Parrot 0.8.2 Feliz Loro Released | parrot.org | 45 TT | 538 RT
Set by moderator on 17 December 2008.
00:03 kid51 joined 00:04 Zaba joined 00:09 AndyA joined
pmichaud november? 00:35
purl november is at www.november-wiki.org/ or use.perl.org/~masak/journal/37212
masak hm. 00:37
hold on.
purl: november is also github.com/viklund/november/
purl okay, masak.
Coke sees rakudoperl.* has been squatted. 00:46
Coke breaks his partcl blog posting record. 00:47
00:54 Zaba_ joined
pmichaud Domain Name: RAKUDOPERL.COM 00:54
Created on: 15-Jan-08
Expires on: 15-Jan-10
Last Updated on: 15-Jan-08
Administrative Contact:
Michaud, Patrick pmichaud@pobox.com
darn those squatters! 00:55
00:55 MariachiElf joined
Coke I was basing that entirely on the content at those sites. 01:00
pmichaud I haven't bothered to make them point anywhere yet. :-|
I registered them to make sure someone else didn't get them. 01:01
01:01 Zaba joined
masak that's the definition of squatting, isn't it? :) 01:04
pmichaud There is that. :-)
I made sure I could get the domains before I chose the name "Rakudo" :-) 01:05
masak :) 01:06
pmichaud aiiiieee! 01:08
(some code just makes me cringe.) 01:09
chromatic Do I have some code for you then! 01:10
01:12 Zaba_ joined 01:16 Zaba joined
pmichaud looks like I have november compiling again -- just waiting for a spectest run to complete and then I can commit. 01:20
Whiteknight well, parrot doesn't even configure with icc, so that answers that question
01:20 Zaba_ joined
masak pmichaud++ 01:20
chromatic I know that ICC had some trouble with CONST_STRING on linebreaks. 01:21
pmichaud I don't know if it _runs_, but it compiles :-)
Coke compiles?
purl SHIP IT
chromatic ... but we could make the CONST_STRING processor check for linebreaks.
masak pmichaud: if it doesn't run, you'll hear from us soon. :)
pmichaud the test_wiki.sh script appears to work.
masak pmichaud: that's how we measure success right now. 01:22
pmichaud congratulations on some good packaging -- the fact that I was able to download and get this working so quickly means someone did some good work. masak++
masak blushes
we are basically a group of young people stumbling around in the dark, doing what we hope is right. 01:23
01:25 bacek_ joined 01:29 TiMBuS joined
dalek r34091 | pmichaud++ | : 01:31
: [rakudo]: Fix handling of $?PACKAGE, @?PACKAGE, $?CLASS, @?CLASS, etc.
: These were pushing null values onto the stacks, which meant they weren't
: being restored properly on exit (as discovered by November++).
diff: www.parrotvm.org/svn/parrot/revision?rev=34091
Coke found a library file that doesn't even compile. :| 01:35
dalek r34092 | Whiteknight++ | : 01:36
: [gsoc_pdd09] updating to trunk r34088 so we can get a clear diff
diff: www.parrotvm.org/svn/parrot/revision?rev=34092
01:39 Zaba joined
Coke allison? 01:41
purl allison is Allison Randal <mailto:allison@perl.org>
Coke anyone know what replaced the recv opcode?
chromatic Should be a method on an IO PMC. 01:42
Coke hurm. Now that we have Socket, do we need runtime/parrot/library/tcpstream.pir ? 01:43
(which doesn't compile)
dalek r34093 | jkeenan++ | : 01:49
: Correct punctuation error in comment.
diff: www.parrotvm.org/svn/parrot/revision?rev=34093
01:53 apeiron joined 01:55 Andy joined
dalek r34094 | chromatic++ | : 01:58
: [src] Fixed all CONST_STRING macros so that they don't break across lines.
: This should make it easier for picky compilers to get the #line numbers right.
diff: www.parrotvm.org/svn/parrot/revision?rev=34094
chromatic From the "Break early, break often" dept. 02:05
dalek r34095 | chromatic++ | :
: [tools] Added an exception to c2str.pl when a CONST_STRING macro occurs after a
: linebreak. The regex to detect this may need tweaking in the future, but this 02:06
: should prevent weird error messages on compilers with very picky #line
: handling.
diff: www.parrotvm.org/svn/parrot/revision?rev=34095
GeJ wonders why his 'compilers/imcc/imcparser.h' always get modified 02:09
chromatic Various maintainers have different versions of yacc/bison installed. 02:10
GeJ Ah, indeed, I define --lex and --yacc in my configure script. 02:12
thanks.
