Parrot 2.10.0 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 16 November 2010.
00:03 dngor_ left 00:06 dngor joined 00:07 seanstickle joined
NotFound Some info about the current GC: If I drop the explicit sweep in winxed examples/fly.winxed memory usage jumps from 0.5% to 11% and it makes clearly viewable pauses every few seconds 00:07
dukeleto NotFound: the gc constants are not tuned well for your machine, by default 00:09
NotFound Maybe the PI constant is not tuned well for my universe %-) 00:10
00:12 donaldh left
cotto_work NotFound: which universe do you live in? Do you have a team name? 00:12
NotFound Anyway, is better thatn the last time I tested, were memory usage constantly increased. 00:13
cotto_work: not sure, but looks like Crysis in infinite earths.
00:25 stilgar left
whiteknight bacek: is that faster than it was last time you showed me? 00:25
00:28 bluescreen joined
cotto_work It looks like it less than 1/2 the speed. 00:28
previously it was ~45% iirc
whiteknight that's absurd 00:29
I vote bacek for king of the universe 00:30
cotto_work gerd++
whiteknight gerd++ indeed 00:32
though every time gerd puts out a release announcement I need to greenlight it on the lists
cotto_work seen gred 00:33
aloha Sorry, I haven't seen gred.
cotto_work seen gerd
aloha gerd was last seen in #parrot 7 hours 49 mins ago saying "I have to leave for fechting my doughter, after that I will look here again, good luck".
cotto_work msg gerd It's not necessary to send announcements for non-critical bugfix releases to lwn. 00:35
aloha OK. I'll deliver the message.
00:41 gerd joined
bluescreen whiteknight -> TT#1854 00:41
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 00:42
00:43 gerd left
whiteknight bluescreen: thanks! 00:49
I have some fixes I'm about to commit in
dalek TT #1854 created by bluescreen++: Re-wrte of the embedded API 00:50
TT #1854: trac.parrot.org/parrot/ticket/1854
bluescreen typo! >D
:D
am I condoned given the fact english is not my mother tongue ?
whiteknight bluescreen: typos are ok 00:54
cotto_work bluescreen: English is bacek_at_work's 4th language 00:57
bluescreen youi thnk sso?
what? trying to make me feel bad cotto_work?
only 2 languages is not enough...? 00:58
cotto_work no, I'm just saying that not everyone is a native speaker
trac.parrot.org/parrot/wiki/ParrotQuotes <-look for "English"
bluescreen I'm wondering what the other 3 are: C, Perl, PIR? 00:59
lol
bacek_at_work Russian, Rude Russian and Very Rude Russian :) 01:00
bluescreen I bet there are some utf8 test cases using " Very rude russian" 01:02
whiteknight unicode CYRILLIC ZHE WITH ANGER 01:04
bluescreen :) 01:07
One of my favorites software quotes is: "Most software today is very much like an Egyptian pyramid with millions of bricks piled on top of each other, with no structural integrity, but just done by brute force and thousands of slaves. " -- Alan Kay
dukeleto bluescreen: that is a great quote. 01:10
01:11 theory left 01:12 seanstickle left
whiteknight bluescreen: I just pushed a bunch of fixes 01:19
bluescreen let's see that... Have you finished Parrot_api_make_interpreter? 01:20
whiteknight I haven't touched it 01:22
bluescreen I'm scare of that one :) it has too many implications 01:23
dukeleto gerd++ 01:24
dalek rrot/gc_ms2_faster_list: e57c949 | mikehh++ | MANIFEST:
re-generate MANIFEST
 01:27
rrot/gc_ms2_faster_list: c70b078 | mikehh++ | src/pointer_array.c:
add missing coda
rrot/gc_ms2_faster_list: 96148ad | mikehh++ | src/pointer_array.c:
add missing ASSERT_ARGS
rrot/gc_ms2_faster_list: 82f9d72 | mikehh++ | src/pointer_array.c:
Add missing documentation

this needs to be checked for completeness and possibly needed additions
dukeleto do we need CLA's for HLL patches on github?
someone sent a pull request to make ecmascript compile on parrot master. It would be nice to merge it. 01:28
whiteknight I imagine a pull request is the same as any other submitted patch 01:30
and we apply patches from people without CLAs all the time 01:31
the licensing info of the repo is clearly indicated
dukeleto before, the committer was the person applying the patch, but now, in git, the person who submitted the patch has their name attached 01:32
but, doesn't matter to me. 01:33
whiteknight that seems to me like only a difference in attribution
dukeleto merged it in: github.com/parrot/ecmascript/commi...b29624fe32 01:35
mikehh msg bacek_at_work still c++ comments and I didn't want to mess with the macro POINTER_ARRAY_ITER(_array, _code) [_array is unwrapped] (well so is _code, but we avoid testing that) 01:36
aloha OK. I'll deliver the message.
dalek mascript: a432c4a | bnoordhuis++ | / (2 files):
Minor changes to make ecmascript compile with the latest and greatest parrot.
