bacek++ | src/pmc/default.pmc:
Create constant metadata for constant PMCs in default PMC.

This is required to preserve GC invariant when "constant" objects are not marked.		 20:37
fbrito Hm 20:38
Coke plobsing: if you're looking for things to rip out, what about "parrot -E" ?
fbrito dukeleto: May I ask something about GCI?
dukeleto: Are you allowed to create new tasks now, in the middle of the contest? 20:43
TypeNameHere_____ aloha msg fbrito yes 21:05
aloha TypeNameHere_____: OK. I'll deliver the message.
fbrito Because there is still lots of space for more unit tests but no task to that 21:06
Coke tries to install something off macports. 21:09
4 hours later...
TypeNameHere_____ hmmm... 21:10
aloha msg aloha Aloha!
aloha TypeNameHere_____: OK. I'll deliver the message.
21:20 rurban_ joined
whiteknight fbrito: ping 21:21
fbrito whiteknight: pong :)
whiteknight I just got in. I'll look at your tests now
21:22 rurban left, rurban_ is now known as rurban
fbrito Ok, thank you. I also have a question to you about the "fill" method with 3 params 21:22
Take a look at that, please: pastie.org/1330731 21:25
Reading the docs (whiteknight.github.com/parrot-linea...ix2d.html) I understood that fill(4, 4, 1) in a 2x2 matrix would resize it to 4,4 and fill it with 1 values, no? 21:26
whiteknight I need to re-read the source to make sure. I don't remember what order the args go in 21:32
fbrito I will take a look in the source-code too 21:35
whiteknight fbrito: fill(1, 4, 4) 21:36
that should resize it and fill it with 1's
fbrito Ah, ok. Should I fork the documentation fixing that? 21:37
whiteknight no, that's okay. I can fix the documentation 21:38
you're welcome to fix it, but it's not part of GCI :)
dukeleto fbrito: yes, tasks can be created at any time 21:39
fbrito I can do that, no problem. I will also add to the documentation that item_at also can receive 3 params to change the value
dukeleto fbrito: if you have ideas for parrot tasks, you can email them to parrot-dev and if we think it makes a good task, a Parrot mentor will add it. How about that? 21:40
whiteknight: be on the lookout for current tasks generating new tasks :)
whiteknight I'm always looking for excuses for new tasks :)
dukeleto whiteknight: I was offline during the holiday, so I have not caught up on GCI stuff
whiteknight dukeleto: that's okay. It's a tricky time 21:41
dukeleto whiteknight: the melange site is wonky. They don't differentiate between "needs action by student" or "needs action by mentor" 21:44
whiteknight: so I have wasted a huge amount of time looking at tasks that "need action" to find out that it requires student action
whiteknight the melange site is wonky for several reasons 21:45
dukeleto whiteknight: so i have resolved to only look at them every few days. If you need a new task approved, just ping me and I will approve it
whiteknight ok
fbrito whiteknight: should I write more tests for my task? I felt bad because it had a lot of "similar" tests (ctrl+c, ctrl+v and change 2 or 3 values) but it was needed
whiteknight fbrito: I need to look at all the tests you wrote 21:46
fbrito Ok, sure. You saw the patch that I attached, right? 21:47
whiteknight I have a lot to look at
fbrito Ok, sorry :D Take your time
whiteknight Give me about an hour to read everything over
fbrito CGI makes me really anxious, ahhaha. sorry
GCI* 21:48
dalek rrot/nwellnhof/gc_dynamic_threshold: b5cc088 | nwellnhof++ | / (10 files):
[gc] Add dynamic threshold to GC MS2

memory_used now also accounts for attribute and other fixed size storage and ignores constant objects. The old code could lead to a lot more memory being used than accounted for, so rather low thresholds could still work very well. To address this, the default dynamic GC threshold has been increased to 75%. Note that these changes have not been applied to the old GC. It still uses the old accounting.
