Parrot 1.8.0 Zygodactyly released | Latest modified TT's: icanhaz.com/parrotbugs | Parrot Languages: icanhaz.com/parrotlang
Set by moderator on 23 November 2009.
dukeleto japhb: just a pointer to other resources 00:00
japhb dukeleto, sure, but the length of the readme is not the question. It's the length of the build, test, and install output that's the problem.
On a fast system, a user wouldn't even know anything was printed out in the first place! 00:01
00:01 joeri joined
japhb Anyway, I'm trying to think of a good behavior that will work for "most" projects. 00:01
dukeleto japhb: nothing needs to be printed during an automated install. i just want a step that someone hand-compiling it can run 00:03
but basically it is the equivalent of "cat README"
japhb ah, I see. './plumage readme kea'
?
dukeleto japhb: now we are cooking with gas! 00:04
japhb: the step would just print the README if it exists
japhb ah, OK, I see
Sure, that makes sense.
dukeleto japhb: is that easy to implement? 00:05
japhb dukeleto, shouldn't be terribly hard. I'm most of the way through my current refactoring run, so features should be coming soon.
dukeleto japhb: sweet! 00:06
japhb And as a matter of fact, it's a good thing you said something now, because it somewhat alters the way I'm thinking about some internals code. :-)
dukeleto japhb: i will translate nqpTAP to setup.pir soon
japhb dukeleto, cool beans
dukeleto japhb: i talked to schwern and got some good hints about how to properly test a test harness 00:07
japhb: so nqpTAP will test itself. i am excited :) 00:08
japhb wheeee 00:09
that's good news
00:09 lucian joined
dukeleto japhb: basically abstract the code that parses the output from the code that prints out the results of the parsing, and then test the parsing results with mocked TAP data 00:10
japhb nodnod
dukeleto japhb: Question: Should i use setup.pir for PL/Parrot? 00:11
japhb What is PL/Parrot again? 00:12
dukeleto japhb: it has to detect if postgres is installed, with a compatible version, and it requires the postgres source
plparrot?
purl somebody said plparrot was the postgres+parrot integration project or github.com/leto/plparrot
dukeleto japhb: embedding Parrot in Postgres as a Postgres Language (PL), so you can write stored procedures in PIR/NQP
japhb Hmmm, good question. What does Postgres use for its build system? 00:13
I mean, by default, I'd probably say "Yes, switch to setup.pir". But if Postgres has some freaky replacement for make, that might be an issue. 00:14
(Or if you needed to be able to do stuff that setup.pir did not support.)
dukeleto japhb: ./configure-style autoconf+make
japhb dukeleto, I guess it would depend on the details. Give it a try and see if it works? 00:15
dukeleto japhb: basically, i need to do what rakudo does. checkout postgres git tree (or tarball) and then put a directory in a certain place, then run make in there, then run a test suite 00:16
PL languages needs to be compiled in a subdirectory of the postgres source 00:17
there may be a way around that, but that is how i do it now
that might be outside the realm of what plumage can do right now 00:18
japhb: postgres specifically requires GNU make 00:19
japhb: how do i say that in the metadata ?
afk, msg me 00:20
00:28 bubaflub joined
dalek rrot-plumage: 4ba740d | japhb++ | :
[CORE] Util.nqp: Add reduce() function; improve spacing between code and...
 01:08
rrot-plumage: bf67382 | japhb++ | :
[plumage] Use map & reduce to improve format of projects command
a: 84c5350 | (Geoffrey Broadwell)++ | plumage/kea.json:
[plumage] fixes for Plumage metadata
 01:13
rrot-plumage: 45c1838 | japhb++ | :
[METADATA] Add metadata for Kea; dukeleto++
japhb msg dukeleto I've got to run now, remind me about the GNU make for Postgres issue next time we're both around at the same time. 01:24
purl Message for dukeleto stored.
01:30 zak_ joined 01:56 cconstantine joined 02:19 TiMBuS joined 02:30 bubaflub joined
dalek a: 952f5e2 | dukeleto++ | plumage/kea.json:
Fix copyright_holder in metadata
 02:31
a: faa14b8 | dukeleto++ | :
Merge branch 'master' of github.com:leto/kea
rrot-plumage: 5e6ce6c | leto++ | :
[metadata] Update the copyright_holder for Kea
 02:36
02:52 joeri joined
dalek TT #1338 closed by coke++: 3 C functions for tcl/tk library 02:52
rrot: r42818 | coke++ | trunk/config/gen/call_list/misc.in:
add some call signatures for vkon++ (TT#1338)
 02:53
Coke japhb: ping 03:02
japhb Coke, pong 03:03
Coke anything worth saving from your POV at:
www.perlfoundation.org/parrot/index...ping_notes
(tene, pmichaud: you too.)
