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!
Set by moderator on 28 November 2010.
dukeleto fbrito: you are ranked 3rd in GCI points currently! www.google-melange.com/gci/program/...le/gci2010 00:00
fbrito oh yeah! :D
whiteknight fbrito: with our help, you could go all the way!
cotto_work Cobol? 00:01
fbrito actually I am the 2nd. just waiting this task to get approved: www.google-melange.com/gci/task/sho...9035408428
whiteknight COBOL: Commit often, but only LOLCODE 00:02
dalek rrot: 8e80030 | fbrito++ | / (6 files):
Renaming Parrot_get_ctx_HLL_namespace() to Parrot_hll_get_ctx_HLL_namespace()
rrot: 35090a7 | fbrito++ | / (6 files):
Renaming Parrot_get_HLL_namespace() to Parrot_hll_get_HLL_namespace()
rrot: f7134de | fbrito++ | / (4 files):
Renaming Parrot_regenerate_HLL_namespaces() to Parrot_hll_regenerate_HLL_namespaces()
rrot: f89a69a | fbrito++ | / (5 files):
Renaming Parrot_get_HLL_name() to Parrot_hll_get_HLL_name()
rrot: 99bcc83 | fbrito++ | / (4 files):
Renaming Parrot_get_HLL_id() to Parrot_hll_get_HLL_id()
rrot: 5a37489 | Whiteknight++ | / (35 files):
merge in changes from fbrito++. GCI task to rename functions in src/hll.c to match naming convnetions
cotto_work Nice. The "Yacc is dead" paper quotes TimToady. 00:04
whiteknight fbrito: I didn't close that task on accident. Can you change it back to NeedsReview?
cotto_work: link?
dukeleto cotto_work: which paper?
cotto_work arxiv.org/abs/1010.5023
posted on LtU and Hacker News recently 00:05
dukeleto cotto_work: cool, thanks
fbrito whiteknight: done 00:06
whiteknight sorry about the confusion
closed
WYSIWYGIYULR(k) 00:08
I like that acronym
GeJ whiteknight: I'm stuck in a living Hell : I have a 2 CPU server with a lod in the two digits; we're changing the office's IP address tomorrow and I've got to review a bunch of servers to check security policies where the IP might have been hardcoded, oh and tomorrow in the morning I'm going under the knife and I have no idea how long until I get back to work. It's going to be a long day. 00:09
dukeleto fbrito: nice work! You are second!
GeJ Oh, and my manager still refuses to hire an Office Linebacker.
dukeleto GeJ: i would take a vacation. 00:10
whiteknight GeJ: Under the knife? Nothing serious I hope
fbrito dukeleto: thank you very much :D
00:10 davidfetter left
GeJ whiteknight: nothing really serious, it's just that it's not the good time. 00:11
whiteknight GeJ: yeah, I definitely understand.
cotto_work: I was talking about making a pure-parrot LALR parser-generator earlier today, but reading this paper is starting to convince me that a derivative parser like this might be better 00:14
the term "parser combinators" is making me salivate 00:15
anybody here have an ACM subscription? 00:18
this one damn paper is from 1964 and STILL costs money to view it
cotto_work whiteknight: you can ask allison when she's around 00:19
dukeleto whiteknight: whatcha need?
whiteknight dukeleto: portal.acm.org/citation.cfm?id=321249
dukeleto has a library card from Cornell
cotto_work nice
dukeleto attempts to figure out how to tell Cornell to tell ACM to give me the paper 00:20
whiteknight no worries, I'll msg allison 00:22
cotto_work I wonder if we could get a subscription for Parrot or PaFo. 00:24
whiteknight that would be nice 00:26
cotto_work msg cotto look into a Parrot/PaFo ACM subscription 00:29
aloha OK. I'll deliver the message.
00:33 preflex left 00:35 preflex joined
dalek rrot: e16b53a | jkeenan++ | README:
Add instructions for obtaining git for Mac OS X. Remove outdated language re
 00:35
whiteknight msg allison if you have a moment, I would really love to get a look at this ACM paper: portal.acm.org/citation.cfm?id=321249 00:38
aloha OK. I'll deliver the message.
00:42 sECuRE left 00:57 kid51 joined 00:58 dmalcolm left, kid51 is now known as kid51_at_dinner
fbrito I saw you guys discussing something about FileHandle... this task www.google-melange.com/gci/task/sho...9104579996 is still valid right? 00:59
s/right/,right/ 01:01
01:03 sECuRE joined
whiteknight fbrito: yes 01:04
fbrito: I have more tasks like that
fbrito ok :D 01:06
dalek rrot: d2a80a7 | nwellnhof++ | t/pmc/stringiterator.t:
[t] String iterator get_*_keyed tests for different encodings
 01:10
fbrito hm, I have questions about this task: www.google-melange.com/gci/task/sho...9104579996 01:13
whiteknight what questions? 01:16
fbrito in step 3: i think that get_fd in stringhandle.pmc already do that 01:17
whiteknight okay. that makes it easy for you 01:18
just double-check it
and if tests for it already exist, that's another check for you
fbrito ok 01:20
wow, those tests written in .t are tricky 01:25
i found is(), ok() and nok() method, but I am looking for something like is_not(-1). can I use test_not(-1)?
whiteknight not()
dukeleto fbrito: isnt() 01:33
cotto ~~ 01:38
isnt_not_not() 01:39
dalek rrot: 856c6e5 | nwellnhof++ | t/op/string_cs.t:
[t] Add tests for conversion to UCS-4
 01:40
whiteknight isnt(), that's what it is 01:42
dalek rrot/embed_api: 7596edd | bluescreen++ | src/main.c:
Make most failing test pass by commeting the error printing routine
 01:44
rrot/embed_api: 372f8f7 | bluescreen++ | src/embed/api.c:
Make bounds runcore available
rrot/embed_api: 704e3c4 | bluescreen++ | src/embed/api.c:
Made another test case pass by putting debug information. Also removed some dependency with embed.c
rrot/embed_api: f686051 | bluescreen++ | include/parrot/api.h:
Made profiling test pass
fbrito i moved get_fd() from FileHandle to Handle, but I'm getting this: 01:53
Method 'get_fd' not found for invocant of class 'Socket'
whiteknight that's weird 01:55
fbrito ah, I think I forgot something really important
wait :D
whiteknight .....