 01:37
cotto ~~ 01:40
whiteknight ...we need to rename that damn compiler
mikehh just call is js 01:41
cotto pjs 01:42
whiteknight ha. ecmascript does build but trying to run the test suite is a nightmare
mikehh better
whiteknight every test aborts
I was thinking "CuckooShrike"
cotto hope you don't get any of that bitrot in your lungs
whiteknight a parrot-like bird that lives on the island of Java
cotto nice 01:43
whiteknight actually, only the Javan Cuckooshrike lives on Java. the rest, not so much 01:44
I've become extremely good at researching Parrots on Wikipedia
bluescreen why do we have interp in longopt.c ? it doesn't seems to do anything with it... besides copying it into the stack many times :D
whiteknight yeah, I think we can remove it now 01:46
I changed all the Parrot_snprintf calls to ordinary snprintf
so now the interp is useless 01:47
bluescreen is there any policy around that there are many aliases to the same thing.. i.e Parrot_Interp .. parrot_interp_t .. Interp *
yes..
whiteknight Parrot_PMC is for use by embedders. PMC* is internal
struct parrot_pmc_t, or whatever it's called, is only used in .h files 01:48
the interp won't be used in the embedding API, only Parrot_PMC
bluescreen but...what about INTVAL, Parrot_Int why should we use one or the other?
whiteknight Parrot_Int is for embedders. INTVAL is internal 01:49
embedders should use the Parrot_* ones, to prevent naming collisions 01:50
bluescreen It makes sense... I think I'll be picking them up as we go along
whiteknight++ for decoupling longopt.c from interp 01:51
whiteknight it's not perfect yet
cotto much needed
whiteknight other programs like pbc_disassemble still expect it to be in libparrot
01:52 theory joined
mikehh whiteknight: Java-moustached Parakeets 01:53
whiteknight mikehh: yeah, doesn't flow off the tongue as well
mikehh whiteknight: anything is better than ecmascript 01:55
bluescreen pbc_dissasemble, pbc_dump, test_main.c and pbc_merge.c are using it
whiteknight mikehh: exactly! 01:56
bluescreen: yeah, it's a mess
bluescreen fortunately the scope is not huge 01:57
whiteknight true 02:00
an ambitious future step might be to move some of these source files out of src/
bluescreen maybe tools... ? 02:04
bacek_at_work ~~ 02:13
whiteknight hello bacek_at_work
bacek_at_work aloha, whiteknight
mikehh, ping
dalek rrot/gc_ms2_faster_list: 0ea7e52 | bacek++ | / (2 files):
Merge branch 'gc_ms2_faster_list' of github.com:parrot/parrot into gc_ms2_faster_list
rrot/gc_ms2_faster_list: 72fc70c | bacek++ | src/pointer_array.c:
Update docs for Pointer_Array.
bacek_at_work mikehh, where I did miss c++ comments?
rrot/gc_ms2_faster_list: 78cb67f | bacek++ | include/parrot/pointer_array.h:
Wrap _array macro parameter.
whiteknight is that merged into trunk? 02:14
bacek_at_work no...
bluescreen whiteknight: let me know what you think of my last commit in your fork... finished the clean up in longopt and its dependencies
whiteknight bluescreen: okay. I'm going to bed now, I'll look at it tomorrow 02:15
Feel free to commit whatever. I'll see it when I get up
bluescreen is it 8pm there, isnt it? 02:16
whiteknight 9:15
bluescreen here is 23
whiteknight I was up most of last night with a sick kid, so I'm going to bed early tonight
bluescreen and havent had dinner yet
whiteknight bluescreen: you're running late!
I'm going to bed now. Goodnight 02:17
bluescreen yeah... that's another story with kids you sleep when you can
bacek_at_work good night, whiteknight
whiteknight :)
02:17 whiteknight left
bluescreen night 02:17
02:18 dngor_ joined, dngor left
mikehh bacek_at_work: // FIXME. Cast to UINTVAL is wrong - line 103 of src/pointer_array.c 02:25
bacek_at_work mikehh, yeah... Found it. 02:26
Unfortunately uintptr_t is in C99. And we have to support VC6 with C89...
02:26 dngor_ is now known as dngor
mikehh bacek_at_work: yeah - going in 2011 soon and they still working with C89 - 22 years :-} 02:28
bacek_at_work mikehh, enterprise level stability! 02:29
Ho-ho-ho. Rakudo's "make spectest" challenge: master vs gc_ms2_faster_list - 34m vs 24m. 02:34
Looks like we can save a lot of time for Rakudo's dev :)
02:35 rurban_ joined 02:38 rurban left, rurban_ is now known as rurban
dalek TT #1855 created by bacek++: Add UINTPTR type into parrot. 02:40
TT #1855: trac.parrot.org/parrot/ticket/1855
rrot/gc_ms2_faster_list: b27a3b8 | bacek++ | src/pointer_array.c:
Fix c++ comment and add reference to ticket number.