dalek tpfwiki: will@coleda.com | Parrot 03:04
tpfwiki: www.perlfoundation.org/parrot/index.cgi?parrot
Coke trac-- # SLOW SLOW 03:08
wtf.
dalek tracwiki: v9 | coke++ | CageTasklist 03:09
tracwiki: trac.parrot.org/parrot/wiki/CageTa...ction=diff
tracwiki: v1 | coke++ | splint 03:12
tracwiki: migrate from <span class="icon">�</span>www.perlfoundation.org/parrot/index...</a>
tracwiki: trac.parrot.org/parrot/wiki/splint...ction=diff
japhb Coke: I've reread the inter_hll_mapping_notes page. What did you mean by "anything worth saving"? You mean, when transferring over to the Trac wiki? (If so, I'd say "Just copy the whole thing -- there's a lot of historical info there".) 03:13
Coke ok.
dalek tpfwiki: will@coleda.com | Parrot 03:14
tpfwiki: www.perlfoundation.org/parrot/index.cgi?parrot
japhb Coke, there is the problem of the RT ticket reference ... but I assume you can just change that to the proper TT
dalek tracwiki: v7 | coke++ | Yapc10Bof 03:16
tracwiki: trac.parrot.org/parrot/wiki/Yapc10...ction=diff
tracwiki: v1 | coke++ | Dtrace
tracwiki: Migrate from <span class="icon">�</span>www.perlfoundation.org/parrot/index...</a>
tracwiki: trac.parrot.org/parrot/wiki/Dtrace...ction=diff
03:18 JimmyZ joined
dalek tracwiki: v125 | coke++ | WikiStart 03:19
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
tracwiki: v1 | coke++ | HLLInteroperability
tracwiki: migrate from <span class="icon">�</span>www.perlfoundation.org/parrot/index...</a>
tracwiki: trac.parrot.org/parrot/wiki/HLLInt...ction=diff
Coke japhb: I'm just moving the wiki items. =-)
japhb Coke, fair enough. And ... 03:20
Coke++ # for that
dalek tpfwiki: will@coleda.com | Parrot
tpfwiki: www.perlfoundation.org/parrot/index.cgi?parrot
tpfwiki: will@coleda.com | Parrot 03:23
tpfwiki: www.perlfoundation.org/parrot/index.cgi?parrot
tracwiki: v126 | coke++ | WikiStart 03:26
tracwiki: old wiki nearly gone.
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
tpfwiki: will@coleda.com | Parrot
tpfwiki: www.perlfoundation.org/parrot/index.cgi?parrot
tracwiki: v127 | coke++ | WikiStart 03:29
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
tracwiki: v1 | coke++ | CFFI
tracwiki: copy from <span class="icon">�</span>www.perlfoundation.org/parrot/index...</a>
tracwiki: trac.parrot.org/parrot/wiki/CFFI?v...ction=diff
tpfwiki: will@coleda.com | Parrot 03:30
tpfwiki: www.perlfoundation.org/parrot/index.cgi?parrot
rrot: r42819 | coke++ | trunk/config/gen/call_list/misc.in:
add another prototype for vkon/TT#1338
 03:34
03:38 davidfetter joined
dalek tracwiki: v128 | coke++ | WikiStart 03:40
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
tracwiki: v129 | coke++ | WikiStart
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
tracwiki: v130 | coke++ | WikiStart
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
Coke tpfwiki? 03:42
tpfwiki is shut down, as far as parrot is concerned. See trac.parrot.org/wiki
dalek tpfwiki: will@coleda.com | Parrot 03:43
tpfwiki: www.perlfoundation.org/parrot/index.cgi?parrot
JimmyZ Coke++ 03:44
03:54 bogen joined
Coke wonders why debug output added to partcl-nqp is not printing. 04:11
04:16 hercynium joined
Coke wonders if anyone else can duplicate his problem with t/sanity.t in partcl-nqp dying on the last test. 04:19
japhb I would, if partcl-nqp had Plumage metadata ... 04:24
japhb looks around innocently
Coke has no idea how to do that. 04:27
I also think it's probably not worth doing, if partcl-nqp is never anything more than a POC. 04:28
(once it's not, it's partcl)
japhb Original partcl has a partcl.json file; presumably just copy that and fix.
nodnod
fair enough
04:31 KatrinaTheLamia joined
Tene Coke: I went to the link you posted, and all I see is "Replace this text with your own". I don't know what you're trying to say to me. 04:49
Coke Tene: you missed the window; moved it over to trac.parrot.org 05:02
Tene Okay.