Also cleanup string compacting code.		 21:49
whiteknight fbrito: yeah, it's a fun new thing. You're doing great though
22:10 nwellnhof joined
whiteknight fbrito: I don't have your changes. Did you delete your repo? 22:10
fbrito yes. please have look at the git formatted patch that I attached 22:11
whiteknight oh, in the ticket?
fbrito yes
whiteknight okay, sorry
fbrito I had problems again with indentation. there were like 8 or 9 commits in my fork just fixing spaces and tabs
so i deleted it :P
dukeleto Has anybody ever played around with sourceforge.net/projects/parakeet/ ? 22:12
whiteknight that's okay. I can fix indentation problems easily
fbrito but in the patch everything is ok with indentation (i hope)
dukeleto: I have not 22:13
dukeleto This conference sounds like it will be very interesting for Parrot hackers: www.cs.technion.ac.il/~erez/vee11/V..._Page.html 22:15
And this looks really cool: www.managedruntime.org/ 22:16
22:18 theory left
whiteknight blah. I think I hate git am 22:18
lucian_ whiteknight: i think it has that effect on people :) 22:19
22:19 lucian_ is now known as lucian
lucian prefers hg 22:19
dukeleto whiteknight: what is your issue? 22:20
lucian: you can use hg as a front end to a git repo 22:21
whiteknight dukeleto: I haven't learned it well enough, and the last few times I have tried have been disasterous
dukeleto whiteknight: tell me what goes wrong, because that is stuff that I need to add to our git docs
whiteknight like "start deleting things out of .git/ in anger" disasterous
dukeleto whiteknight: oh boy. Don't manually delete files in .git! I will definitely add that to the docs :) 22:22
lucian dukeleto: yeah, that too
dukeleto: but you still need to know git
whiteknight: there's a nice paper about the internals of git
it's useful
whiteknight fbrito: tests look good. I just committed them and will push in a minute 22:23
Task successful
fbrito Ok, thank you very much :)
dukeleto stands back in awe as the GCI train goes full steam ahead
whiteknight it is quite inspiring 22:24
dukeleto whiteknight: stackoverflow.com/questions/4294438...it-and-how <-- have you seen this? 22:26
whiteknight no 22:27
I'll put in an answer for that one later
22:31 lidden left, fperrad left
whiteknight I have to touch up some GCI-related PLA stuff first, before it slips my mind 22:32
lucian dukeleto: well, python isn't a supported language at all 22:33
i've just been playing with pynie and it's bitrotted. doesn't even run/build
dukeleto lucian: there is much interest in making pynie active again. Perhaps you can start it up. 22:36
lucian: everyone in here will help you. We have good docs about deprecations now on our wiki. 22:37
lucian dukeleto: yeah, i'd like to. i'm just a bit short on time right now
i'll try and squeeze in a few hours/week
dukeleto: as a python dev, the main stumbling block is the lack of objects 22:39
if i can get pynie running and boostrapped up to object + parrot interop, other python devs will be much more likely to contribute
(we have a possibly irrational aversion towards all things perl and NQP is just too close) 22:40
plobsing not close enough imho
22:41 bacek left
plobsing lucian: objects may be hard to get working pythonically. cardinal ran into some issues trying to do similar things. 22:41
we do however have a planned object system refactor in december that should help 22:42
dalek rrot: e72cca0 | (Gerd Pokorra)++ | NEWS:
remove some whitespace
 22:43
NotFound If you want to bootstrap, you can write a python compiler that generates pir, and bootstraping it with any other python. 22:44
lucian NotFound: yeah, that's also a choice
dukeleto lucian: chromatic (and myself) plan on giving Parrot a proper meta object protocol
lucian but i guess I'd rather reuse parrot-side work
dukeleto: plobsing: I'm not confident that it can be abstract enough and useful enough at the same time
object systems seem just too different to me 22:45
dukeleto lucian: parrots meta object protocol and objects are a bit wonky, imho
lucian: i think a proper meta object protocol can be used to implement any object oriented system
lucian dukeleto: for example, would that support multiple dispatch, a la CLOS?
dukeleto lucian: yes
lucian ok, then it's probably general enough 22:46
dukeleto lucian: since Parrot support it now, are newer MOP must also support it
lucian dukeleto: i know you can write a multi dispatch object system on parrot, but will it interoperate ok with other objects?