 02:41
rrot/gc_ms2_faster_list: 3f626ce | bacek++ | / (56 files):
Merge branch 'master' into gc_ms2_faster_list

Conflicts:
  \tMANIFEST		 02:56
rrot: 33f0e17 | bacek++ | tools/dev/mk_gitignore.pl:
Remove old mk_gitignore tool to generate .gitignore from svn-based MANIFEST.SKIP. We are in git now
 02:59
dukeleto bacek_at_work: i am sure they will like the sound of that 03:09
bacek_at_work dukeleto, I hope so :) 03:15
03:19 bluescreen left
mikehh bacek_at_work: I am getting a g++ error -> src/pointer_array.c:143:55: error: invalid conversion from ‘void**’ to ‘int’ 03:22
bacek_at_work mikehh, can you try this: *(int*)ptr = (void**)(((UINTVAL)self->next_free) | 1); 03:23
just add parenthesis around (UINTVAL)self->next_free 03:24
dalek rrot: 2897df3 | jkeenan++ | tools/dev/mk_gitignore.pl:
Merge branch 'master' of git@github.com:parrot/parrot
 03:27
rrot: 649a9d5 | jkeenan++ | MANIFEST:
Update MANIFEST to reflect deletion of tools/dev/mk_gitignore.pl.
rrot: 5ac740e | jkeenan++ | t/steps/auto/ (2 files):
Skip tests when .git not present (tarball).
03:35 bluescreen joined
mikehh bacek_at_work: nope that doesn't work - src/pointer_array.c:143:57: error: invalid conversion from ‘void**’ to ‘int’ (moves it from pos 55 to pos 57) 03:36
bacek_at_work: maybe it don't like the 1 03:37
bacek_at_work: gcc says - src/pointer_array.c:143:16: warning: assignment makes integer from pointer without a cast 03:43
dukeleto looks like we don't have a tag for 2.10.1 03:44
aloha, msg gerd can you push a tag for 2.10.1 ? 03:45
aloha dukeleto: OK. I'll deliver the message.
bacek_at_work mikehh, what is command to configure parrot to us g++? 03:52
dukeleto bacek_at_work: you want --cc=.. and --ld=.. and maybe --ccflags=... 03:56
bacek_at_work dukeleto, --link actually
dukeleto bacek_at_work: ah! 03:57
dalek rrot/gc_ms2_faster_list: aa53013 | bacek++ | src/gc/gc_ms2.c:
Fix c++ compilation
rrot/gc_ms2_faster_list: ee296a7 | bacek++ | src/pointer_array.c:
Fix c++ compilation
bacek_at_work mikehh, looks like fixed :)
dukeleto bacek_at_work: you might like github.com/leto/Util/blob/master/bin/new_parrot 03:58
bacek_at_work: ccache speeds up things quite a bunch 03:59
bacek_at_work: but i don't have it on every machine
bacek_at_work dukeleto, I use ccache all the time
dukeleto bacek_at_work: good to hear :)
mikehh bacek_at_work: I use -> perl Configure.pl --test --cc=g++ --cxx=g++ --link=g++ --ld=g++ (plus other options as necessary)
dalek rrot/gc_ms2_faster_list: 68e4b2e | bacek++ | t/src/pointer_array.t:
Include embed.h to run tests on c++ build.
 04:01
04:03 plobsing left
mikehh bacek_at_work: ok that built ok with g++ - t/src/pointer_array.t - Failed test: 2 [at t/src/pointer_array.t line 93. - Exited with error code: [SIGNAL 11]] looks like a segfault to me 04:16
that's on Ubuntu 10.10 (64 bit) 04:19
04:23 plobsing joined 05:37 bluescreen left
bacek_at_work mikehh, ping 06:12
can you provide output of prove -v t/src/pointer_array.t?
dalek rrot: 440216f | plobsing++ | t/steps/auto/ (2 files):
[codetest] trailing space
 07:18
07:46 jan left
cotto any objections to holding the pds in #parrotsketch? 08:08
plobsing I don't know. that seems like such a sketchy place ;) 08:09
moritz if you want any other channel publicly logged, just talk to me 08:11
cotto msg kid51 do you have some kind of agenda for the pds? I just ask because you had a pretty good one for the pdx meeting.
aloha OK. I'll deliver the message.
cotto moritz, good to know
08:36 theory left
mikehh bacek_at_work: there's nothing really informative about it, but here it is 09:18
nopaste "mikehh" at 192.168.1.3 pasted "gc_ms2_faster_list branch prove -v t/src/pointer_array.t" (31 lines) at nopaste.snit.ch/25765
09:49 whiteknight joined
he Can someone please do 09:51
sed -e 's/freebsd/netbsd/ <config/gen/platform/freebsd/sysmem.c > config/gen/platform/netbsd/sysmem.c
and commit the result?