Is it different from this page: trac.parrot.org/parrot/wiki/HllInteroperability ? 05:03
ah, probably.
Oh, nice, this page also exists and is different: trac.parrot.org/parrot/wiki/HLLInteroperability 05:04
That's... LTA
Coke someone could merge those pages, sure. 05:05
on the plus side, I just killed a wiki. 05:06
Tene Oh, that's the page you just made 05:07
yay death!
Coke IZIT? whee.
hey, you know nqp-rx. maybe you can find my problem with partcl-nqp. =-) 05:08
Tene is it insufficient pie? 'cos I just solved that problem at home.
Coke (grammar.pm's 'eq' operator is generating: The opcode 'iseq_i_s' (iseq<2>) was not found. Check the type and number of the arguments )
Tene Go ahead. I'll look at it for a bit, but I make no guaranteed about my attention span.
Coke Tene: I wish!
Tene ew. that's stupid. 05:09
lemme verify, then find out why.
building...
Coke the last test in t/sanity.t is causing this. (the result of the string compare of [string compare] with 0
(granted, that should be a numeric compare.
(but either eq or ==, you get a very similar error.) 05:10
Tene oh, um, where's the partcl-nqp source? 05:11
Coke partcl-nqp?
purl well, partcl-nqp is github.com/partcl/partcl-nqp
Tene coke?
purl well, coke is Will Coleda <mailto:will@coleda.com> or perpetually annoyed. or magical ticket robot or (if Diet), something that turns into formaldehyde in my blood.
Coke I sure do.
Tene need to update parrot first... 05:13
that's at LEAST 5 attention span points gone. ;)
Coke yah, I needed to grab the try/CATCH/CONTROL stuff tene++ added!
Tene that jerk
nopaste "coke" at 72.228.52.192 pasted "tene: found this.." (4 lines) at nopaste.snit.ch/18883 05:26
Coke so, that's wierd. 05:27
Tene ack, sorry, phone call 05:29
Coke what? where's my free tech support? =-)
Tene you're not the one who called. =-) 05:30
Coke oh. 05:36
didn't pmichaud say something about constant zeroes getting dropped from the ast? 05:37
Tene oh, could be.
Coke that would explain why that iseq isn't comparing against a 0. =-)
Tene so put it in a PAST::Val
yeah, if you change it from 0 to 1, it doesn't crash.
So use Val
Wow, this problem was easy to solve.
Coke where do I put the Val? 05:38
Tene You should ask me for help more often.
Eh?
Coke you say, put it in a Past::Val... I can't put that Past::Val in the tcl code... where does it go?
Actions.pm // integer ?
Tene Coke: I haven't read any of pmtcl. 05:39
but whatever AST has a 0 in it
Coke I think you're assuming I know what I'm doing. =-) 05:40
nopaste "coke" at 72.228.52.192 pasted "something like this? (except this fails)" (13 lines) at nopaste.snit.ch/18884
"tene" at 24.10.252.130 pasted "look at what 'puts 0' compiles to" (10 lines) at nopaste.snit.ch/18885 05:41
Tene see the [0] => 0
?
Coke yes.
Tene yeah, something like what you posted. 05:42
Coke that fails with: 05:44
Class 'Regex;Match' not found
Tene it's also not what's used there.
Coke needed to add a :returns('Integer'), apparently. 05:46
(only found I needed that by googling for Past::Val.new =-) 05:47
fixed. tene++ 05:49
Tene :D
Coke well, worked-around, anyway. 05:50
now I can run t/sanity.t ... but am getting this on almost every other test file: 05:52
The opcode 'isne_i_s' (isne<2>) was not found. Check the type and number of the
arguments
purl somebody said arguments was what's left after action selection
Coke Oy.
(e.g. t/tcl_misc.t) 05:54
dalek rtcl-nqp: 7de9dfe | (Will Coleda)++ | (2 files):
update [catch] to use new try/catch/control
rtcl-nqp: 3dd2bc8 | (Will Coleda)++ | src/PmTcl/Actions.pm:
Work around a bug where == 0 was dropping the 0 in the generated PIR.
Coke unfortunately, I think nqp-rx itself is subject to this bug. 05:55
(or not. but bah.) 05:57
looks like constant strings are also getting dropped. 06:05
so [something] eq "something else" is dying with an iseq $I..., $S... that is missing the second string.
msg pmichaud I just committed a workaround to partcl-nqp that seems to resolve the issue I had with t/sanity.t; but this same problem of constants getting dropped is also happening in most other .t's, esp with string comparisions. [something] eq "something else" is dying with an iseq $I..., $S... that is missing the second string. 06:07
purl Message for pmichaud stored.