01:58 bluescreen left 01:59 nwellnhof joined
dalek rrot: 4c93dc1 | nwellnhof++ | / (13 files):
[str] Replace STRING_iter_set_position with STRING_iter_skip
 02:02
dukeleto fbrito: Socket inherits from FileHandle 02:11
whiteknight Okay, I've created 3 tasks today. I need 2 more 02:13
dalek rrot: 1f7b157 | nwellnhof++ | t/op/stringu.t:
[t] Test ord with different Unicode encodings
 02:14
whiteknight dukeleto: trac.parrot.org/parrot/wiki/APIFuncRenaming 02:15
02:15 bluescreen joined
fbrito dukeleto: really? I assumed that it inherits from Handle 02:16
"This method should be moved to the Handle class, and inherited by all Handle types (StringHandle and Socket)." 02:17
dalek rrot: 8159152 | nwellnhof++ | t/op/stringu.t:
[t] Test another malformed UTF-8 string
fbrito isn't the task about moving get_fd from FileHandle to Handle so that Socket can "use" it? 02:18
02:22 bacek left
fbrito whiteknight: on the task, step 3 and 4: do you really mean StringHandle, right? Not FileHandle? 02:23
whiteknight let me look
FileHandle has a valid file descriptor
StringHandle does not
so get_fd on StringHandle should return -1 02:24
fbrito ahh, ok
dalek rrot: f9f96c9 | nwellnhof++ | t/op/stringu.t:
[t] Test illegal UTF-8 escape char
 02:25
tracwiki: v18 | whiteknight++ | BuildTaskList 02:26
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
fbrito but I still can't understand this: (23:11:06) dukeleto: fbrito: Socket inherits from FileHandle 02:27
nwellnhof Socket doesn't inherit from FileHandle
dukeleto fbrito: perhaps i am wrong 02:28
dukeleto can actually be wrong.
nwellnhof pmclass Socket extends Handle
fbrito ah, ok :D. no problem
cotto Socket and FileHandle inherit from Handle
dukeleto nwellnhof++
fbrito so I am back to my original problem. I moved get_fd() to Handle but it is saying that Socket doesn't have this method... 02:31
cotto fbrito, can you nopaste? 02:32
fbrito hehe, I know that this question will sound really stupid but I have to ask: when Socket inherits from Handle, it "automatically" make all methods available, right?
02:32 kid51_at_dinner is now known as kid51
cotto yes 02:33
and any VTABLE functions that Socket doesn't implement directly come from Handle
whiteknight fbrito: commit what you have and I can look at the code 02:34
fbrito ok, thank you very much. just a second... I wrote some methods just returning some int values on Handle, FileHandle, StringHandle and Socket so I can try some tests on my own :D 02:36
whiteknight ok 02:39
dalek rrot: 7df9d84 | nwellnhof++ | t/op/stringu.t:
[t] More tests for illegal char escapes
 02:42
dukeleto nwellnhof++ # i like tests 02:43
nwellnhof my plan is to get the test coverage of the code in src/string/encoding/ to a reasonable level. 02:44
the upcase_first and downcase_first string functions are completely unused in Parrot. 02:45
dukeleto nwellnhof: do any HLLs use them? what about in examples/ ? 02:48
nwellnhof we don't have an upcase_first opcode and there's no other code that makes use of these functions. 02:49
bacek_at_work nwellnhof, add string dynops for it? 02:51
fbrito NOOOOOOOOOOOO. LOL. I was always running "make" and trying to run individual tests with the parrot installed on my system, and not the one that I had just compiled 02:53
nwellnhof bacek: yes, why not. but it would be nice if there was a HLL that would use them. 02:54
bacek_at_work nwellnhof, rakudo?
02:57 bluescreen left
nwellnhof yes, rakudo could use it. it's needed for proper Unicode support. 02:58
Infinoid whiteknight: ping 03:03
whiteknight Infinoid: pong 03:07
Infinoid whiteknight: I'm reading over the IO task list, and had a couple questions.
Sounds like the guts are moving into .c files instead of PMCs 03:08
whiteknight fbrito: ah, that's one lesson you have to learn the hard way
Infinoid but ... what tracks the open handle at the API level? will it be referred to by fd (int), or some struct, or what?
whiteknight Infinoid: much of that task list hasn't changed since last time you were working on it
Infinoid: src/pmc/handle.pmc is the abstract parent type. It has a PIOHANDLE attribute, which is platform-defined 03:09
Infinoid Ok 03:10
I think I'm mostly looking at "Change IO API to not use PCCINVOKE calls. Method calls on IO PMCs should call the API functions, not the other way around"
whiteknight so that gets passed around the functions in src/io/*
Infinoid Which sounds great, but I'm wondering what it will look like
Ok, so it's a PIOHANDLE typedef?
whiteknight Infinoid: the open/close opcodes used to call the Parrot_io_ functions, which turned around and called the methods on the various PMC types 03:11
but the open/close opcodes are gone, and we always call the methods directly now
if you look at Parrot_io_close() function, you see the fallback that still attempts to call a method
...which is stupid, because we called the method to get there in the first place 03:12
Infinoid Yeah :)
Though I suppose it makes polymorphism for things like StringHandle happen without needing a big series of if statements
whiteknight If anybody ever exercised that code path, I suspect the infinite loops would be wonderful
Infinoid (I'm assuming stringhandle doesn't call an underlying close) 03:13
whiteknight We aren't being intelligent right now about how the methods are implemented
the internals all need to be moved around to account for the new architecture
it used to be opcode->API->Methods, now it's Methods->API, but the API hasn't really changed much 03:14
Infinoid If you have an understanding of the new architecture, I'd love to bug you occasionally with a few more questions 03:15
whiteknight that would be fine
cotto is the new architect though, so I'm sure he has some input
Infinoid Cool
03:19 preflex left 03:21 preflex joined
dalek rrot/tt532_headerizer_refactor: df60083 | jkeenan++ | / (4 files):
Refactor some code from within

into Parrot::Headerizer::Functions::add_headerizer_markers(). Code itself is working, but not correctly tested yet in 01_functions.t		 03:35
rrot/tt532_headerizer_refactor: 6cc3b2c | jkeenan++ | lib/Parrot/Headerizer/ (2 files):
Delete commented-out code.
 03:39
kid51 So, in the brave new world of git, do we have anything to automatically identify files which should be added to MANIFEST or MANIFEST.SKIP (like the former tools/dev/mk_manifest_and_skip.pl)? 03:41
plobsing kid51: it works the same way (iirc dukeleto++ updated the tools) 03:47
kid51 Hmm, I guess it's not working in my branch because my branch dates from the long-ago-Subversion days. 03:48
fbrito Yes :D another Parrot GCI task completed 03:49
cotto fbrito++ 03:50
fbrito right now I have 2 pull requests pending 03:52
whiteknight I'm going to have to look at it tomorrow morning. Right now I need sleep
fbrito++
fbrito ok, no problem :)
whiteknight goodnight
03:52 whiteknight left
fbrito good night 03:56
ops, too late
dukeleto fbrito: i will take a look 04:02
kid51: merge master into your branch "git checkout $branch && git merge master" to get the new git-aware mk_manifest_and_skip.pl 04:03
fbrito dukeleto: ok, but only do that if you really want. I have no problem waiting until tomorrow :) 04:05
dalek rrot: 31f6c2e | jkeenan++ | config/gen/makefiles/root.in:
Apply patch submitted by fbrito++ in trac.parrot.org/parrot/ticket/1876. Have

Configure.pl and 'make clean' is intended to remove files created specifically by
  'make'.		 04:09
kid51 fbrito: trac.parrot.org/parrot/ticket/1876#comment:5 04:13
fbrito ok! thank you :D 04:14
kid51 fbrito: Prepare a patch to apply your data to CREDITS
fbrito github.com/parrot/parrot/commit/05...a46eecd30d 04:15
dukeleto fbrito: i see only 1 pull request 04:16
fbrito I have just closed the one related with the Makefile 04:17
dukeleto fbrito: ok
fbrito: i am checking the other pull request
fbrito this is the task link: www.google-melange.com/gci/task/sho...9104579996 04:18
step 3 and 4 were already implemented
dukeleto fbrito: i am compiling and testing now 04:22
kid51 fbrito: Okay, I was looking at CREDITS in a branch. I see that you are already listed in CREDITS in master.