(well, add the missing ' after netbsd/, of course) 09:52
whiteknight he: done 10:00
dalek rrot: c35e562 | Whiteknight++ | config/gen/platform/netbsd/sysmem.c:
adding in a netbsd sysmem file for he++
rrot: 28b24ab | Whiteknight++ | / (18 files):
Merge branch 'master' of github.com:parrot/parrot
whiteknight ah damnit. forgot the --rebase
he whiteknight: thanks! 10:03
10:06 mikehh left 10:08 contingencyplan left 10:32 moritz left, moritz joined 10:34 rurban_ joined 10:38 rurban left, rurban_ is now known as rurban
zby seen chromatic? 10:42
aloha chromatic was last seen in #parrot 5 days 20 hours ago joining the channel.
10:42 fperrad joined 10:51 particle left 10:53 particle joined 10:55 mikehh joined
mikehh bacek_at_work: no problem with 32-bit, just the 64-bit 11:03
11:07 dip left, dip joined
mikehh gc_ms2_faster_lists branch - All tests PASS (pre/post-config, make corevm/make coretest, test, fulltest) at 68e4b2e - Ubuntu 10.10 i386 (g++-4.5) 11:28
gc_ms2_faster_lists branch - t/src/pointer_array.t FAILs test2 on Ubuntu 10.10 amd64 (g++-4.5 and gcc-4.5) 11:30
tadzik ~~ 11:55
12:03 mikehh left 12:19 whiteknight left
dalek rrot/gc_ms2_faster_list: b4e4e81 | plobsing++ | src/pointer_array.c:
fix pointer arrays where sizeof (int) != sizeof (void *)

also silence warning about undefined operation on self->chunks		 12:23
plobsing t/src/pointer_array.t should pass now on 64 bit 12:24
12:43 mikehh joined 12:56 bluescreen joined
dalek TT #1856 created by doughera++: Build error in 2.10.1 13:06
TT #1856: trac.parrot.org/parrot/ticket/1856
13:08 jsut_ joined
Coke Ops/Emitter compilation slllllow. 13:11
13:13 jsut left
dalek rrot/gc_ms2_faster_list: a393669 | mikehh++ | src/pointer_array.c:
Will not work with g++ - use UINTVAL instead of int - we need an int type here
 13:25
13:26 PerlPilot joined 13:30 PerlJam left 13:51 whiteknight joined 13:57 darbelo joined
mikehh gc_ms2_faster list branch: 14:07
All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1178) fulltest) at a393669 - Ubuntu 10.10 amd64 (g++-4.5 with --optimize)
we need to bump VERSION - rakudo requires 2.10.1 to build 14:14
14:22 mikehh_ joined 14:26 mikehh left 14:32 PerlPilot is now known as PerlJam 14:33 mikehh_ is now known as mikehh
mikehh opbots, names 14:33
whiteknight does rakudo actually build on 2.10.1? Andy's last message to the list suggests otherwise 14:34
dalek TT #1856 closed by gerd++: Build error in 2.10.1 14:40
TT #1856: trac.parrot.org/parrot/ticket/1856
PerlJam www.parrot.org/download gives parrot-2.10.0.tar.gz as the current developer release. 14:47
14:51 mikehh left 14:52 PerlJam left, PerlJam joined
whiteknight fixed. Thanks PerlJam++ 14:55
14:59 mikehh joined
Coke are we going to do a 2.10.2? that'd be a record! ;) 15:13
whiteknight I don't think so. It looks like gerd updated the 2.10.1 tarball in place 15:14
Coke sneaky. Was the tag updated? 15:15
particle sigh 15:16
mikehh we all need to make sure that master reflects the changes as well 15:19
moritz the last parrot 2.10.1 I built identified itself as 2.10.0
moritz would vote for a policy to leave releases immutable 15:20
and if something's screwed up, just make another point release
you might get a .2 or .3 at some point, but it's less confusing
mikehh plus branches (what I am testin' now - gc_ms2_faster_list) 15:22
15:22 notbenh left
whiteknight we all knew that the first release after the migration would have some roadbumps 15:24
an immutable release policy is not a bad one, no
mikehh just a few potholes :-}
moritz that's fine by me 15:26
it's why I'm trying to make constructive arguments, and not ranting senselessly about errors :-)
PerlJam It would certainly preserve history 15:29
i.e., if someone had a problem with a parrot tarball, you can easily see that they got one that was broken by looking at the point release. If you update in place, you can't tell. 15:30
But in either case, the solution would be the same: download the tarball and try again
(the right tarball :)
whiteknight it wasn't a bugfix or anything in the tarball. I think gerd missed a step and corrected it
so it's not really a new release
mikehh I think it was the ./ops2c --core step 15:33
whiteknight yeah, that's what his last message said 15:34
mikehh lemme look
whiteknight On the bright side, 2.11 should be much easier :)
mikehh yeah, that's what it was 15:35
moritz <devil's advocate>why isn't there a single release script that enters the version in all necessary files, and runs all the necessary steps?</*> 15:36
mikehh 'cause nobody done wrote one :-} 15:37
moritz so it's no misterious "it requires human interaction" thing?