Coke gives up 06:10
davidfetter leto, ping 07:27
07:48 cotto joined
davidfetter 'sup cotto 07:48
cotto hi davidfetter
davidfetter what's new & good? 07:49
cotto leftovers
purl leftovers are easy to pick through
cotto if you'd asked about new and slow, I'd say my nqp port of pprof2cg
davidfetter heh 07:50
cotto After talking to pmichaud about the acrobatics that nqp has to do to be usable, I'm wondering if assignment semantics shouldn't be the next big overhaul for parrot. 07:51
As with most things Parrot, they're a mess. 07:52
davidfetter in the military, they call this a, "target-rich environment" ;)
cotto It's a good thing I'm not cutting the 2.0 release, because that'd be its name. 07:55
davidfetter heh
i suppose you could lobby whoever *is* cutting it... 07:57
cotto better to try to fix the bird 07:58
or at least compl^H^H^H^H^H^Hraise awareness 07:59
08:01 fperrad joined
davidfetter usually uses ^W for these kinds of things 08:01
08:04 cotto_w0rk joined 08:08 fperrad_ joined 08:25 lasko joined 08:27 lasko left 08:35 fperrad joined 08:39 eternaleye joined 08:47 iblechbot joined 09:59 jan joined 10:29 zak_ joined 10:40 kthakore joined 11:10 fperrad joined 12:31 lucian joined 12:51 payload joined 13:13 whiteknight joined 13:14 JimmyZ joined 13:24 lucian joined
whiteknight good morning #parrot 13:33
13:36 lucian joined
JimmyZ good morning, whiteknight. 13:52
whiteknight hello JimmyZ
JimmyZ howdy whiteknight ;) 13:53
JimmyZ had been waiting for whiteknight's new blog for a week. 13:55
whiteknight yeah, I've been quite busy this week and am actually wrting one now 13:56
JimmyZ whiteknight++ for new things. 13:57
whiteknight thanks!
14:11 joeri joined 14:30 payload left
pmichaud I think I removed the "constant zeros being dropped from the ast" -- although it's possible I missed some. 14:33
The new strategy is to drop undef nodes from the tree.
oooops 14:39
I did remove constant zeros... but apparently I forgot to push the changes.
so when other changes were made, my updates were lost. process fail (on this end) 14:42
Infinoid Cool, I have a little free time. What's the #1 thing dalek is lacking at the moment? 14:57
15:00 Psyche^ joined
dalek p-rx: 8468ad8 | pmichaud++ | src/ (3 files):
Remove .peek_ast, .ast returns Undef if not set.
 15:01
p-rx: 32ed0f5 | pmichaud++ | src/HLL/Actions.pm:
Switch to use definedness instead of truthiness for adding nodes.
p-rx: 4c8bc06 | pmichaud++ | :
Merge branch 'master' of git@github.com:perl6/nqp-rx
p-rx: fa71e47 | pmichaud++ | src/stage0/ (3 files):
Cleanup stage0 files from conflicted merge.
p-rx: b24fb5f | pmichaud++ | :
Merge branch 'master' of git@github.com:perl6/nqp-rx
p-rx: cd02597 | pmichaud++ | src/stage0/ (4 files):
Update bootstrap.
15:06 patspam joined
dalek rrot-linear-algebra: 0ffd2c3 | Andrew++ | README:
add info about building/installing on fedora
 15:18
15:19 joeri joined
dalek trixy: 09092ae | Andrew++ | t/syntax/system_bang.t:
fix the system_bang.t test file to use proper numbering
 15:28
trixy: 81b6bbe | Whiteknight++ | t/functions/abs.t:
fix t/functions/abs.t test to properly mark TODO a test that doesn't work
 15:34
15:38 cognominal joined
whiteknight always forgets git config on a new system 15:38
dukeleto whiteknight: that is why i have my .gitconfig in git, so on a new system i just git clone that repo and then symlink ~/.gitconfig and friends into it 15:39
whiteknight: i have every config file imaginable in git and I symlink into it
whiteknight well look at you, mister smarty-pants
dukeleto whiteknight: i even have a script to generate the symlinks!
whiteknight: so you say that vcgc should be real-time capable? 15:40
whiteknight my personal laptop crashed quite badly last week. I'm currently using VirtualBox on my work computer to test out alternatives
dukeleto whiteknight: i really want to hack on a real-time gc for parrot, but starting off seems like a big jump. do you have any tips? how do start implementing and test a new GC? 15:41
dalek rrot: r42820 | pmichaud++ | failed to fetch changeset:
[nqp]: Update with latest bootstraps (reincorporating ast/undef changes)
whiteknight dukeleto: I'm actually starting that very project myself 15:42
so soon I want to put together a branch and start hacking on it publicly
pmichaud message Coke see if latest partcl (with PARROT_REVISION bump) fixes the problems with pct. Looks like it was a git push brain-o on my end. 15:43
purl Message for coke stored.
dalek rtcl-nqp: 3eddba2 | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION to get nqp-rx fixes.