CLA? 04:23
aloha, CLA?
aloha kid51: I have no idea.
kid51 stupid bot
fbrito: Should you ultimately wish to apply for a commit bit, please fill out a CLA: www.parrot.org/files/parrot_cla.pdf 04:27
kid51 must sleep
04:28 kid51 left
fbrito bye 04:28
cotto aloha, cla is www.parrot.org/files/parrot_cla.pdf
aloha cotto: Okay.
fbrito argh, again I was too late
dukeleto fbrito: compiled ok. running a smoke test now 04:29
fbrito: looks good! 04:45
fbrito oh yes:D
sorear I think you'll have to wait a bit 04:46
depending on where you are
fbrito what? 04:47
sorear the CLA
fbrito I am not sure but I don't think that I should apply for a commit bit 04:49
dalek rrot: 8d9d806 | fbrito++ | src/pmc/ (2 files):
Moving get_fd() from filehandle.pmc to handle.pmc

Signed-off-by: Jonathan "Duke" Leto <jonathan@leto.net>		 04:51
rrot: 70992f3 | fbrito++ | t/pmc/socket.t:
Test: Socket get_fd() should not return -1

Signed-off-by: Jonathan "Duke" Leto <jonathan@leto.net>
dukeleto fbrito: i would say have fun with GCI first. Think about it later.
dalek rrot: 93b2c43 | fbrito++ | t/pmc/filehandle.t:
Forgot to tell FileHandle test that I added a new one

Signed-off-by: Jonathan "Duke" Leto <jonathan@leto.net>
rrot: 7a4bef0 | fbrito++ | t/pmc/filehandle.t:
Test: FileHandle get_fd() should not return -1

Signed-off-by: Jonathan "Duke" Leto <jonathan@leto.net>
dukeleto figured out how to do Signed-Off-by lines
dukeleto just learned that "git apply --check foo.patch" checks if a patch will apply. Cool! 04:52
fbrito: what was that task link again? 04:53
fbrito www.google-melange.com/gci/task/sho...9104579996
I really need to learn how to use Git. I can do only the basic 04:54
And some times I still have to open a GUI interface (git-cola) to do somethings :P
dukeleto fbrito: you know a lot more than a lot of people :)
fbrito and a lot of people know a lot more than me 04:55
haha :D
dukeleto fbrito: those are the people to hang out with :)
fbrito totally true! 04:56
cotto dukeleto, can you add Signed-off-by documentation somewhere?
fbrito there is even this quotation "You are the average of the five people you spend the most time"
plobsing fbrito: can you run 'make codetest' before you commit? some of your code is not in line with our formatting policy (long lines mostly). 05:00
fbrito of course! sorry, didn't know of that 05:01
dukeleto fbrito: i didn't do a "make codetest" either 05:02
dukeleto slaps his own wrist
fbrito: i hadn't heard that quote before
fbrito: you can haz another finished task 05:03
fbrito++
karma fbrito
aloha fbrito has karma of 19.
dalek rrot: 1d7dded | plobsing++ | / (419 files):
update C file vim modelines

Specifies 2-space label outdent as per coding standard. Satisfies TT # 1547.
command used:
  > perl -Mlib=lib -MParrot::Distribution -E 'say $_->path for Parrot::Distribution->new->get_c_language_files' | xargs sed -i "s/vim:\\(.*\\):$/vim:\\1 cinoptions='\\\\\\\\:2=2' :/"		 05:04
rrot: 6dd9f88 | plobsing++ | / (141 files):
Merge branch 'master' of github.com:parrot/parrot
dukeleto cotto: i will add it to the docs. it has the side effect that github pull requests don't close themselves, but I think that is reasonable.
fbrito: do you have more tasks to hack on? 05:08
fbrito there is this task on warning experimental features: www.google-melange.com/gci/task/sho...9062052689 05:09
05:10 nwellnhof left
bacek_at_work plobsing, trac.parrot.org/parrot/ticket/1864 05:10
plobsing bacek_at_work: oops. 05:13
I like 2 space outdent anyways. 05:14
dukeleto fbrito: i just approved another IO task 05:15
fbrito: i am trying to. melange doesn't want me to
plobsing serves me right for opting to fix a random ticket I guess
dukeleto fbrito: it is there now.
05:17 hercynium left
dukeleto fbrito: just made another 05:18
fbrito ok, I will take a look :D
dukeleto rather, i approved some written by whiteknight
fbrito: one of them looks fun, the other is easy :) 05:19
fbrito the IO task looks interesting. I want to take it :) www.google-melange.com/gci/task/sho...9104725162 05:21
dukeleto fbrito: and just added another about adding readline to Socket PMC 05:22
fbrito: and it is yours :)
fbrito yes!! 05:25
but now I need to get some sleep
dukeleto fbrito: yes. talk to you later, and good luck!
fbrito thank you very much for approving the tasks, the pull requests and helping me
really. thank you :) 05:26
good bye 05:27
dalek TT #1547 closed by plobsing++: Mass-update vim modelines
TT #1547: trac.parrot.org/parrot/ticket/1547
05:27 fbrito left 06:09 TypeNameHere_____ joined 06:49 contingencyplan left
dalek rrot: 4dc8e16 | cotto++ | docs/project/core_inclusion.pod:
[project] initial version of core inclusion document
 06:57
rrot: f085399 | cotto++ | src/pmc/handle.pmc:
Merge branch 'fixing_doc' of github.com/fernandobrito/parrot into fernandobrito-fixing_doc
 07:09
07:26 luben joined 07:40 theory left 07:46 fperrad joined 08:04 uniejo joined 08:13 bacek joined
bacek aloha, humans 08:35
09:08 bacek left
dalek rrot: 85c73bf | fperrad++ | runtime/parrot/library/ (2 files):
[io] readall with encoding
 09:13
09:17 mikehh left 09:20 mikehh joined
dalek a: 7bf1fd1 | fperrad++ | dynext/pmc/luastring.pmc:
s/Parrot_get_ctx_HLL_namespace/Parrot_hll_get_ctx_HLL_namespace/
 09:47
09:52 bacek joined
dalek lscript: 5b96a08 | fperrad++ | dynext/pmc/wmlsbytecode.pmc:
s/key_append/Parrot_key_append/
 10:03
lscript: 1067c5d | fperrad++ | wmlscript/wmlsstdlibs.pir:
now encoding is mandatory
rrot: cda6fd7 | mikehh++ | MANIFEST:
re-generate MANIFEST
 10:13
rrot: e3c8a49 | mikehh++ | src/global_setup.c:
fix =item for c_function_docs
rrot: 9067976 | mikehh++ | docs/project/core_inclusion.pod:
fix codetest failure - linelength
rrot: 7e679e6 | mikehh++ | src/hash.c:
fix codetest failure - linelength
rrot: 735e267 | mikehh++ | src/hll.c:
fix codetest failure - linelength
rrot: 7979a2d | mikehh++ | src/namespace.c:
fix codetest failure - linelength
rrot: 997f960 | mikehh++ | src/ops/experimental.ops:
fix codetest failure - linelength
rrot/generational_gc: f39bc7b | bacek++ | / (6 files):
Copy WIP gc_ms2 into gc_gms to simplify merging with master
rrot/generational_gc: 1376dda | bacek++ | / (1911 files):
Merge branch 'master' into generational_gc

Conflicts:
  \tMANIFEST
  \tconfig/gen/makefiles/root.in
  \tsrc/gc/gc_ms2.c
  \tsrc/io/buffer.c
  \tsrc/pmc/class.pmc
  \tt/op/gc-non-recursive.t		 10:20
bacek Niiice! 1911 files! 10:21
git++ for so small merge conflicts!