15:37 theory joined
moritz that's actually good news for me :-) 15:38
whiteknight If anybody wants to write such a script, they would get a whole bag of karma from me 15:39
PerlJam moritz seems to be volunteering :)
whiteknight PerlJam: yeah, that's what I heard
:)
PerlJam moritz++ (for asking the right questions if nothing else) 15:40
moritz PerlJam: indeed, but I don't promise timely delivery (and if anybody beats me to it, they'll get karma from me too :-)
mikehh we also could do with some tests to check that everything has been done, over and above running build up to fulltest 15:42
PerlJam reads parrot's release manager guide
15:42 Andy joined
PerlJam Step 2 would be *greatly* simplified by a nice script 15:45
it's a 10 part (!) step 15:46
whiteknight I'm 2.11, cotto has 3.0, and mikehh I think has 3.1, and kid51 has 3.6
so if anybody else wants to volunteer to jump in that list and show us all how it's done, please do
15:53 mikehh left
whiteknight some days I really like WPF. Some days I really hate it 15:59
moritz wpf? 16:00
16:02 dmalcolm joined
whiteknight Windows Presentation Foundation. It's the new UI garbage from MS 16:03
moritz the context sensitive ribbons and all that? 16:04
I use office rather seldom, and I'm not an expert at all. All in all it seems to make finding stuff easier 16:05
but maybe they could have achieved the same thing by reorganizing the modules from scratch too
16:08 hercynium joined 16:10 seanstickle joined
whiteknight The ribbons are just one widget. WPF is the entire framework for all user interfaces 16:16
16:18 contingencyplan joined 16:32 mikehh joined
dukeleto 'ello 16:53
whiteknight good morning, duke
dukeleto what is the good word?
whiteknight "Potato"
I just bumped up my RAM limit on my VM. I'm going to see if I can build Rakudo with it now 16:58
it hung with 512Mb. I'm hoping it does better with 1024
dukeleto prefers "Tomato", because those pay my bills 17:04
whiteknight really? where do you work? 17:07
ah, nevermind. I'm looking at your profile on linked in 17:08
THE INTERNET KNOWS ALL
dukeleto lollerskates
whiteknight: i work on solgenomics.net, which is a site that aggregates bioinformatics data about solanaceae plants. Tomatos are the biggest cash crop, but potato is in that family as well 17:09
whiteknight: people know a lot more about the tomato genome than the potato genome, though. 17:10
17:15 Kovensky left
whiteknight that explains why you're connected to a bunch of solgenomics projects on github. 17:16
blah. compiling perl6.c failed with clang. Could not allocate enough mem 17:19
dukeleto whiteknight: yeah, I converted them over to git as well. They have 18 interdependent repos. A lot more fun than parrot.git :) 17:21
Is it just me or has Parrot development sped up since it went on Github?
dukeleto is looking at github.com/organizations/parrot and seeing lots of action
whiteknight yes, it has been brisk. I don't know if the slowdown before the migration was because of the migration itself 17:22
dukeleto or just the time of year. I think that slowdown was seasonal
but this speed up in activity is related to us being on github and it being easy for people to watch, fork and submit patches 17:23
whiteknight I do sincerely hope that this is the case, and that we are going to see a permanent activity boost from it 17:24
17:33 Kovensky joined 17:34 particle left
cotto_work ~~ 17:39
whiteknight hello cotto_work 17:41
whiteknight tries building Rakudo again, this time with gcc 17:42
mikehh whiteknight: what configure params do you use for clang (or link) 17:48
whiteknight nothing special, just --cc, --cxx, and --link 17:58
okay, Rakudo built fine with the newest 2.10.1 tarball
18:04 tcurtis joined 18:16 Kovensky left 18:17 particle joined
mikehh we seriously need to incorporate 2.10.1 into master 18:22
Coke ah, yah, master is still reporting 2.10.0-dev, neh? 18:23
18:26 Kovensky joined 18:27 jan joined 18:35 rurban_ joined
dalek rrot: fe5b6cd | NotFound++ | t/ (23 files):
change set_addr to set_label in exeption handlers in test files, TT #218
 18:35
TT #1852 closed by doughera++: Configure.pl errors in 2.10.0 18:36
TT #1852: trac.parrot.org/parrot/ticket/1852
18:38 rurban left, rurban_ is now known as rurban
Coke mikehh: it's just a matter of merging prepare 2_10_0, no? 18:39
er, *prepare_2_10_0
cotto_work NotFound: did you warn rakudo about that change or check to see if it'd affect them? 18:44
NotFound cotto_work: given that we are using the deprecated form even in our repo, I don't think we should kill it right now. Sure, warning people will be good. 18:47
I've even noticed that I'm also using the deprecated for in winxed.
form
cotto_work It looks like they use set_addr in three places.