dalek r34096 | jkeenan++ | : 02:14
: In r30478 ( 2008-08-23 ), I changed the name of t/tools/ops2pmutils/ to
: t/tools/ops2pm/. I did not change all instances of the old name in code and,
: as a result, the tests in t/tools/ops2pm/*.t have not been run as part of 02:15
: 'make buildtools_tests' or 'perl Configure.pl --test' since then. This fixes
: that, for the most part. But in the interim changes in the way ops codes are
: renumbered have broken the tests in t/tools/ops2pm/05-renum_op_map_file.t, so
: mosts of its tests are TODOed out.
diff: www.parrotvm.org/svn/parrot/revision?rev=34096
02:28 geof joined
dalek r34097 | util++ | : 02:41
: [codingstd] Added coda to Rakudo's Exception.pir
diff: www.parrotvm.org/svn/parrot/revision?rev=34097
r34098 | chromatic++ | :
02:41 ChrisDavaz joined
dalek : [ops] Made some temporary PMCs used in comparison opcodes *real* temporary PMCs. 02:41
diff: www.parrotvm.org/svn/parrot/revision?rev=34098
r34099 | pmichaud++ | : 02:47
: Remove obsolete rakudo "main" branch.
diff: www.parrotvm.org/svn/parrot/revision?rev=34099
kid51 Do we have a Google Groups set up to archive the parrot-tickets list? 03:02
I know we have this for parrot-dev: groups.google.com/group/parrot-dev/topics
pmichaud We have lists.parrot.org/pipermail/parrot-tickets/ 03:03
but I don't know if a Google Group has been set up yet.
kid51 pmichaud: Thanks. That's good enough for now. 03:05
03:22 Zaba joined
Coke missed kid? 03:30
(have a patch to mostly rip out smoke. will stick it on TT#69 momentarily. 03:31
Coke would love it if TEST_JOBS worked on "smolder_tests"
03:40 Zaba_ joined
Coke You know, I think it might be nice if we combined parrot-dev and parrot-tickets. 03:53
04:02 Andy joined 04:13 Zaba joined 04:22 jimmy joined 04:26 Zaba joined 04:29 tetragon joined 04:30 vaidyagi joined 04:46 Zaba_ joined 05:33 cotto joined 05:52 NotFound joined 05:54 Maddingue joined 05:57 Themeruta joined 05:59 Maddingu1 joined 06:11 TimToady joined
cotto (french toast)++ 06:14
GeJ As a French, I can only agree with you. 06:23
06:27 ask_ joined, Maddingue joined, Hadi joined 06:28 Hadi left 06:33 Hadi joined 06:34 Hadi left
cotto As a French, what do you call it? 06:41
GeJ "Pain Perdu", word-for-word "lost bread" or more accurately "wasted bread" 06:48
Thinking of it : salvage wasted bread by dipping in a mix of mil, egg and sugar = french dish 06:49
cutting-and-slicing potatoes, dropping in boiling oil = french dish...
why do the most simple things get french-ed? 06:50
GeJ goes looking for other "french $something" recipes 06:51
cotto kiss?
purl kiss is Keep It Simple, Stupid!
GeJ like we invented tonguing, right :) 06:52
06:53 Theory joined
GeJ Theory: how's the trip in France going? 06:53
Theory GeJ: Very nicely, thanks! 06:54
A little cold, but it got a bit warmer, and now it's much colder in Portland so I'm glad to be here. :-) 06:55
jimmy november?
purl i guess november is at www.november-wiki.org/ or use.perl.org/~masak/journal/37212 or github.com/viklund/november/
GeJ Theory: seen you've been in Paris (or are you still), did you go by the "Galeries Lafayette"? Your little one might enjoy it. 06:57
Theory We 06:58
We're back in Paris Jan 2-4. Then go home on the 5th.
or 5eme, I should say
Theory googles Galeries Lafayette
dalek r34100 | allison++ | : 06:59
: Creating a branch for the first round of GC refactoring.
diff: www.parrotvm.org/svn/parrot/revision?rev=34100
Theory GeJ: More like my wife would enjoy it! :-P 07:00
GeJ I asked because I saw her last blog entry :) 07:01
Theory ah 07:02
GeJ Around Christmas, they have these use displays filled with animated toys.
Theory ah!