10:28 fperrad left
dalek rrot/generational_gc: 0b47fa2 | bacek++ | src/gc/gc_ms2.c:
Fix build after merge.
 10:29
10:30 fperrad joined
mikehh I'm getting a failure in post-config tests -> t/steps/gen/platform-01.t - Failed test: 11 10:44
dalek rrot: 6e047aa | mikehh++ | config/gen/platform.pm:
replace CODA to reflect changes in coda
 11:05
rrot: 156a8cf | mikehh++ | config/gen/platform.pm:
forgot to change Copyright date
11:37 cognominal left
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1424) fulltest) at 156a8cf - Ubuntu 10.10 i386 (g++-4.5 with --optimize) 11:41
11:47 uniejo left 11:49 cognominal joined
dalek rrot: 26904d0 | jkeenan++ | docs/project/core_inclusion.pod:
review: github.com/parrot/parrot/commit/26904d02df		 13:19
13:35 whiteknight joined
dalek rrot/get_fd_method: 1429e1c | fbrito++ | src/pmc/ (2 files):
Moving get_fd() from filehandle.pmc to handle.pmc
 13:41
rrot/get_fd_method: 78fb393 | fbrito++ | t/pmc/socket.t:
Test: Socket get_fd() should not return -1
rrot/get_fd_method: eed7d38 | fbrito++ | t/pmc/filehandle.t:
Forgot to tell FileHandle test that I added a new one
rrot/get_fd_method: 8e50ded | fbrito++ | t/pmc/filehandle.t:
Test: FileHandle get_fd() should not return -1
13:41 Patterner left, plobsing left, Psyche^ joined, Psyche^ is now known as Patterner
whiteknight blah, I didn't realize dukeleto++ already pulled those 13:42
dalek rrot/fbrito_style_fixes: 67fa9c6 | fbrito++ | src/ (4 files):
Fixing lines longer than 100 columns

Those lines were written by myself and found later with "make codetest" command. Thanks to plobsing for teaching me this command on IRC.		 13:43
rrot/fbrito_style_fixes: 1e410d3 | fbrito++ | src/global_setup.c:
'make codetest' was giving errors on that
rrot/gci_io_close: e844c4c | fbrito++ | src/pmc/ (2 files):
Moving close() from FileHandle to Handle
 13:45
rrot/gci_io_close: 0428783 | fbrito++ | src/ (2 files):
Moving close() from Socket to Handle
rrot/gci_io_close: 4a7b901 | fbrito++ | src/ (2 files):
Moving close() from StringHandle to Handle
rrot/gci_io_close: b57705c | fbrito++ | src/io/api.c:
Changing some variable names from code that I've copied

Had to include pcm_socket because it couldn't find macro PARROT_SOCKET
rrot/gci_io_close: b13bbe8 | fbrito++ | src/io/api.c:
Reused documentation of old close() in the new one
rrot/gci_io_close: 4a3d5e7 | fbrito++ | src/io/api.c:
This would not work outside StringHandle
13:45 contingencyplan joined
Coke when adding #includes, be sure to also fixup the dependencies. 13:48
(you can trust most of the warnings on .c files from tools/dev/checkdepend.pl) 13:49
whiteknight heh. We may need to make fbrito a committer just to avoid having to merge in all his patches 14:01
do we have an age limit on CLA submissions? 14:03
msg nwellnhof I'm seeing errors in t/op/string_cs.t this morning. Looking through the recent commit list your branch merge looks the most related. Any ideas? 14:31
aloha OK. I'll deliver the message.
dalek rrot: 9984719 | Whiteknight++ | src/ (5 files):
Merge branch 'gci_io_close'
 14:32
14:35 fbrito joined 14:38 jsut joined
Coke whiteknight: Not as such, but since we're basically asking them to sign a contract, maybe. 14:41
Did someone recently remove include/parrot/list.h ? 14:42
(or, rather, remove a need to use it everywhere?)
14:43 jsut_ left 14:44 preflex left 14:48 preflex joined
whiteknight Coke: I have no idea about list.h. 14:50
Coke: Yeah, there probably is an age limit on binding contracts. I don't think we'll be able to have any GCI students (< 18 yrs old) sign them
Coke I need to get checkdepend running with codetest so we stop regressing.
whiteknight or maybe they would need a parent to sign them
Coke: Yes, that would be a great idea 14:51
Coke whiteknight: to be sure, You'd have to check with the lawyers. Who would then probably want to charge you money.
whiteknight Coke: to be safe, we simply disallow it for youngsters
Coke git question: I have 2 un-pushed commits on my local master. how I do push only 1 of them? 14:59
PerlJam Coke: I would guess that it's: git push origin <SHA1-of-commit-you-want-to-push> 15:06
Coke: but I've never done this before
15:11 fbrito left 15:12 fbrito joined
Coke webgui++ 15:15
15:15 nwellnhof joined
nwellnhof Coke, do you still need help with your git question? 15:21
15:22 allison joined
dalek rrot: e483ba1 | NotFound++ | t/pmc/complex.t:
unskip Complex div by zero tests
 15:23
Coke nwellnhof: I will try perljam's suggestion later. for now, $DAYJOB 15:26
NotFound This causes a failure in JSON tests: "Longitude" => -122.026020, -> "Longitude" => -122.02602, 15:37
Note the trailing zero. 15:38
not ok 59 - another example taken from the RFC # TODO check number precision
But the RFC says nothing about that, is just an example of valid JSON, it doesn't says it should stringify that way.
tools.ietf.org/html/rfc4627 15:39
dalek rrot: f35ee58 | NotFound++ | t/compilers/data_json/to_parrot.t:
fix and unTODO a json test after chekcing that the reason failing was not valid
 15:47
NotFound Fixed
15:48 dmalcolm joined
PerlJam Coke: PerlJam's suggestion doesn't work (I just tried it) 16:06
16:08 hercynium joined
Coke (my question would be mooted by someone trying the "allow a space in install dir" patch on windows. ;) 16:16
nwellnhof Coke, do you want to push the first or second commit? 16:20
whiteknight: ping 16:27
fbrito hello... I am doing a GCI task to add some experimental warnings (similar to deprecated warnings) and I need to know if there is any convention of putting those warnings on the end of the method, on the begging, etc 16:33
At the moment I am adding it in the first line of the method with a blank line after (github.com/fernandobrito/parrot/co...85dba5f8). Please, if you have any comments about that let me know 16:37
whiteknight nwellnhof: pong
fbrito: it should be after all variable definitions 16:38
nwellnhof whiteknight: what errors are you seeing in string_cs.t?
fbrito whiteknight: ah, ok... thank you :D
whiteknight nwellnhof: t/op/string_cs.t (Wstat: 256 Tests: 55 Failed: 1) 16:39
Failed test: 26
Non-zero exit status: 1
trans_encoding_s_s_i utf8 to ucs4
nwellnhof whiteknight: do you have the output of that test? 16:41
whiteknight # got: '0
# ucs4
# 3
# '
# expected: '1
# ucs4
# 3
# ' 16:42
nwellnhof Ah, I see what's probably wrong. What platform/compiler do you use? 16:45
Coke nwellnhof: the second one.