 15:44
whiteknight dukeleto: I'm starting in two phases: First, I'm going to write a simple GC that uses a series of linked lists to represent the different "colors" of objects. It's basically going to be similar to our current GC with slightly better sweep performance (I hope) 15:45
second stage is to copy that first core and add threading support to turn it into vcgc 15:46
I have an idea for a third stage too: I think we can modify the VCGC algorithm to be more aggressive about using generations for increased mark performance 15:47
15:54 patspam joined
dukeleto whiteknight: where do i learn about vcgc? 16:00
whiteknight: i am on board to help you, if it is real-time capabale, then parrot on RTEMS needs to use it
whiteknight let me look up the paper 16:01
dukeleto whiteknight: do you want to work on a git branch? or are you staying in svn-land?
whiteknight whichever, I'm slowly becoming proficient with git
dukeleto whiteknight: you can branch from the 'upstream' branch on my parrot github repo
whiteknight okay, I might try that 16:02
dukeleto whiteknight: that directly corresponds to 'trunk' on parrot svn
whiteknight: the master branch has a bunch of merge commits, but is the same content
whiteknight: but mostly you want my 'upstream' branch
whiteknight what is upstream? is that the svn repo? 16:03
or, the closest mirror of it?
dukeleto git remote add parrot git@github.com:leto/parrot.git
git checkout -b upstream parrot/upstream
whiteknight portal.acm.org/citation.cfm?id=301589.286878
16:04 Psyche^ joined
whiteknight that's the ACM page, doesn't have free full-text access 16:04
doc.cat-v.org/inferno/concurrent_gc...ent_gc.pdf
dukeleto whiteknight: git clone git@github.com:leto/parrot.git
whiteknight: then it will be called "origin" and you would do : git checkout -b upstream origin/upstream 16:05
whiteknight okay, I will try that a little later
dukeleto whiteknight: that is a repo that you can commit back to, create branches, etc
whiteknight: just don't commit to upstream ! :)
whiteknight so when I commit to that repo, it commits to Parrot SVN>
?
dukeleto whiteknight: no, it is uni-directional
purl okay, dukeleto. 16:06
dukeleto whiteknight: i just don't have any access restrictions to keep 'upstream' a pure mirror
whiteknight: nothing goes back to parrot svn
whiteknight: i use it to work on public git branches
whiteknight dukeleto: okay then. So I create a branch there, it's purely a git branch and has no real relationship to Parrot SVN? 16:07
dukeleto whiteknight: which can then be squashed down to 1 svn commit when they are reading to go in core. or whatever
whiteknight: correct.
whiteknight: i like that people can comment on github commits and have a nice web interface to review them
whiteknight I'll have to think about it. I might prefer to use an SVN branch just to keep the commit history in the Parrot repo 16:08
but that's a small nit
you are right though, github has a lot of benefits as well
dukeleto whiteknight: yes, up to you.
whiteknight: if it is going to be a long-lived branch, it might be painful to do it in svn 16:09
whiteknight I'll try to get something up this weekend with my current, modest, progress
true
if we have a git branch, would there be any kind of way we could iterate over the commits there and "replay" them to an SVN branch?
my biggest hangup is keeping small commit-sized diffs available for inspection in the Parrot repo 16:10
a huge whole-branch dump is hard to dig through
dukeleto whiteknight: what about: there is an svn branch that has stable stuff added it to it occasionally, but a public github branch is where all the gory details happen?