Are they removed before Jan 2? 07:03
GeJ They might still be there. I haven't spend a Christmas in Paris for 4 years now. I couldn't tell for sure.
jimmy why does dalek lost anythin before ':' ? 07:04
Theory you should leave a comment to that effect. :-)
GeJ But I remember bugging my parents to go there every week-ends. 07:05
I must confess that even later I always could find an hour or so at lunch break to pay a visit. 07:06
Theory good to know 07:07
Don't think we'll be back in Paris before xmas
07:37 Zaba joined
cotto msg chromatic I'd appreciate you double-checking my conclusion on CID #161. It looks like a false positive but I might be missing something. 07:47
purl Message for chromatic stored.
07:49 iblechbot joined, funktio joined
GeJ Theory: hum, that's strange... it looks like you were there already. :-/ 08:01
08:09 elmex joined 08:31 TiMBuS joined
Theory GeJ: I thought that the rotunda looked familiar. We didn't really seen any displays, though. A just played in the toy section of the store. 09:02
09:05 Zaba joined 09:25 alvar joined 09:31 Hadi joined, Hadi left 09:49 Hadi joined 09:54 Hadi left 10:00 bacek joined 10:07 jimmy joined 10:18 tomyan joined 10:48 Hadi joined, Hadi left 11:06 kj joined 11:35 Zaba joined 11:39 gaz joined
ask_ is there a policy on adding (or not adding) email addresses in the CREDITS file? 11:54
moritz ask_: I think major contributors should have an email address in CREDITS
ask_: but I have been told that it's not necessary for occasional contributors (especially those without commit access) 11:55
ask_ only counting those with commit access then? Anyway - I'm (via making the git repository) completing the CREDITS file with everyone ever doing a commit.
I don't know if I can put the changes into the repository as it's adding a bunch of email addresses. 11:56
moritz ask_: maybe ask on the list
ask_ what's the list these days?
:)
moritz parrot-dev@lists.parrot.org 11:57
afk
ask_ thanks, sent. 11:59
12:00 iblechbot joined 12:30 kj joined 12:34 tomyan left 12:47 jan joined
dalek r34101 | kjs++ | : 12:53
: [pirc] refactor of codesegment creation which is needed to get things working in future.
: + add runtime checks for lin.scan.reg.alloc; only if it's requested create the data structures. The tests are worth it, as they are only simple checks (probably 1 or 2 machine instructions), whereas data structure creation will invoke memory allocators, which is much more painful.
diff: www.parrotvm.org/svn/parrot/revision?rev=34101
12:55 Coke joined 12:59 Theory joined, riffraff joined
dalek r34102 | kjs++ | : 13:01
: [pirc] add a synopsis section to bcgen.c. This presents overview of how to use the ByteCodeGENerator file.
diff: www.parrotvm.org/svn/parrot/revision?rev=34102
riffraff whis there a way or best practice to define a PGE rule/token to match a big list of words (aptly put in valid.txt) ? 13:02
kj riffraff: what comes to mind would be to match a hash (or maybe array? not sure...) However, that's not implemented yet 13:05
riffraff I see thanks
13:07 tetragon joined
dalek r34103 | kjs++ | : 13:07
: [pirc] Remove old hand-written version of pirc. No point in keeping it; should we decide that a recursive-descent version is the way to go, I'd rather start over with experience and lessons learned with pirc/new.
: + update MANIFEST
diff: www.parrotvm.org/svn/parrot/revision?rev=34103
r34104 | kjs++ | : 13:10
: [pirc] put some stuff into the Memory Hole. and other updates.
diff: www.parrotvm.org/svn/parrot/revision?rev=34104
r34105 | kjs++ | : 13:13
: [pirc] remove old documentation about pirc/src. + update MANIFEST
diff: www.parrotvm.org/svn/parrot/revision?rev=34105
13:23 Theory joined
dalek r34106 | kjs++ | : 13:25
: [config] fix pirc.in; it referred to src/pirlexer, should be new/~
diff: www.parrotvm.org/svn/parrot/revision?rev=34106
13:42 riffraff joined
dalek r34107 | kjs++ | : 13:42
: [config] remove a reference to non-existing file in pirc.in
diff: www.parrotvm.org/svn/parrot/revision?rev=34107
Coke I don't see ask_'s email. 13:44
dalek r34108 | kjs++ | : 13:45
: [pirc] refactoring of grammar + action code.
diff: www.parrotvm.org/svn/parrot/revision?rev=34108
Coke hurls xkcd.com/519/ for Andy 13:46
13:47 pdcawley joined
dalek r34109 | kjs++ | : 14:05
: [pirc] shortcuts and layout updates.
diff: www.parrotvm.org/svn/parrot/revision?rev=34109
14:08 tomyan joined 14:17 Andy joined 14:22 dprelec joined 14:25 dprelec left 14:37 Zaba joined 14:38 alvar joined
Coke msg Andy perlbuzz.com/how-to-contribute.html doesn't actually say where to send suggestions. 14:50
purl Message for andy stored.