... to answer an earlier question.
nwellnhof Coke: OK, that's more difficult. There's more than one way to do it. git is a lot like perl in that regard. 16:46
What I would do: 16:47
1) git branch topic_branch
2) git reset --hard HEAD~2
3) git cherry-pick <hash of commit to push>
4) git push origin master
That moves the two commits in a new branch, resets master, picks one commit and pushes it. 16:48
Coke nwellnhof: I could always just use git rebase -i to reorder the commits, if that would help. 16:49
nwellnhof Coke: That's another possibility. If you want to push the first of two commits, the following should work: 16:50
git push origin <hash>:master
But I never tried that. 16:51
Coke where hash is (now) the first commit? 16:52
dalek rrot/mappedbytearray: 5d42d2d | NotFound++ | src/pmc/mappedbytearray.pmc:
add method get_utf8 to MappedByteArray
nwellnhof Yes, but you should try a dry run first.
nopaste "coke" at 192.168.1.3 pasted "non cage warnings" (6 lines) at nopaste.snit.ch/26296 16:54
Coke urk. bad sha1. moment...
dalek rrot: e483077 | Coke++ | config/gen/makefiles/root.in:
fix some makefile dependencies.
 16:55
Coke nwellnhof++
nwellnhof yes, looks good. 16:56
Coke PerlJam++
17:09 theory joined
fbrito whiteknight: ping 17:26
Coke wonders if "Reverend Chip" == our chip. 17:29
cotto_work ~~ 17:33
seen kid51 17:35
aloha kid51 was last seen in #parrot 13 hours 7 mins ago saying "fbrito: Should you ultimately wish to apply for a commit bit, please fill out a CLA: www.parrot.org/files/parrot_cla.pdf".
dalek rrot: 35b5125 | cotto++ | docs/project/core_inclusion.pod:
add a couple more questions from kid51++ to core_inclusion.pod
 17:39
tracwiki: v48 | coke++ | GitMigration 17:40
tracwiki: already did a release, need for dry run gone.
tracwiki: trac.parrot.org/parrot/wiki/GitMigr...ction=diff
tracwiki: v49 | coke++ | GitMigration
tracwiki: trac.parrot.org/parrot/wiki/GitMigr...ction=diff
cotto_work msg kid51 If you have any more questions to add to core_inclusion.pod, feel free to add them directly. We can always clean them up later if you don't have the tuits atm. 17:49
aloha OK. I'll deliver the message.
17:54 nwellnhof left
Coke trac.parrot.org/parrot/wiki/GitMigration - now that this is nearly done, would it make sense to move the remaining tasks into tickets? 17:59
(or at least have duke & cotto clean this page up?)
18:04 Kovensky left 18:05 cognominal left, cognominal joined
whiteknight Coke: I'm always in favor of making dukeleto and cotto do more work :) 18:07
fbrito whiteknight: hey :D 18:08
Coke I was never a big fan of having tasks in a wiki page. 18:11
whiteknight hello fbrito
Coke: Yeah, I don't think it's the ideal medium. But we don't really have another good way on trac to link together related tickets and put them in an order 18:12
fbrito About my task... 18:13
I did a "cat DEPRECATED.pod | grep warning" and started to put some warning messages here and there 18:14
whiteknight codicesoftware.blogspot.com/2010/11...ching.html
fbrito I mean "grep experimental"
whiteknight right
that list changes frequently, so try to cover what is there and don't worry abut things that change 18:15
I will be making changes later today, and I will use your new function on it
fbrito pastie.org/1335961
whiteknight dukeleto: codicesoftware.blogspot.com/2010/11...ching.html 18:17
fbrito: that looks like a good list. I would call that complete if everything works 18:18
fbrito Ok, but I still have run all tests and test the warning message with "parrot -w" 18:19
When I am finished I will let you know. Thanks :)
18:25 allison left 18:28 Kovensky joined
dukeleto is at a Galois Tech Talk about the Rubinus Virtual Machine 18:31
and I just heard a small jab at Parrot, methinks. 18:32
whiteknight blah 18:33
Coke hell, /i/ jab at parrot.
dukeleto Coke: i am glad that he *mentioned* parrot ;) 18:34
Coke "then they mock you..." 18:35
dukeleto++
whiteknight I don't mind a few jabs at Parrot. We aren't exactly cutting a record-breaking development pace 18:46
=I estimate Parrot is about 5 years behind where it should be, in terms of development and capability 18:47
but we have been picking up a lot of steam in the past year or so, so I'm encouraged by that
atrodo whiteknight> Development pace becomes difficult when you have "customers" 18:48
whiteknight atrodo: No question. 18:49
atrodo Pesky customers if you ask me. Depending on what I've made available for free 18:50
whiteknight But in the past few months, since 2.6 came out, we've been making very good development pace and still have several customers
We spent about a year and a half not making any substantial interface changes, only cleaning up and refactoring internals 18:51
with that behind us, we can start adding lots of new features
18:51 dngor left
atrodo whiteknight> Actually, I think having Rakudo in the pre-release stage has helped a lot too 18:52
dukeleto "Rubinius has had roughly 100-150 contributors in the last 4 years" 18:53
Coke whiteknight: the past few months have also pretty much lost you another customer. YMMV.
dukeleto Coke: which customer? 18:54
atrodo I would guess he's referring to partcl 18:55
Coke ayup 18:56
whiteknight Coke: also, no question.
dukeleto Coke: why did Parrot lose Partcl? Do you think continuous integration would have prevented that from happening? 18:57
I do.
Coke it would have helped a lot, yes. 18:58
18:59 dngor joined
dukeleto We need continuous integration, badly. 18:59
moritz call me stupid, but what exactly is "continuous integration"?
Coke I already spent a huge amount of time in the past year fixing a bunch of stuff parrot changed/bitrotted/deprecated after months of not touching it, and got things working. It then happened again and I just haven't had the energy to fix it. (esp. given the one or 2 things that were reported were not fixed) 19:00
moritz: like what taptinder does.
but for HLLs.
cotto_work moritz: automatic building and testing at regular intervals, ideally after each push
whiteknight I had said it pretty frequently for some time: Parrot really has historically had too many problems both internally and externally to be reliable as a base for many projects. 19:03
And I sincerely believe that the deprecation policy aggravated that
and the lack of continuous integration too
19:03 simcop2387 left
sorear Oh, the dep policy is a new evil? 19:03
whiteknight sorear: the deprecation policy has always been evil 19:04
cotto_work whiteknight: do you see it as being something that needed to get worse before it could get better or is it just something we shouldn't have gotten into?