dukeleto lucian: these are good questions to ask. You might want to send an email to parrot-dev explaining what you need to hack on Pynie
lucian dukeleto: i'm not sure what I need really, I don't know parrot well enough
objects in python are basically dicts with a few extra tricks 22:47
Tene lucian: Anything that generates PAST, you can re-use the parrot toolchain
lucian everything else in CPython is an optinisation
22:47 bacek joined
lucian Tene: right 22:47
Tene lucian: Would Perl 6 grammars be off-limits too?
lucian Tene: i think grammars are arcane enough for it not to matter 22:48
Tene lucian: You could write a standard parrot-style compiler, but with the NQP parts switched out for some minimal proto-python.
lucian Tene: yeah, but that's more work
i can stand writing some perl, but it won't be quite so hackable by python devs
i guess i can boostrap it with nqp and figure it out later 22:49
dalek TT #173 closed by bacek++: lexicals not found in PIRC
TT #173: trac.parrot.org/parrot/ticket/173
lucian Tene: but you gave me an idea. I could try to reuse CPython's parser + AST and transform that to PAST
Tene lucian: So you want to bootstrap as much as you can, so you have a minimal proto-python in NQP or whatever, and implement the rest in that is fine
lucian: That's also a workable approach
lucian Tene: bootstrapping is just a bit weird in parrot 22:50
Tene lucian: s/ in parrot// :)
NotFound I've been able to create a language and bootstrap it in two monts. Weird is not the word I'll choose. 23:02
lucian NotFound: ok then, unusual. and I haven't done it before, either 23:04
NotFound If a lot of pyhton devs can'r beat me, too bad :P
lucian NotFound: are there a lot? afaict it's just me :P
NotFound lucian: there is a foundation or something, isn't it? ;)
lucian NotFound: i doubt they care about parrot 23:07
NotFound: it's the (very) backup plan for a future runtime 23:08
NotFound Maybe the problem is too much targets: cpython, jpython, ironpython...
lucian NotFound: it's not a bug, it's a feature :P 23:10
NotFound: there are two good future C-level platforms: CPython+llvm and PyPy
PyPy has the extra bonus of being metacircular and with the potential to obsolete jython and ironpython
NotFound I think allison said something about using PyPy to target parrot.
lucian NotFound: it doesn't quite work 23:11
but a lot of PyPy work will be useful to pynie, since they rewrote a lot of stuff to pure python
dalek rrot-linear-algebra: 33272b0 | rofflwaffls++ | t/methods/complexmatrix2d/ (10 files):
Added ComplexMatrix2D test cases (GCI).
 23:15
rrot-linear-algebra: 1de7876 | Whiteknight++ | pmcs/examples/nummatrix2d.md:
nothing
rrot-linear-algebra: 44f605c | Whiteknight++ | /:
Merge branch 'master' of github.com:Whiteknight/parrot-linear-algebra
rrot-linear-algebra: b8016b8 | Whiteknight++ | t/methods/nummatrix2d/ (6 files):
Apply new tests from Fernando Brito as a result of his GCI work. Adds substantia test coverage for NumMatrix2D
rrot-linear-algebra: fb9ebd0 | Whiteknight++ | t/methods/complexmatrix2d/ (3 files):
Fix a gemm test which got borked somehow. Fix whitespace
rrot-linear-algebra: 1a72e1f | Whiteknight++ | t/methods/complexmatrix2d/ (2 files):
some fixes. TODO one test that requires attention later
23:17 lucian left
whiteknight dukeleto: can you look at github.com/parrot/parrot/pull/7? 23:23
It looks decent to me, but I would like a second pair of git-expert eyes on it to make sure it's good
23:34 PacoLinux left
NotFound I've used msysgit, it allowed me to do the win32 part of MappedByteArray 23:40
Don't pushed with it, though. Just copied the patch to a linux box.
bluescreen whiteknight: ping 23:47
dukeleto whiteknight: in the pull request, the debian line is wrong 23:50
whiteknight: should be "apt-get install git-core"
whiteknight: the git directions don't look exactly right, but they are a good start 23:51
whiteknight: we should create a task to change our readme into markdown, so it is displayed nicely on github
dukeleto goes out to dinner
23:58 nwellnhof left
fbrito if you are going to create such a task (readme to markdown), please tell me :) 23:59

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!