Coke msg andy xkcd.com/519/ ; awesome webcomic gives perl some love. 14:51
purl Message for andy stored.
moritz and I thought that Mr. Munroe went on to python...
14:51 gryphon joined 15:05 iblechbot joined
Coke proposes parrot split up it's tests by PDD coverage, ala perl6. 15:09
bah. "its"
15:19 Theory joined
Coke thinks he just got another 204 partcl spec tests working. 15:22
s/working/counted/ 15:23
allison++ # GC refactor!
I expect a 50% improvement in run time! =-)
if I have one part of a repository checked out in dir A, and another in dir B, can I do a commit of both simultaneously and have it be atomic? 15:24
15:26 donaldh joined
jhorwitz i don't think so, but try it. worst it will do is say a dir is not part of the repo. 15:27
Coke (google code keeps the wiki in a top level svn dir, next to trunk. I wanted to commit the wiki update in the same commit with a fix to trunk) 15:28
jhorwitz a noble goal 15:29
Coke too noble, not worth it. =-)
yay, 203 more tests. (off by one error in my estimate. =-)
203/2341 15:30
purl 0.0867150790260572
15:31 Lorn joined
Coke yay, a 9% jump. =-) 15:31
(not bad, the diff was to add one line.)
nopaste "tewk" at 155.97.237.62 pasted ":nsentry, perl6 failure, ideas?" (12 lines) at nopaste.snit.ch/15038 15:36
moritz tewk: is that in a branch? 15:37
tewk nope I can put it in one, or paste a diff, :vtable and :method no longer get put in the ns, so it is probably related to that. 15:40
or I could push it to github 15:41
moritz I can't answer your question either way; if it where in a branch, a missing or wrong update from trunk might have explained it 15:42
tewk Another reason why git is better than svn :) 15:43
dalek r34110 | fperrad++ | : 15:46
: [win32] setup
: - now, update %PATH% when installing
diff: www.parrotvm.org/svn/parrot/revision?rev=34110
Coke msg andy in my defense, I mentioned it here before it was posted on perlbuzz. =-) 16:05
purl Message for andy stored.
dalek r34111 | fperrad++ | : 16:55
: [lisp]
: - fix opcode removed
diff: www.parrotvm.org/svn/parrot/revision?rev=34111
17:01 gaurav joined 17:11 tomyan left 17:15 martin joined, davidfetter joined 17:23 chromatic joined 17:27 Andy joined
dalek r34112 | pmichaud++ | : 17:29
: [rakudo]: spectest-progress.csv update: 250 files, 5213 passing, 0 failing
diff: www.parrotvm.org/svn/parrot/revision?rev=34112
17:30 Zaba joined
dalek r34113 | chromatic++ | : 17:30
: r34132@waterwheel (orig r34101): kjs | 2008-12-19 04:52:21 -0800
: [pirc] refactor of codesegment creation which is needed to get things working in future.
purl i already had it that way, dalek.
dalek : + add runtime checks for lin.scan.reg.alloc; only if it's requested create the data structures. The tests are worth it, as they are only simple checks (probably 1 or 2 machine instructions), whereas data structure creation will invoke memory allocators, which is much more painful.
: r34133@waterwheel (orig r34102): kjs | 2008-12-19 04:59:46 -0800
: [pirc] add a synopsis section to bcgen.c. This presents overview of how to use the ByteCodeGENerator file.
: r34134@waterwheel (orig r34103): kjs | 2008-12-19 05:06:39 -0800
: [pirc] Remove old hand-written version of pirc. No point in keeping it; should we decide that a recursive-descent version is the way to go, I'd rather start over with experience and lessons learned with pirc/new.
: + update MANIFEST
: r34135@waterwheel (orig r34104): kjs | 2008-12-19 05:09:14 -0800
: [pirc] put some stuff into the Memory Hole. and other updates.
: r34136@waterwheel (orig r34105): kjs | 2008-12-19 05:11:51 -0800
: [pirc] remove old documentation about pirc/src. + update MANIFEST
: r34137@waterwheel (orig r34106): kjs | 2008-12-19 05:24:06 -0800
: [config] fix pirc.in; it referred to src/pirlexer, should be new/~
: r34138@waterwheel (orig r34107): kjs | 2008-12-19 05:41:10 -0800
: [config] remove a reference to non-existing file in pirc.in
: r34139@waterwheel (orig r34108): kjs | 2008-12-19 05:44:37 -0800
: [pirc] refactoring of grammar + action code.