whiteknight: we can keep as much commit data as we want
whiteknight okay, I can be fine with that. Opportunity for me to learn a new workflow
dukeleto whiteknight: we can definitely replay every git commit as a svn commit. that is basically what git-svn does by default, without rebasing 16:11
whiteknight I'll try to get some things set up tonight when I get some free time
awesome
unfortunately my free time now is completely shot and only in sporadic bits and pieces
dukeleto whiteknight: the vcs you start with doesn't matter. just get something started :) 16:12
whiteknight AND my damn personal computer crashed too, so I have to waste time fixing that
on the bright side, I've learned how to use VirtualBox more effectively, so I should be able to set up virtual environments for a number of OSes to do more parrot testing 16:13
dukeleto whiteknight: that sounds like awesomesauce. a quick intro to virtualbox for parrot testing would be an awesome lightning talk or blog post 16:17
whiteknight: i converted Kea over to the PIR-based build system of fperrad yesterday, have you seen it? 16:18
whiteknight: you might want setup.pir for matrixy/PLA or some other stuff
16:26 cognominal joined
whiteknight I definitely want to update PLA and matrixy to use setup.pir 16:34
or, soemthing like it
I haven't seen your work on Kea yet, I've been in the hospital since wednesday 16:35
lots of backlogging to do
come home to like 300+ emails
16:36 cconstantine joined
whiteknight an intro to virtual box might make a nice blog post, yes. It really is quite amazing how quickly I've been able to get Parrot building and running on a few different systems already 16:37
mostly linux though, I couldn't get FreeBSD to install on it, and I haven't even downloaded opensolaris yet
dukeleto whiteknight: i would like to try pc-bsd on virtualbox, if that is possible 16:41
whiteknight: sorry to hear about your hospital visit. you have a lot of catching up to do!
parrot-users is starting to get lots of talk too, which is cool to see 16:42
16:44 patspam joined 16:45 theory joined 16:48 whiteknight joined
dukeleto whiteknight: kea mostly is in the prototype stage, now that it has a working build system, i can do the SMOP to make it actually work 16:55
whiteknight: i have some friends that are core factor devs, so they give me moral support :) 16:56
looks like setup.pir only works with Parrot 1.8.0 and above? 16:59
dalek a: 7200283 | dukeleto++ | README:
Update README with parrot link and minimum version requirement
 17:02
17:04 patspam joined
whiteknight I've only taken a cursory glance at Factor, I should look at it more 17:21
dukeleto whiteknight: "forth with batteries" is a good phrase 17:23
whiteknight: it has a gui IDE and good 3rd party library support
whiteknight: and pretty much lives up to being a "practical stack language" 17:24
whiteknight cool 17:28
I never did look at forth much either, which is weird since I tout myself as an "embedded guy"
dukeleto whiteknight: i never messed with forth much either. what i know about forth is from factor 17:29
whiteknight ok
Tene fperrad: if I give you commit privs on gil and steme can you just commit those changes yourself? 17:39
fperrad Tene, yes I can 17:42
17:42 Coke joined
Coke finally signs back on after the power outage. 17:42
whiteknight Tene: Gil?
purl Gil is, like, wearing eyeliner? or the currency in Final Fantasy or python's global interpreter lock which is horrible blip.tv/file/2232410
Tene fperrad: anything else here you want commit privs to: github.com/tene ? 17:43
whiteknight Ha! my first thought was that it was an FF reference
Tene whiteknight: a Generic Imperative Language that I wrote for a presentation a couple of years ago
whiteknight awesome
Tene fperrad: added to both. is the distutils.pir commit after I migrated steme to nqp-rx? 17:44
fperrad Tene, nothing else
Tene I need to make plumage metadata for perl6-sqlite...
17:45 zak_ joined
fperrad Tene, distutils handles nqp-rx 17:46
17:46 lucian joined
Coke pmichaud: I'm rebuilding now, thanks. 17:50
I admit, I can't see where you made your fix in perl6/nqp-rx. looks like you just updated the stage0 files. 17:51
(after I had already updated them after tene's stuff. So color me confused if this works. =-) 17:52
ah. github.com/perl6/nqp-rx/network clarifies things a bit.
pmichaud: that is much better, thank you. 17:56
pmichaud++ 17:57
dalek rtcl-nqp: aaf5b44 | (Will Coleda)++ | src/PmTcl/Actions.pm:
remove workaround; latest version of nqp-rx handles this as expected.
 18:03
18:08 fperrad joined 18:17 patspam joined 18:18 s1n left 18:20 zak_ joined 18:28 Zak joined 18:34 pdcawley joined
dalek ptap: 166d29f | dukeleto++ | (3 files):
Get rid of unused files and improve .gitignore
 18:41
Tene fperrad: I pushed your steme changes.
18:42 tetragon joined 18:48 hercynium joined 19:06 bogen joined 19:21 theory joined 19:23 NotFound joined
dukeleto what does "error:imcc:The opcode 'split_p_s_ic' (split<3>) was not found." actually mean again? 19:31
ic is .. ? 19:32
plobsing integer constant maybe?