19:04 simcop2387 joined
whiteknight cotto_work: It's been getting much better 19:04
Coke I disagree. As a user, I find the deprecation policy mostly helpful.
because without it, I'd have drowned ages ago.
whiteknight Coke: the deprecation policy itself isn't the problem. It's the underlying system that's the problem. The deprecation policy puts weird restrictions on the system, and says the broken shit needs to remain broken shit 19:05
You can't take an ad hoc, organic, ugly pile of features, call them an "API" and then make a blanket guarantee that you support every piece of it
Coke I find this conversation no more helpful than it was when we had it the last time. 19:06
dukeleto I meant "continuous integration of HLLs on core"
whiteknight Coke: you take an unstable system, add partcl with "months of not touching it", and then a deprecation policy that says breakages happen in large waves, and you have a recipe for disaster 19:08
Parrot is becoming significantly more robust and stable, but the damage for many people is already done
Coke when the breakages are documented, it's not a problem. 19:09
I can then go through the list and find them.
and fix them.
the problem, IME, is more when someone changes something without realizing (caring?) that it's actually changing functionality and therefore it is never documented like "official" deprecations. 19:10
whiteknight That's exactly what I am talking about. We don't have a list anywhere of what, exactly, is part of our API and is covered
Coke for example, see TT #....
whiteknight and we certainly don't have 100% test coverage of the API list we don't have
and when something breaks, we don't know, because we didn't know we weren't supposed to break it
Coke trac.parrot.org/parrot/ticket/1811
dukeleto "Rubinius currently only supports UTF8"
Coke So, to sum up, IME, the deprecation policy is not the problem. 19:11
whiteknight I doubt highly that the deprecation policy will change at this point. so it's a huge waste of my time to argue about it's merits
dukeleto I think all these problems are natural growing pains for something like Parrot.
Coke Hopefully cotto will fix everything now that he's the architect. 19:12
dukeleto +1 # no more beating up on the dep policy. It accomplishes nothing.
Coke (which, btw, got FAR less coverage than I would have expected)
dukeleto Coke: ?
Coke dukeleto: ??
whiteknight that's a tall order for anybody. cotto isn't going to fix the world.
Parrot is fixing itself naturally
but that brings me back to my thesis: Parrot as a project is 5 years behind where it really should be, in terms of development 19:13
Coke I think you're being overly generous. ;)
dukeleto Coke: why do you think cotto will "fix everything" ?
whiteknight that slowdown means we deserve jabs from people like the Rubinius folk
Coke dukeleto: <sigh>
I don't. 19:14
dukeleto Coke: heroic efforts are unsustainable. We need to learn how to communicate and hack together more effectively, as a group.
whiteknight dukeleto: he was being sarcastic
dukeleto whiteknight: indeed we do.
Coke dukeleto: sorry. will refrain from humorous statements in channel in future.
cotto_work dukeleto: that's what I'm going to try to encourage. 19:15
dukeleto Coke: why? You seem to be so good at them...
whiteknight The "architectural team" approach I think will be far more important than any one person taking the lead role
certainly more sustainable 19:16
19:16 Coke left
cotto_work dukeleto: where are you? 19:16
dukeleto cotto_work: in downtown Portland, at Galois 19:17
dukeleto is learning about the Rubinius GC
cotto_work nice. Make sure to blog about it. 19:18
dukeleto blarg 19:20
cotto_work Well, that's a good start.
whiteknight make sure to blarg about it
cotto_work you can take notes during the presentation and post them unedited if tuits are an issue 19:21
19:22 nwellnhof joined
dukeleto cotto_work: i always take notes, but blag toasts don't happpen so often 19:22
Rubinius uses LLVM to turn their bytecode to LLVM IR 19:24
cotto_work sounds familiar 19:25
19:26 dngor_ joined
whiteknight We'll be there soon enough 19:26
cotto_work I've passed the point where I worry about where we aren't. 19:27
It's not a fruitful exercise. 19:28
19:29 dngor left
cotto_work speaking of fruitful exercises, #ps in 63 19:30
whiteknight cotto_work: Well, I do find it instructional to take an honest look at the project and take stock of it 19:31
dukeleto I think we can learn alot about how to use LLVM by seeing how Rubinius does it. I've learned a bunch just at this talk
whiteknight Parrot has a lot of potential, and I mean that in like the physical sense of Potential Energy: I sincerely believe we're coming to a critical mass point where we either go big or go home 19:32
There are a hell of a lot of good ideas in the codebase, a hell of a lot of features, and several great devs. If we can tie things together and build on what we have, we win 19:33
dukeleto whiteknight: Let's Do It. 19:34
whiteknight Look at it this way: We get a pluggable, modern optimizations system. We rework our bytecode to be more performance-conscious, then we implement a good JIT 19:35
suddenly we're back towards the top of the pack in terms of runtime performance
Fix the object model to cut out inefficiencies there, and make it flexible to get the Python, PHP, and JavaScript people the basis they need 19:36
Add good thread support, and suddenly we're more feature-rich than most modern dynamic language interpreters 19:37
dukeleto I think Rubinius is our biggest competitor and we should compare ourselves to them to see where we fall short. 19:38
19:38 dngor_ is now known as dngor
dukeleto whiteknight: i agree with everything you just said. 19:38
whiteknight: but what is the optimal order to accomplish those things? 19:39
whiteknight right, and those are just the major points on the roadmap, that takes us to the 4.0 release or just after
I think the order I mention there is good. Lorito -> JIT -> Threads is really a decent flow 19:40
the object metamodel can happen at any time, but the earlier the better
dalek rrot: cf42ee5 | nwellnhof++ | src/string/encoding/ucs4.c:
[str] Hopefully fix UCS-4 test failure discovered by whiteknight++
cotto_work dukeleto: privmsg ping
whiteknight nwellnhoff++ # Let me pull and see what happens
TimToady "rework our bytecode to be more performance-conscious" --> "install an adequate native type system" 19:45
atrodo TimToady> Parrot's native type system is inadequate? 19:48
19:51 Tommy joined
cotto_work TimToady: what would an adequate type system look like? 19:54
whiteknight TimToady: preaching to the converted. But our native type system is pretty deeply ingrained. It's not something we can really change at this point 19:57
if we pull out the rug like that, we will lose Rakudo as users of Parrot, no questions asked
cotto_work: If you read any good papers, send a link my way. I love reading them 19:58
cotto_work whiteknight: I've got the yacc paper and "prototypes with multiple dispatch" on my current list 19:59
19:59 bacek left 20:00 Tommy left
whiteknight I read the yacc paper too. Dense stuff. I had to thumb through my copy of the dragon book, and I still don't think I took it all in 20:01
cotto_work Nice. I'm excited about it now.
whiteknight what I do know is that we need to find somebody smarter than myself to implement a derivative-based parser generator
much, much smarter than myself
cotto_work sounds like a job for pmichaud
whiteknight when he started talking about a "lazy stream of parse trees", I thought to myself "I don't know what any of those words mean, but I MUST HAVE IT" 20:02
cotto_work goes off to hunt a wild sandwich
whiteknight Util: ping
Util whiteknight: pong 20:03
whiteknight Util: I had to make a lot of changes to pbc_to_exe in the embed_api branch. I'm definitely going to want you to take a look at it eventually, for sanity
Util whiteknight: Is today too early for me to have a initial look? 20:04
whiteknight Util: no, today is mostly fine. I need to make some changes to the error-handling portion of the generated C code, but those are relatively minor
Util whiteknight: I will look at it now. 20:05
whiteknight Util++
NotFound I'd better kill pbc_to_exe 20:07
whiteknight ...!!
NotFound We don't need a fakacutable for that work, 20:08
dukeleto aloha msg bacek_at_work have you ever heard of the Immix GC (Blackburn + McKinley, PLDI 2008) ? It is what Rubinius uses. 20:09
aloha dukeleto: OK. I'll deliver the message.