: r34140@waterwheel (orig r34109): kjs | 2008-12-19 06:04:16 -0800
: [pirc] shortcuts and layout updates.
: r34141@waterwheel (orig r34110): fperrad | 2008-12-19 07:45:10 -0800
: [win32] setup
: - now, update %PATH% when installing
: r34142@waterwheel (orig r34111): fperrad | 2008-12-19 08:54:27 -0800
: [lisp]
: - fix opcode removed
: r34143@waterwheel (orig r34112): pmichaud | 2008-12-19 09:28:23 -0800
chromatic Huh. That's not right.
dalek : [rakudo]: spectest-progress.csv update: 250 files, 5213 passing, 0 failing
:
diff: www.parrotvm.org/svn/parrot/revision?rev=34113
r34114 | pmichaud++ | :
: [rakudo]: Add another spectest file (S06-other/main.t)
diff: www.parrotvm.org/svn/parrot/revision?rev=34114
Coke ah, dalek isn't broken. 17:38
chromatic is. =-)
tewk pmichaud: ping 17:39
Coke grumbles. 17:59
GAH. 18:00
t_tcl/expr.test was hanging after test # foo; except it wasn't hanging, it was just skipping several gazillion tests, and by default, tcl test is silent about skips. 18:01
(this wouldn't have been an issue except for how slow we are... bah) 18:03
chromatic Isn't it faster than yesterday? 18:19
Coke yes; but it is still embarrasingly slow compared to tcl. 18:20
tcl runs t_tcl/expr.test in 1.2seconds. It's going to take at least 30m on for partcl.
(I'll have a better time estimate once this run finishes) 18:24
though I fear I'm going to run out of memory first. 18:26
is there a theoretical upper limit as to how much memory parrot will try to allocate? 18:27
or will it just happily continue until the OS barfs?
pmichaud S26?
purl rumour has it S26 is perlcabal.org/syn/S26.html
chromatic Coke, it will gorge itself until the OS says no more.
eric256 is there a way to warn from parrot? like 'die' but you know...without the whole dieing bit ;) 18:28
chromatic printerr 18:29
Coke chromatic: can you posit a way for me to tell where my memory is going? 18:30
I presume it's part memory leak, part keeping a ref when I shouldn't.
eric256 thanks ;) 18:31
chromatic Valgrind can show memory leaks and Callgrind/Kcachegrind can show you what allocates memory. 18:33
pmichaud RT slow. :-( 18:38
Coke 2.6gb and climbing 18:42
martin eric256, $*ERR.say: 'some warning'; comes close 18:43
eric256 martin: thanks i meant in parrot so printerr is perfect ;) 18:44
Coke gives up.
(trying to run expr.test)
pmichaud rakudo: warn "This is a warning!" 18:47
polyglotbot OUTPUT[This is a warning!ā¤]
pmichaud eric256: try "warn" :-)
oh, from parrot.
rakudo: q:PIR { printerr "This is a warning!" } 18:48
polyglotbot OUTPUT[This is a warning!]
PacoLinux Coke: I'm testing partcl spectest - now in t_tcl/ioUtil.test and I'm using 846/848 Mb (total of all the procs of the machine)
pmichaud sad but true... polyglotbot can execute limited PIR statements :-P
eric256 is debuging the object creation with parent attributes the old fashion brute force way
pmichaud eric256: jonathan++ and I are expecting to refactor that code soon 18:49
well, "soon" == "within the next two weeks"
you're welcome to hack on it, though. Just note that the WHENCE properties are currently not being constructed properly. 18:50
(currently WHENCE is being built as a Hash, where it really ought to be a Closure)
afk, more errands 18:51
eric256 okay 18:52
Coke PacoLinux: patches desperately welcome.
eric256 well most likely i wont be able to fix anything anyway ;)
18:53 ffwonko joined 18:55 ff-wonko joined
nopaste "Coke" at 72.228.52.192 pasted "partcl memory issues" (4 lines) at nopaste.snit.ch/15042 18:57
Coke pmichaud: any "make PGE go faster or use less memory" tasks you can offload?
chromatic: kcachegrind is GUI? 19:01
Coke wonders if he can run that on feather and export the display back to his windows box.
chromatic Yes. 19:02
Coke hurm. looks like I could also theoretically run it on my mac.
chromatic If you bzip the callgrind.out.xxxxx file, I can run it too. 19:03
bzip and transfer, I mean.