Ah. I thought you removed it instead of just updating some tests. 18:48
NotFound I'm just removing usages.
cotto_work ok. Thanks for doing that. 18:50
NotFound Is google code down? Some of its web pages are fine, but I can't aceess svn repo or project pages. 18:56
Coke NotFound: looks like the svn is nonresponsive, yes. 18:58
darbelo NotFound: Looks like it's down, yes.
mikehh Coke: dunno, does that have the 2.10.1 additions? 19:00
dalek TT #1850 closed by jkeenan++: config failure post 2.10.0 19:02
TT #1850: trac.parrot.org/parrot/ticket/1850
NotFound Looks iit's alive again. 19:10
PerlJam mikehh: looks like it does
git diff master...origin/prepare_2_10_0 # is how I looked 19:11
dalek nxed: r685 | NotFound++ | trunk/winxedst (2 files):
replace deprecated set_addr usage with set_label in exception handling
 19:12
NotFound cotto_work: you can tell rakudo guys to look at TT #1857 19:14
19:15 seanstickle left
PerlJam mikehh: oh, master already appears to have the same fix only slightly better 19:17
dalek nxed: r686 | NotFound++ | trunk/pir/winxed_ (2 files):
update installable files
TT #1857 created by NotFound++: Change remaining deprecated usages of set_addr to set_label
TT #1857: trac.parrot.org/parrot/ticket/1857
mikehh I mean like VERSION in master is 2.10.0, we need it to be 2.10.1 etc 19:24
19:32 AzureStone left 19:37 AzureStone joined
dukeleto what now? 19:42
moritz fix it. 19:45
19:45 muixirt joined
PerlJam Oh, I see what you're saying. git-describe is RELEASE_2_10_0-94-gfe5b6cd in master right now 19:47
and this probably results in what moritz was saying earlier about 2.10.1 reporting itself as 2.10.0 19:49
(maybe)
(I don't know for sure the parrot generates its version string from the release tag) 19:50
s/the/that/
Given that step 2 of the release manager's guide is 10 steps of "make sure you change the version here, here, here, and here" that odds aren't good that it's generated, but I hold out some bit of hope :) 19:53
dukeleto PerlJam: that stuff needs to be fixed. That can all be automated. 19:56
PerlJam: i think we may want to look at making a Dist::Zilla plugin for parrot release managers
PerlJam: we could probably automate 98% of the release manager guide
PerlJam dukeleto: +1 that would be awesome
cotto_work That's would be amazing. I'd love to be able to input some initial information, kick it off and come back an hour later to a fully tested and upload-ready tarball. 19:58
dukeleto cotto_work: yep. We need to make releasing parrot stupidly simple. 20:00
cotto_work: for instance, we need to automate running the test suite from a release tarball in a temp dir. That would have caught the bug that required 2.10.1 to be released 20:01
cotto_work yes
dukeleto and that temp dir must not be a subdir of the git repo, which was not specified in the release manager docs, because nobody knew that would matter.
tcurtis dukeleto: That wouldn't quite have caught the bug. IIRC mikehh also noticed that running the configure with --test caused the bug to not occur. 20:03
Coke dukeleto: it says 'into another area', which could be clearer. 20:04
(but covers the intent, IMO. and that's from 2007) 20:05
but automation++
dalek rrot: 7059420 | NotFound++ | src/ops/core.ops:
clarify the pod doc of constants for interpinfo argument, TT #104
 20:08
tcurtis VERSION appears to still contain 2.10.0. 20:10
github.com/parrot/parrot/blob/RELE..._1/VERSION
dukeleto blarg 20:11
dalek TT #104 closed by NotFound++: interpinfo.pasm constants don't match docs in core.ops 20:20
TT #104: trac.parrot.org/parrot/ticket/104
Coke NotFound is on a roll. 20:21
NotFound We have too much stalled tickets. 20:22
whiteknight NotFound++ # Thanks for looking at the set_addr tickets 20:27
set_addr sucks because anybody can cause a segfault by passing in a random value to it
getting rid of it and only using set_label to set valid label values will be a big help in preventing segfaults
NotFound It sucks in a lot of ways, but we must clean up that usages before anythieng. 20:29
whiteknight yes 20:31
NotFound BTW I think we should put set_label and get_label out of experimental.ops 20:34
whiteknight +1
are they well-tested?