NotFound Ah, you mean changing the generated code... 20:10
whiteknight yes, the generated code changes 20:11
Util NotFound: could you clarify "for that work" for me? 20:22
whiteknight plobsing++
let us all increase the awesome
msg nwellnhof thanks, your latest commit fixed my bug 20:28
aloha OK. I'll deliver the message.
NotFound Util: we can just use the pbc. Is not a tool intended for casual users. 20:30
ps time? 20:31
whiteknight indeed
20:34 tcurtis joined 20:40 bacek joined
dalek tracwiki: v71 | cotto++ | ParrotQuotes 20:55
tracwiki: trac.parrot.org/parrot/wiki/ParrotQ...ction=diff
Util whiteknight: Your pbc_to_exe looks sane so far. 21:27
All the calls joined with && should each get their own error block, but you probably know that.
whiteknight Util: yeah, prototype 21:28
whiteknight leaves 21:29
21:29 whiteknight left 21:30 dngor left
dukeleto This is a really interesting read about Oracle + Hudson: www.hudson-labs.org/content/whos-driving-thing 21:32
Util re: C99 - As of last month, M$ says "we do not have plans to add C99 support to the compiler"
connect.microsoft.com/VisualStudio/...99-support
Also, connect.microsoft.com/VisualStudio/...9-standard 21:33
atrodo Wow, how very, typical, of them
Util Thanks for submitting this suggestion. I've resolved it as Won't Fix, because we currently have no plans to implement C99 Core Language features. While we recognize that a few programmers are interested in those features, --snip-- 21:34
NotFound "a few" 21:36
atrodo "a few", as in everyone who doesn't like C++ 21:38
NotFound I think they also don't support recent C++ 21:39
21:41 plobsing joined 21:44 perlite left 21:45 perlite joined
bacek_at_work aloha, humans 21:45
dukeleto bacek_at_work: werd 21:46
bacek_at_work dukeleto, whut?
Util multi sub aloha () {...}; # Hello or goodbye?
dukeleto bacek_at_work: i sent you an aloha message 21:47
bacek_at_work dukeleto, yes, reading now.
dukeleto Someone is starting to hack on Lua on Rubinius: github.com/Twisol/lupin
NotFound <NotFound> There is no info in ParrotDeprecations for the functions recently renamed. 21:48
cotto_work NotFound: how many of the are api functions? 21:49
those are the only ones we care about that
nwellnhof And the renaming broke the Rakudo build.
cotto_work hears a chainsaw in the disance
NotFound cotto_work: I think the purpose of that page is to help, not to dicuse minutiae-
We need to document that things, they are used, being api blessed or not. 21:50
cotto_work NotFound: the API functions are the only ones we claim to support and what we recommend that external users use. 21:51
dukeleto bacek_at_work: i am going to try to get my hands on that Immix GC paper. Let me know if you beat me to it.
NotFound cotto_work: then we don't support anything.
There are no PARROT_API decorated functions.
bacek_at_work dukeleto, hackage.haskell.org/trac/ghc/wiki/C...e/GC/Immix this one?
cotto_work fbrito: ping 21:52
NotFound I though the purpose of the Deprecation wiki pages was to help. For that goal we need to put changes in things that were actively used in the real world.
dukeleto bacek_at_work: yes, i think they are talking about the same thing. I want to read the original paper.
bacek_at_work dukeleto, www.cs.utexas.edu/users/speedway/Da...i-2008.pdf this one? 21:53
dukeleto bacek_at_work++
bacek_at_work: i would love to hear your thoughts about what our current GC can learn from Immix 21:54
cotto_work NotFound: some of fbrito's changes clearly went in prematurely. We need to revert some and go through the proper deprecation cycle.
dukeleto cotto_work: which changes?
cotto_work whichever broke the Rakudo build 21:55
NotFound cotto_work: I don't think we should revert. Better provide a compatibility layer during the deprecation period.
cotto_work That's also an option.
bacek_at_work which changes?
cotto_work a better one, even
bacek_at_work: some function renaming 21:56
NotFound That's the problem. We need a list of that changes.
bacek_at_work cotto_work, it's covered by #443
cotto_work in that case we need notifications 21:57
bacek_at_work best option - provide patches for rakudo.
NotFound That's what I hate in our policy. We just add a note, and some day we break things without providing a way to be ready. 21:58
22:00 lucian joined, Kristaba joined
NotFound Anyway, fixing the code is good, but we should put a list of that renaming in the Deprecation pages. 22:00
Kristaba Hi everyone.
dukeleto Kristaba: howdy! 22:01
dalek rrot: 4b895ad | nwellnhof++ | / (2 files):
[str] Rewrite Unicode upcase, downcase and titlecase

Set correct result strlen. Fix downcase and titlecase to combined chars.		 22:02
dukeleto NotFound: perhaps we need a program that reads PIR/PASM/C and then lists out all the deprecated things that are used, with links to the deprecation pages about them 22:04
NotFound: we need a better way for users to know what deprecations are
NotFound dukeleto: users? *I* need it now.
dukeleto Perhaps Parrot needs something like Package::DeprecationManager search.cpan.org/~drolsky/Package-De...Manager.pm 22:05
that evolved out of Moose deprecations biting people
NotFound: i hear ya. I am in the same boat.
NotFound: i am sure PL/Parrot is getting the rug ripped out from under it since i haven't hacked on it in a few months
22:05 dmalcolm left
NotFound What happened to STRING_validate? 22:06
22:08 bluescreen joined
fbrito cotto_work: pong 22:09
nwellnhof NotFound: use STRING_scan instead 22:10
cotto_work fbrito: some of the functions you renamed broke the Rakudo build. Could you add deprecation notices for any supported functions that were renamed?
dalek rrot: f10c1d5 | nwellnhof++ | t/pmc/stringiterator.t:
[t] More tests for string iterators and encodings
 22:11
cotto_work Our support policy is docs/project/support_policy.pod and deprecations are listed on ParrotDeprecations on the wiki
22:11 dngor joined
NotFound nwellnhof: What is the current status on that? Do we need to validate after creating a string, or is automatically done? 22:12
cotto_work It'd also be nice to submit a patch for Rakudo. It shouldn't be any harder than it was to change the function names in the first palce.
fbrito cotto_work: oh :O. ok, but I am leaving right now and will be back only in a couple of hours
nwellnhof NotFound: it's done automatically if you use str_new_init
22:12 Hunger left
cotto_work no worries 22:12
NotFound nwellnhof: ah, good.
Kristaba Sorry to ask a question without become acquainted with you better before, but I work on PLA and I have a (probably stupid) problem when I try to run the test.
I have an error message : "Contextual $*FileSystem not found" 22:13
dukeleto fbrito: this will be another task for you 22:14
Kristaba I'm not an usual user of Perl and Parrot, so it's maybe a blunder from me :/
dukeleto cotto_work: you need to make a task for fbrito to work on that
cotto_work: or you can tell me the gist of it and I can create it 22:15
Kristaba: i have run into that error as well
NotFound Kristaba: I had that problem one time, but don't remember what was the cause.
dukeleto Kristaba: it means the wrong version of NQP or Kakapo is being used, I think
NotFound Maybe Kakapo related, yes. 22:16
dukeleto Kristaba: whiteknight is the best person to ask about that, but he isn't here right now. Have you followed the instructions for PLA? It requires a special version of Kakapo
Kristaba: if you send an email to parrot-dev with all your pertinent info (OS version, parrot version, PLA version, etc) and say that you are a GCI student working on a task, I am sure someone will help you 22:17
Kristaba dukeleto: it's strange, because I built the two (Parrot and Kakapo) today :o
Oh.
dukeleto Kristaba: also, sometimes the wrong version of Parrot is being used. Did you install parrot recently?