Coke let me try to get it running locally on teh mac.
Coke peers at brent.izolo.com/blog/?p=4
19:05 ruoso joined 19:06 ruoso_ joined 19:08 Theory joined
Coke (this will be very nice if it works.) 19:09
Coke apparently has not updated fink in aeons. 19:10
The following package will be installed or updated: kcachegrind 19:16
The following 92 additional packages will be installed:
moose:cpan::kcachegrind:fink? 19:20
19:20 Whiteknight joined
Whiteknight I borked my laptop MBR, accidentally deleted my main partition while trying to fix it (thanks GRUB FAQ!) and had to reinstall everything 19:22
19:25 ask_ joined
Coke oof 19:27
fink is finally done /downloading/ the 92 packages, and is now building them. =-) 19:41
chromatic What I like about Mac OS X is that it just works, as long as His Steveness allows you to do something, and that His Steveness punishes you for deviating from that. 19:42
Mostly I like that because I don't have to suffer that anymore.
Coke I am sure there is some cool mac tool for doing what I want here. 19:45
Whiteknight Well, it was my fault, I tried to add a third boot option because I wasn't satisfied with the first two 19:46
but the third one didn't work out, so I deleted the partition, and it took GRUB with it 19:47
chromatic I think there's an iTunes plugin for Kcachegrind actually.
Coke cannot install kcachegrind. 19:50
Coke has X running on os x, and will try the feather trick. 19:51
chromatic: is kcg available on feather already? 19:53
(wow does that font bring back memories) 19:55
lu_zero feather trick? 19:57
Coke ssh -X'ing to feather. 19:58
lu_zero works nicely
Coke slow as hell, but otherwise yes. =-) 19:59
chromatic I don't use feathre; I don't know.
lu_zero likes better macosx just for adium and linux for the rest
Whiteknight one day, maybe Parrot will bring platform ubiquity
lu_zero also uses gentoo in both of them
Coke Whiteknight: I highly doubt that. =-) 20:00
lu_zero Whiteknight I hope not =P
Whiteknight aim for the skys, damnit!
Coke I'd like to actually hit the 1.0 milestone, thanks. =-)
lu_zero would love to have parrot bundled in firefox and webkit
but before that have it installable =P 20:01
Tene Coke: if you'd rather ssh to my laptop, I can install kcachegrind here.
Probably faster.
Coke I can do that, but partcl will kill your laptop. no? =-)
Tene It will? 20:02
Coke I'd like to have it setup somewhere where I can use it whenver; thanks though.
Tene Ah, okay. I thought you were just analyzing one trace.
lu_zero Coke otherwise you may use virtualbox and make it explode for you =) 20:03
lu_zero likes better kvm
Infinoid kvm works well.
lu_zero Infinoid on macosx? 20:07
Infinoid lu_zero: nah, gentoo linux 20:10
Coke gives up on kcachegrind and mallocdebug and kicks something. 20:16
PerlJam Coke: tcl still problematic with newer parrots? 20:19
lu_zero Coke dtrace won't help you?
Coke PerlJam: problematic? just (@#*&$# slow. 20:21
now that I can actually /run/ things again, the fact that they are painfully slow is far more obvious. =-) 20:22
20:22 gmansi joined
chromatic Coke, do you have a callgrind profile? 20:22
20:26 mj41 joined
chromatic TimToady, ping 20:29
TimToady you pang? 20:30
Coke chromatic: no. I'll get one later. 20:34
chromatic TimToady, did you get the contract?
TimToady on whom?
PerlJam If Larry is Lurch, who plays the other members of the Addams Family? 20:35
chromatic On the Masterminds interviews.
TimToady no, but Glo speaks French
snail? 20:36
purl snail is an invertebrate, right?
20:36 Theory joined 20:37 mberends joined
chromatic I believe I mailed it to you with the most recent interview questions. 20:37
TimToady when?
chromatic 12-05 20:38
purl 7
chromatic "Masterminds Contract"
TimToady maybe it went to the North Pole...
chromatic I'll re-send. 20:39
There. Now I resent it. 20:40
So to speak.
purl absolutely, don't you know
Whiteknight is behind on his blogging. Hiveminder emails are getting more and more angry sounding 20:48
20:52 contingencyplan joined
dalek r34115 | Whiteknight++ | : 21:01
: [pdd09gc_part1] separating the MS collector into it's own file, sanity checking a change that I want to make to trunk later
diff: www.parrotvm.org/svn/parrot/revision?rev=34115
r34116 | moritz++ | : 21:03
: [rakudo] make ceil(), floor() etc. return Int values, not Num. cspencer++,
: patch courtesy by Cory Spencer (cspencer at sprocket.org). Closes RT #61526.