Tene yes 19:33
dukeleto i think it is a string, but it thinks it is an ic. what does that mean? 19:35
Tene can you show me what's generating that?
why you think it's a string? 19:36
dukeleto Tene: it seems to have been a heisenbug. i am glad it is gone :) 19:39
Tene: oh wait!
Tene: there it is :)
Tene: actually, it is being cause by an "unexpected namespace" error earlier on 19:40
Tene: gist.github.com/15cd14223da25fe4f661 19:42
Tene: and github.com/leto/tapir/blob/master/l...Parser.pir 19:44
Tene: when I put all that code in one file without abstracting out into separate files, it works. that is "tapir.pir" in the main directory 19:45
Tene: also, you have a commit bit on tapir 19:46
plobsing dukeleto: I don't think you can nest .sub declarations in PIR 19:49
Coke is this valid nqp? my ($cond,$body) = @args; 19:51
you used to be able to next .sub's, but not for some time.
*nest 19:52
(I mean my ($cond,$body) := @args; 19:53
dukeleto plobsing: where is my nesting? 19:54
Coke: are you getting an error? 19:55
plobsing t/01-parse_plan.t .includes 'lib/Tapir/Parser.pir'
Coke dukeleto: yes.
plobsing I assume of course that .include does verbatim inclusion and not something fancier
Coke but relying on the runtime to determine if something is legal is susceptible to crappy coding. =-) 19:56
dukeleto Coke: nopaste?
purl nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at www.extpaste.com/ or paste.scsys.co.uk (for #catalyst, #dbix-class, #moose and others) or gist.github.com/ or paste or gtfo or tools/dev/nopaste.pl or trac.parrot.org/parrot/browser/tru...nopaste.pl
dukeleto plobsing: you are probably right
plobsing: i should be using a pbc
Coke Unable to parse blockoid, couldn't find final '}' at line 340
dukeleto plobsing: but i wonder how to do what I want without making pbcs
Coke: that means "don't do that"
plobsing .include outside of the main sub
would probably work 19:57
Coke dukeleto: see my previous comment.
dukeleto Coke: or a missing quote or something nqp cannot parse
Coke right. since it could be any number of things I screwed up, It would be nice to know if it was expected to work. =-)
dukeleto plobsing: you were right! i fixed stuff. thanks for your help 20:00
Coke: NQP does not have any syntax sugar
Coke: start from that premise, and you will not be disappointed :)
Coke: also, $ and @ don't actually mean anything in NQP, then are just mini-namespaces. you can put an array in a $foo and an array in a @bar. That will be the next thing that will confuse you ;) 20:01
s/then are/they are/
s/an array in a @bar/a scalar in a @bar/ 20:02
dukeleto at least succeeded in spreading confusion
tapir has a passing test! 20:03
dalek tracwiki: v112 | dukeleto++ | Languages 20:09
tracwiki: trac.parrot.org/parrot/wiki/Langua...ction=diff
Tene Huh, that actually worked... 20:11
dukeleto test_more.pir prefers to be imported into a :main function, because it exports things to the calling namespace 20:13
other things prefer to be outside a :main function 20:14
20:20 xenoterracide joined
dukeleto plobsing++ # tapir thanks you 20:28
20:30 iblechbot joined 20:36 zak_ joined 20:37 davidfetter joined
fperrad ping Tene 20:40
purl I can't find Tene in the DNS.
20:47 mj41_ joined
Coke ping www.spork.com 20:47
purl 10 packets transmitted, 10 received, 0% packet loss, time 9015ms, rtt min/avg/max/mdev = 5.348/5.556/5.839/0.168 ms
Coke gets a pull request on github from fperrad. 20:48
Coke wonders how that works.
dukeleto Coke: there is a web-based github UI for merging in forks branches 20:50
Coke: it just cherry-picks every commit, and tells you if it is likely to give conflicts
Coke: it is ok in a pinch, but i prefer command line tools. it is very shiny though 20:51
Tene fperrad: pong 20:52
fperrad Tene, in setup.pir (steme), could you fill the fields license_type & license_uri 20:53
Coke fperrad: is there an NQP version of the PIR you created? 20:54
(for partcl-nqp)
Tene fperrad: Yes, I will.
fperrad Coke, no, because NQP have not yet a hash constructor 20:56
Coke like pir::new__ps('Hash') ?
(which admittedly is a little wordy) 20:57
fperrad Coke, something like this : 20:58
setup(
:pbc_pir( {
'Math/Random/mt19937ar.pbc' => 'Math/Random/mt19937ar.pir',
} ),
:prove_exec( get_parrot() ),
:inst_lib( 'Math/Random/mt19937ar.pbc' ),
);
21:00 MoC joined
fperrad Tene, the next step is : 21:00
$ mkdir plumage
$ parrot setup.pir plumage > plumage/steme.json
Tene fperrad: done and pushed. thanks. 21:05
Coke incoming. 21:07
purl duck!