Kristaba: did you install the version that you compiled for PLA?
Kristaba dukeleto: Yes I did 22:18
dukeleto Kristaba: i can give you meta-info, but I've never been able to get the PLA test suite to work for me. I stopped at the problem you are having now :)
Kristaba: which means you are asking very good questions
Kristaba: welcome to the Parrot Community. Don't be afraid to ask any questions, we won't bite :)
Kristaba dukeleto: thank you, I'll try some others things before try to contact Whiteknight ;) 22:20
dalek Some 'NotFound' person just gave me a 7102 byte commit packet. They need to be more careful in the future. 22:23
cotto_work dukeleto: the gist is to either create a wrapper around the supported functions whose names changed or submit a fix to any HLLs/libraries that are using them.
dukeleto Kristaba: i think the version of Kakapo and/or nqp-rx that PLA is using is not right. Telling whiteknight all the version numbers involved will probably tell him what he needs to know
cotto_work: that is pretty vague and open-ended 22:24
dalek: you talk now?
NotFound Nice message.
Anyway, careful or not, mappedbytearray is merged into master. 22:25
dukeleto cotto_work: fixing each HLL would be a seperate task 22:26
cotto_work: i think you should talk to whiteknight about that stuf
mikehh NotFound: some codetest failures, working on it 22:29
Kristaba dukeleto: In fact, I was wrong, my Kakapo version is the last build, whereas LPA need the "TBA" version of Kakapo. But what does TBA means? 22:30
dalek rrot: 5f23cc5 | nwellnhof++ | t/ (3 files):
[t] More string tests to improve test coverage
 22:31
cotto_work dukeleto: I'll make sure it doesn't get dropped. 22:35
msg cotto Make sure Rakudo et al get fixed after fbrito's changes.
aloha OK. I'll deliver the message.
22:37 plobsing left
NotFound mikehh: my fault, forgot to check some of the last changes. 22:39
dukeleto Kristaba: there should be a kakapo branch or version tag that PLA needs 22:40
Kristaba: which directions are you reading?
jnthn cotto_work: To what degree does core_inclusion.pod include things like nqp-rx? 22:50
(That is, do the criteria apply to that, or are they more for things that get compiled into Parrot itself?) 22:51
Kristaba dukeleto: It's that I thought but I see no tag or branch that contain "tba"
cotto_work jnthn: it's relevant to potential future components, not software that's already included. 22:52
jnthn cotto_work: nqp-rx could count as both of those since it's going to end up with dynops/dynpmcs landing at some point.
dalek rrot: a61f2c4 | NotFound++ | DEPRECATED.pod:
set experimental status of MappedByteArray, TT #1877
 22:53
jnthn Not sure if that counts as "future componetns" or just an increment of existing software though :)
dalek rrot: 944a850 | mikehh++ | src/pmc/mappedbytearray.pmc:
fix coda to conform to new coda
 22:54
rrot: 6ef4fda | mikehh++ | src/pmc/mappedbytearray.pmc:
fix indentation of pre-processor directives to pass codetest
rrot: 1ccf9fc | mikehh++ | docs/project/core_inclusion.pod:
fix codetest failure - linelength
jnthn (Of course, the new bits are actually underlying some much deeper changes.) 22:55
cotto_work jnthn: are these dynops and dynpmcs something I'd know about if I were caught up on 6guts? 22:56
jnthn cotto_work: I've not talked in depth about the Parrot mapping of 6model a great deal.
cotto_work: But they're just part of implementing that. 22:57
cotto_work: Mostly, I'm finding it easier to do the first cut in a smaller/managed environment then port it to Parrot.
cotto_work .net is a bit more stable right now 22:58
jnthn Aye :)
It's a lot more -Ofun to target.
Part of that is that I can do what I want in the runtime layer though.
Without having any deprecation issues to care about.
NotFound origin/codestring-deprecation-prep This branch is dead, isn't it? 22:59
fbrito cotto_work: how sould I solve the problem you mentioned? should I submit a patch for Rakudo updating the functions that I renamed or should I create wrappers with depreciation warnings?
jnthn cotto_work: Anyway, I guess that at some level, when nqp-rx with 6model lands, and if it gets bundled with Parrot, then at some level it has to pay some attention to deprecation policy. 23:00
cotto_work fbrito: I'd add wrapper functions first.
fbrito ok
jnthn s:1st/at some level//
:) 23:01
NotFound fbrito: our policy is not very good. We have items deprecated for long time without providing a replacement. That means that when is finally deleted or renamed no one can be prepared in advance.
nwellnhof fbrito: here is the official deprecation howto: trac.parrot.org/parrot/wiki/HowToDeprecate
jnthn Anyway, I need tuits and to stop feeling ill too much of the time so I can actually get more progress in on 6model stuff...
cotto_work NotFound: agreed. It's not good to have some functions that might disappear at any time continue existing for more than a couple deprecation cycles. 23:02
NotFound cotto_work: that is just ugly, but not having a working replacement before killing time is not just ugly, is disturbing. 23:04
dalek TT #1877 created by NotFound++: MappedByteArray experimental PMC 23:05
TT #1877: trac.parrot.org/parrot/ticket/1877
cotto_work I wonder how feasible it'd be to get rid of all our deprecations within a set period of time.
at least, all those older than x 23:06
NotFound cotto_work: no way. For example we have a deprecation notice for exception handlers.
And no one has even imagined a replacement.
Personally, I don't want to provide one, if that means opening the door for killing all existing code in one release. 23:07
dukeleto Kristaba: tba means "To Be Announced" 23:09
NotFound I'll suggest to delete all eligibility notes for items that don't have yet a suggested replacement.
dukeleto Kristaba: those docs are old. is that in the README ?
23:09 hercynium left
Kristaba dukeleto: Oh, yes of course, my english is so bad! 23:14
dukeleto: The README says nothing about the Kakapo version, only the RELEASES file refers to this 23:16
dalek TT #557 closed by nwellnhof++: ucs2_decode_and_advance conditionally defined on ICU 23:21
TT #557: trac.parrot.org/parrot/ticket/557
23:22 tcurtis left
dalek rrot: 8efa315 | dukeleto++ | docs/project/git_workflow.pod:
[docs] Add docs for merging Github pull requests
 23:23
dukeleto cotto_work: ^^^
23:29 nwellnhof left
cotto_work dukeleto: thanks 23:30
dalek TT #934 closed by nwellnhof++: PGE build leads to assertion failure on MacOS X? 23:37
TT #934: trac.parrot.org/parrot/ticket/934
23:42 nwellnhof joined 23:46 tcurtis joined 23:49 whiteknight joined 23:51 Hunger joined
dalek TT #904 closed by nwellnhof++: PGE trigger a "Segmentation fault" on ppc 23:54
TT #904: trac.parrot.org/parrot/ticket/904
23:54 plobsing joined 23:58 TiMBuS_ is now known as TiMBuS 23:59 fperrad left

Powered by Cro and the Raku® Programming Language.

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

Thank you!