NotFound A lot of tests use it. A lot more after my recent related commit. 20:36
inline op get_addr(out INT, invar PMC) { 20:37
void *ptr = VTABLE_get_pointer(interp, $2);
$1 = (INTVAL)ptr; 20:38
}
inline op get_label(out INT, invar PMC) {
void * const ptr = VTABLE_get_pointer(interp, $2);
$1 = PTR2INTVAL(ptr);
}
This different casting has some rationale?
whiteknight no. Did I write all that? I thought there was some bounds-checking in there for valid labels 20:40
NotFound BTW it makes any sense having invar LABEL parameters? There are non-const labels? 20:42
ack 'invar LABEL' src/ops/*.ops gives only one usage: set_addr(invar PMC, invar LABEL) 20:47
Andy just for grins, do it as 'invar\\s+LABEL' 20:48
I want to make a flag for ack that says "Turn any literal space in the regex into a \\s+"
NotFound Done, and also with dynops. Still only usage.
20:49 zby_home joined, bluescreen left 20:55 seanstickle joined, seanstickle left
Coke /x 20:56
cotto_work /y 21:00
darbelo /z 21:01
cotto_work ohai darbelo 21:02
atrodo /aa
darbelo ponders the use of a _work suffix
dukeleto /aleph_N 21:03
21:04 bluescreen joined
GeJ Good Matin everyone. 21:06
dalek tracwiki: v5 | plobsing++ | ParrotDeprecationsFor3.0 21:08
tracwiki: Add deprecation notice for .nci_call and .meth_call special forms
tracwiki: trac.parrot.org/parrot/wiki/ParrotD...ction=diff
rrot: 0994a6b | plobsing++ | / (7 files):
remove deprecated .nci_call PIR form

this is simply an outdated alias for .call		 21:12
rrot: 9bdc32d | plobsing++ | / (8 files):
eliminate deprecated .meth_call PIR form

this is simply an outdated alias for .call
rrot: 323a3c2 | plobsing++ | t/compilers/imcc/syn/objects.t:
fix tests depending on .meth_call

as recomended, these now use .call
rrot: 2eaf4b8 | plobsing++ | MANIFEST:
mk_manifest_and_skip
whiteknight Hello GeJ 21:16
dalek rrot: 94fcb44 | plobsing++ | DEPRECATED.pod:
retire completed deprecation.pod entries for .nci_call and .meth_call
21:18 davidfetter_vmw joined
whiteknight plobsing++ 21:20
several people are kicking ass today
dukeleto stands back and watches as the Git Times Roll 21:21
davidfetter_vmw puts on some old Cars tune 21:22
whiteknight it's not just git. We have some fantastic coders in our community 21:23
dalek tracwiki: v20 | plobsing++ | ParrotDeprecations 21:24
tracwiki: Add removal note for .nci_call and .meth_call special forms
whiteknight git is just the vehicle through which the awesomeness manifests
dalek tracwiki: trac.parrot.org/parrot/wiki/ParrotD...ction=diff
tracwiki: v21 | plobsing++ | ParrotDeprecations
tracwiki: correct typo
tracwiki: trac.parrot.org/parrot/wiki/ParrotD...ction=diff
TT #1623 closed by plobsing++: [DEPRECATED] .nci_call
TT #1623: trac.parrot.org/parrot/ticket/1623
TT #1624 closed by plobsing++: [DEPRECATED] .meth_call
TT #1624: trac.parrot.org/parrot/ticket/1624
dukeleto whiteknight: i can't hear you over all the activity in here today ;) 21:25
whiteknight yeah, it's more fun just to watch
davidfetter_vmw heh 21:28
whiteknight but now, I'm off. Later 21:30
21:30 whiteknight left
dukeleto davidfetter_vmw: how is the new $job? 21:35
21:35 perlite_ joined
dukeleto davidfetter_vmw: what can you tell us about that doesn't require an NDA? ;) 21:35
21:39 perlite left, perlite_ is now known as perlite
davidfetter_vmw dukeleto, we're doing Very Cool Stuff™ with PostgreSQL 21:43
and i'm not travelling quite as much, although that may change next year
21:46 zby_home left 21:49 donaldh joined 21:55 donaldh left 22:02 bluescreen left 22:16 bluescreen joined 22:23 fperrad left
bacek_at_work ~~ 22:29
plobsing++ # fixing gc_ms2_faster_list 22:30
GeJ aloha: clock? 22:32
aloha GeJ: GeJ: LAX: Thu, 14:32 PST / CHI: Thu, 16:32 CST / NYC: Thu, 17:32 EST / UTC: Thu, 22:32 UTC / LON: Thu, 22:32 GMT / BER: Thu, 23:32 CET / TOK: Fri, 07:32 JST / SYD: Fri, 09:32 EST
cotto_work :}~~ 22:33
22:34 benabik joined 22:39 benabik left
dalek rrot/origin/gc_ms2_faster_list: 0dade28 | chromatic++ | /:
Merge branches 'master', 'master' and 'master' of github.com:parrot/parrot
 22:53
rrot/origin/gc_ms2_faster_list: 5c682b1 | chromatic++ | src/gc/fixed_allocator.c:
[GC] Rearranged Parrot_gc_fixed_allocator_allocate.