diff: www.parrotvm.org/svn/parrot/revision?rev=34116
21:04 Eevee joined 21:05 alvar joined 21:09 Theory joined
GeJ good morning everyone 21:19
pmichaud good morning, GeJ 21:22
21:36 donaldh joined 21:44 mj41_ joined 21:52 gryphon joined 21:57 tetragon joined
dalek r34117 | util++ | : 22:08
: [codingstd] Fixed cut-and-paste POD error and added =back to PCT's PAST/Compiler.pir
diff: www.parrotvm.org/svn/parrot/revision?rev=34117
r34118 | util++ | : 22:11
: [codingstd] Fixed typo in POD in IMCC's parser_util.c
diff: www.parrotvm.org/svn/parrot/revision?rev=34118
22:15 Theory joined
dalek r34119 | util++ | : 22:34
: [codingstd] Removed leftover =back in POD of io/api.c
diff: www.parrotvm.org/svn/parrot/revision?rev=34119
22:37 ElPenguin joined 22:38 Whiteknight joined
bacek Good morning 22:50
Whiteknight good morning bacek 22:51
bacek Any particular tasks for rambling developer? :) 22:52
Whiteknight depends, what kinds of stuff do you like to develop? 22:53
dalek r34120 | moritz++ | : 22:55
: [rakudo] another file for spectest.data: test passing of optional parameters
diff: www.parrotvm.org/svn/parrot/revision?rev=34120
bacek Whiteknight: not GC :) 22:57
Whiteknight that's ok, GC isn't really programming: it's black magic 22:59
chromatic I thought it was a Shinto Shrine. 23:00
bacek Whiteknight: it's not quite true. But it's very low-level stuff. And my C-fu isn't high
pmichaud It's right next to the shrine for "git".
Whiteknight but git doesn't require you to sacrifice a pound of flesh to work on it 23:01
bacek heard rumours about moving parrot to git
chromatic If only git required a pound of flesh.
"Dear git, I don't want to write my own version control system using your plumbing. I just want to create a patch! Is that so hard?" 23:02
cotto as in a pound of smoked salmon? That doesn't sound so bad.
bacek chromatic: it is... easy.
chromatic I'd believe it were easy if any Git tutorial gave an explanation of how to do it without first explaining category theory, the Riemann hypothesis, unlambda, and the sexy beauty of the git filesystem, which apparently you really really need to understand before running git-diff. 23:05
moritz chromatic++ 23:06
Tene I didn't learn about any of that until after I had been using git for a while.
chromatic I ran into this when trying to patch Perl 5, which previously meant running rsync to get bleadperl and manually moving and diffing the files that you'd modified, then concatenating individual file diffs into a patch file. 23:07
Turns out that was *easier* than generating a single diff from git.
I don't want to publish my own tree.
I don't want to maintain my own branch.
I don't want to freshen my breath, whiten my teeth, and increase my magnetism with members of the appropriate gender. 23:08
I want to check out a project, modify a few files, and generate a patch.
bacek chromatic: www-cs-students.stanford.edu/~blynn.../ch02.html
chromatic: git clone some-repo; <hack hack hack>; git diff . 23:09
chromatic That doesn't answer my questions though.
bacek November?
purl November is at www.november-wiki.org/ or use.perl.org/~masak/journal/37212 or github.com/viklund/november/
Tene I don't see any questions beyond 'how do I generate a diff?' 23:10
chromatic In my experience, that's missing the magic git add foo.c and git ci foo.c.
Then once you've done that, git diff starts generating multiple patch files for you, which is also not what I wanted.
Tene If you're not wanting to deal with git at all, then don't, and just 'git diff' will show all changes. 23:11
chromatic In my experience, it didn't.
bacek chromatic: interesting...
PerlJam chromatic: if you did "git add" first, then you want "git diff --cached" (weird, I know)
Tene If you can reproduce, I'm interested to see. I can't. I rely on that behavior of git regularly.
chromatic Now it may have been that the (severe lack of comprehensible) instructions for checking out the Perl 5 sources made me do something far different and wrong, but that's my experience. 23:12
bacek chromatic: anyway. You can always squash few commits into one
chromatic In particular, I think they had me create a local branch.
pmichaud while we're on the topic of git, there has been some discussion about moving rakudo into a git repository
bacek chromatic: you always creates local branch with git
Tene If you've been doing commits, 'git diff origin' or whatever will show a diff of every difference you've committed.
bacek chromatic: git just works in this way 23:13
chromatic But git didn't just work.
git diff origin didn't just work either.
pmichaud there's also discussion that perhaps someone else can take the lead for making that happen. Anyone who might be interested, contact me. (But do it later, as things are a bit hectic here now.)