Coke msg pmichaud now have one of the original partcl tests passing... 21:08
purl Message for pmichaud stored.
dalek rtcl-nqp: faf635b | (Will Coleda)++ | CREDITS:
Add credits file from partcl, plus some new entries for this version.
rtcl-nqp: ebae878 | (Will Coleda)++ | src/PmTcl/Commands.pm:
implement [error] - make [catch] respect CONTROL_ERROR; add args-error to [while]
rtcl-nqp: b1b5426 | (Will Coleda)++ | src/PmTcl/Commands.pm:
first pass of [append]
rtcl-nqp: f133de3 | (Will Coleda)++ | src/PmTcl/Commands.pm:
[for] returns '';
rtcl-nqp: 3f3c9a4 | (Will Coleda)++ | src/PmTcl/Commands.pm:
add args error to [incr]
rtcl-nqp: 6ebafe5 | (Will Coleda)++ | src/PmTcl/Commands.pm:
add args-handling to [llength]
rtcl-nqp: 9aeb25c | (Will Coleda)++ | build/Makefile.in:
Add another test to make test!
fperrad ping japhb 21:12
purl I can't find japhb in the DNS.
21:13 tetragon joined
Coke another test file... 21:24
21:25 plobsing joined
dalek rtcl-nqp: e9f559a | (Will Coleda)++ | (3 files):
Recover t/cmd_split.t ...

and force tcllist to cover array type used by parrot's split.		 21:25
rtcl-nqp: 232402e | (Will Coleda)++ | (2 files):
reclaim t/cmd_join.t -
 21:36
21:39 patspam joined
Coke anyone know how to specify a key in npq's pir:: syntax? 21:41
dalek l: d5c4641 | fperrad++ | (7 files):
update infrastructure with setup.pir (distutils)
l: e002e93 | fperrad++ | (13 files):
refactor with opcode load_language
Coke (trying to delete a hash entry)
l: 8b2f130 | fperrad++ | plumage/gil.json:
add a Plumage description
21:44 tetragon joined 21:48 nopaste joined
Tene c'mon nopaste... 21:48
Bah, bot fail.
nopaste.snit.ch/18891
I got a basic sort of macros working in my scheme compiler. 21:49
Coke evil.
Tene Yeah.
I need to get that into a more reasonable form, but... it works. 21:50
fperrad Tene, Gil is pushed, in setup.pir, could you confirm (or modify) the fields license_type, license_uri & copyright_holder
Coke hurm. anyone know how to access a NQP variable from inside a Q:PIR block?
(tempting to just write this whole sub in pir.) 21:52
fperrad ping dukeleto 21:54
purl I can't find dukeleto in the DNS.
Tene Coke: $P0 = find_lex '$foo'
Coke: know how to get a class object in NQP? Foo::Bar.add_method() complains of Method 'add_method' not found for invocant of class '' 21:55
fperrad: that license is fine. Thanks.
Coke hurm. no.
er, no I don't know how. 21:56
tene++ that worked. 21:57
nopaste "tene" at 24.10.252.130 pasted "Coke: the action method for that macro syntax" (11 lines) at nopaste.snit.ch/18892 21:58
22:07 chromatic joined
dalek tracwiki: v113 | fperrad++ | Languages 22:09
tracwiki: trac.parrot.org/parrot/wiki/Langua...ction=diff
nopaste "tene" at 24.10.252.130 pasted "A more realistic-ish macros example" (14 lines) at nopaste.snit.ch/18893 22:16
Tene purl: msg pmichaud Check out github.com/tene/steme/commits to see the start of my macros impl. and a test demonstrating it. 22:18
purl Message for pmichaud stored.
dalek l: b6ad877 | fperrad++ | setup.pir:
chmod +x setup.pir
 22:22
l: 441068a | fperrad++ | .gitignore:
update .gitignore
22:24 theory joined 22:33 nopaste joined
dalek a: 2389ba7 | fperrad++ | setup.pir:
chmod +x setup.pir
 22:34
a: ebbb281 | fperrad++ | .gitignore:
update .gitignore
22:36 tetragon joined 22:42 Zak joined 22:44 lucian_ joined 23:16 nopaste joined 23:28 Whiteknight joined 23:35 theory joined 23:47 Whiteknight joined
Whiteknight finally has a workig computer again 23:48
23:52 davidfetter joined 23:56 redbrain joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!