This makes the most likely condition occur first.
rrot/origin/gc_ms2_faster_list: 34e9497 | chromatic++ | src/gc/fixed_allocator.c:
[GC] Simplified pool_allocate().
rrot/origin/gc_ms2_faster_list: 5a6044b | chromatic++ | src/gc/fixed_allocator.c:
[GC] Made get_free_*_item() funcs decrement counts.

Now those accessor functions track the counts of free items in the pool.
rrot/origin/gc_ms2_faster_list: 4dc014e | chromatic++ | src/gc/gc_ms2.c:
[GC] Removed unnecessary MS2 callbacks.
rrot/origin/gc_ms2_faster_list: 7247760 | chromatic++ | src/gc/gc_ms2.c:
[GC] Inlined Parrot_pmc_destroy() in MS2 sweeper.

It's crazy how much of an optimization this is.
22:55 chromatic joined
chromatic Okay, now that's not the branch I intended. 22:55
darbelo I particularly like the "Merge branches 'master', 'master' and 'master'" part :) 22:57
chromatic No kidding!
bacek_at_work chromatic, now we have TWO origins! :)
chromatic Apparently I have a local branch that tracks a remote branch but has no connection to it.
bacek_at_work chromatic, wanna me to merge it? Or you'll do it by yourself? 22:59
chromatic I'm trying to figure out how to check out that branch locally, then cherry pick my commits. 23:00
23:00 Khisanth left
bacek_at_work git checkout -b tmp origin/origin/gc_ms2_faster_list 23:01
chromatic One too many origins?
bacek_at_work yes
you need 2 of them :) 23:02
23:04 bluescreen left 23:06 bluescreen joined
chromatic Rebuilding to test the cherry picks now. 23:08
23:10 darbelo left, darbelo joined
chromatic Ah, I have to push to the appropriate branch. 23:12
23:14 tcurtis left 23:19 lucian joined
chromatic Let's see if this makes more sense. 23:26
dalek rrot/gc_ms2_faster_list: dfbc072 | chromatic++ | src/gc/fixed_allocator.c:
[GC] Rearranged Parrot_gc_fixed_allocator_allocate.

This makes the most likely condition occur first.
rrot/gc_ms2_faster_list: 98c21fc | chromatic++ | src/gc/fixed_allocator.c:
[GC] Simplified pool_allocate().
rrot/gc_ms2_faster_list: 43952b4 | chromatic++ | src/gc/fixed_allocator.c:
[GC] Made get_free_*_item() funcs decrement counts.

Now those accessor functions track the counts of free items in the pool.
rrot/gc_ms2_faster_list: f16f1ac | chromatic++ | src/gc/gc_ms2.c:
[GC] Removed unnecessary MS2 callbacks.
rrot/gc_ms2_faster_list: 14e8987 | chromatic++ | src/gc/gc_ms2.c:
[GC] Inlined Parrot_pmc_destroy() in MS2 sweeper.

It's crazy how much of an optimization this is.
dukeleto rides the commit wave 23:28
23:29 kid51 joined 23:30 jsut joined 23:31 darbelo left 23:32 Khisanth joined 23:35 jsut_ left
davidfetter_vmw wouldn't those technically be pushes? 23:35
davidfetter_vmw puts on some salt'n'peppa
dukeleto cranks the speed metal to average out davidfetter_vmw's music 23:38
davidfetter_vmw mmm...speed metalz :d
not a salt'n'peppa fan, dukeleto ?
23:38 hercynium left
cotto_work a balance must be maintained 23:39
For every action there is an equal and awesomer reaction.
dukeleto davidfetter_vmw: i like some of their stuff ;) 23:40
23:42 whiteknight joined
davidfetter_vmw o/` ooh, baby, baby o/` 23:42
23:47 dngor left, dngor joined
dalek rrot/gc_ms2_faster_list: 390038d | bacek++ | / (42 files):
Merge branch 'master' into gc_ms2_faster_list

Conflicts:
  \tMANIFEST		 23:49
rrot/gc_ms2_faster_list: 1de38be | bacek++ | src/gc/fixed_allocator. (2 files):
Rerun headerizer
dukeleto i really hate MANIFEST and MANIFEST.SKIP 23:52
they conflict on every merge, because they have timestamps that are always on the same line
whiteknight I've always hated MANIFEST 23:53
23:55 kid51_ joined 23:56 dngor_ joined 23:57 dngor left, muixirt left, kid51 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!