Tene If you used 'git-format-patch', it would generate multiple patches. 'git diff' will just make a plain patch.
chromatic git diff made multiple patches.
I accept that either git or Obama is the second coming for various values of messianic properties, but you're describing things it didn't do for me. 23:14
That's my frustration.
PerlJam chromatic: show us a transcript :)
Tene What you're describing is quite different from my experience. Every time I have multiple commits, I rely on 'git diff origin' to show me the full differences, in one patch, of everything I'm committing. 23:15
Whiteknight whoa whoa whoa, back it up a bit! Can git really increase your magnetism to members of the opposite gender?!?
PerlJam Whiteknight: no, only an appropriate gender, where you don't get to decide what it appropriate/
Tene Every time I have changes that I haven't committed, I use 'git diff' before committing. It always behaves appropriately.
nopaste "bacek" at 87.250.243.86 pasted "Simple work with git for chromatic" (21 lines) at nopaste.snit.ch/15043
chromatic I'd love to show you a transcript, but git diff origin mysteriously started working for me.
PerlJam Whiteknight: unfortunately, it's usually males as they become in awe of your git-fu.
chromatic: pebkac then! :) 23:16
Whiteknight meh, I take what I can get
chromatic I have no doubt that it's a wonderful system, but I find it incomprehensible. 23:17
How do I revert local changes to a file?
Tene checkout
purl hmmm... checkout is still running
Tene git checkout file
PerlJam chromatic: git is a bit impenatrable initially. Once you get over the learning hump of understanding the index and how all of the commands interact with it, its gets *much* easier. 23:18
chromatic I'm sure it does.
Infinoid ...and once you get used to the weird interactions between stgit and git-svn...
chromatic I don't mind if it has a learning curve. I've just never seen a tutorial that assumed I want to hack on a project for a couple of hours without having to learn its internals. 23:19
Maybe you do have to learn its internals to use it. That's fine, I suppose.
Tene chromatic: want me to write one for you?
PerlJam chromatic: if you haven't seen it already, Randal's git google techtalk is fairly good. 23:20
Tene: And put it on the wiki! :)
chromatic "Using git with Parrot" would be very handy.
bacek chromatic: "git crash course for svn users"?
chromatic It will likely come in useful for Rakudo.
PerlJam chromatic: www.youtube.com/watch?v=8dhZ9BXQgc4 <-- merlyn's techtalk
chromatic bacek, I tried that one, and it didn't help. Maybe that's just me.
Tene chromatic: preferred tutorial format? 23:21
chromatic Tene, assume I'm a non-committer who wants to download the project, stay up to date, and occasionally submit a patch.
PerlJam wait ... does parrot actually have an "official" git repo? 23:22
Tene Okay. I'll write that tonight.
PerlJam Or are we still talking git-svn here? 23:23
Tene driving home now.
chromatic Assume git-svn for now, but things should apply pretty well if and when Rakudo moves to git. 23:24
bacek isn't commiter, who downloaded project, stays up to date and occasionally submit patches 23:25
Using git-svn
PerlJam too
(though I do that last bit far less frequently than I should)
chromatic I thought you had a commit bit, PerlJam. 23:26
PerlJam well ... I *am* a committer, but I have submitted patches on ares where I was unsure of the patch.
a/ares/areaas/
grr
chromatic Want to work on a GC?
PerlJam GC is completely opaque to me other than general concepts. 23:27
chromatic Then you're qualified.
PerlJam heh
pmichaud anyone who claims to understand GC is lying. :-)
chromatic Actually, we could still use people reviewing TODO/SKIP tests to make sure they all have RT or TT numbers.
... if anyone's looking for something to work on today. 23:28
PerlJam I'd like to work on something where I feel like I'm able to make progress. Unfortunately, I think I'm the only one who can figure that out :) 23:30
chromatic We don't know our current state of what we expect in the tests, and it is a milestone task, so it is progress... at least from the project's point of view. I can't give you internal motivation. 23:32
bacek, you're also welcome to help there.
AS IS ANY LURKER HELLO WAKE UP EVERYONE! 23:33
Whiteknight not me, I've got too much other stuff to do 23:35
Infinoid Lurk, lurk. 23:48
23:48 ruoso_ joined 23:57 eric256 left

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!