geoffb's column: www.oreillynet.com/pub/au/2333 | Pugs 6.2.10 coming soon | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 5 October 2005.
autrijus :) 00:00
Shillo geoffb: Failing to mount a share would confuse Visual Studio. -That- sucked.
Shillo helpfully tries to build with PUGS_EMBED, too... 00:01
geoffb clkao, btw, if you read my O'Reilly blog from wed, you may have noticed that SVK is the only VC client I said was "excellent"
joao svk assumes that we have a svn remote server, or am I wrong?
clkao i know :P
joao: yes, it works best for that. but you can also branch from cvs or p4 00:02
Shillo I kinda like arch, except for a bit of a learning curve.
geoffb clkao, I hate SVK because I still have to have a darcs client . . . but I've said that before. :-)
joao Southen_, geoffb need to convince his boss to use svn, I guess :)
s/Southen_/clkao 00:03
geoffb joao, $work = $contract. Less sway there, since they can't bug me all the time for help for "free"
svnbot6 r7361 | autrijus++ | * install and register Pugs to cabal via "make install".
Shillo Still burps itself on missing parrot headers, then continues to build...
Built. I'm way too tired to check whether the parrot actually got embedded. :) 00:09
Night, all!
geoffb night, Shillo
autrijus night Shillo. thanks!
Shillo Actually PGE does run.
joao 5.8.1 is not supported? 00:11
autrijus joao: hm? I think it is supported, tho I'm not sure. 00:12
joao I just don't get the error, then :9 00:14
Generating precompiled Prelude... Can't exec "./pugs": No such file or directory at /System/Library/Perl/5.8.1/IPC/Open3.pm line 241.
oh well, i'll go to bed, good evening! 00:17
autrijus joao: the thing would be the error before the "Generating" line 00:18
good night :)
joao autrijus, I didn't have the latest svn release; I'm building again, let's see... 00:19
autrijus oh ok... wait me a sec 00:20
please up to r7362 00:21
joao hmmmm. ok 00:24
svnbot6 r7362 | autrijus++ | * trigger rebuild for dist too. 00:25
joao I definitely have to update my environment :)
autrijus "make unoptimised" is your friend :) 00:26
joao I'm using it
autrijus (and it can't help that I just doubled the make time. temporarily though)
joao But I was refering to update development tools, for instance
autrijus ah.
joao ghc still 6.4
geoffb As a side note, new config.yml keys aren't added or mentioned by Makefile.PL.
joao perl 5.8.1 ;)
geoffb I moved my config.yml aside and reran Perl Makefile.PL to get the new keys, but people shouldn't have to think of that 00:27
autrijus geoffb: indeed.
geoffb: slap a -M protection to Makefile.PL?
-M __FILE__ against -M 'config.yml'
in config.yml 00:28
or compare it against util/config-template.yml
in inc/PugsBuild/Config.pm
improvise :)
geoffb autrijus, nodnod
Just started a build. Going to see if that passed before mucking with anything
autrijus ok.
svnbot6 r7363 | autrijus++ | * try to be a bit more portable with GNU ar. 00:30
joao still building :-) 00:41
svnbot6 r7364 | autrijus++ | * switch to an autoconf-like format for Cabal building. 00:42
geoffb Perl 5 embedding causes build to explode
autrijus not surprisingly, but ack 00:43
svnbot6 r7365 | autrijus++ | * bump version to 6.2.10 as release is imminient. 00:48
geoffb autrijus, I'm doing the -M hack and a little cleanup in the surrounding function. Is there time to get it in before you cut? 00:49
autrijus sure 00:50
I need a sleep between this and 6.2.10.
;)
and I'm still wrangling with cabal. close, though 00:51
mostly just don't want to double the maketime.
joao autrijus, for how long are you awake? 00:52
autrijus joao: I don't know, close to 24 00:53
joao autrijus++ # pugs is built!
autrijus joao: oooh
joao autrijus, you deserve some sleep :)
autrijus not going to do that without a working cabal :)
joao Being so many time awake would drive me crazy. 00:55
autrijus hm, what makes you think I'm not crazy :) 00:57
joao Oh, I forgot the "too" :) It would drive me crazy, too. :)
geoffb OK, tested and committed, now trying build sans PUGS_EMBED 00:59
svnbot6 r7366 | autrijus++ | * depend on "Win32" and "unix" packages correctly according to platform.
r7366 | autrijus++ | * verison normalization.
r7367 | geoffb++ | PugsBuild::Config: some cleanup in sub read, and warn if config.yml is older than template
joao waves 01:01
offby1 waves back 01:02
autrijus waves 01:03
theorbtwo particles. 01:04
nothingmuch oof, /me needs parametrized types in perl 5 01:07
it's nice that perl 5 can let you do all this stuff without too much effort
but it would be nice to get shortcuts occasionally
good night everyone 01:10
geoffb gnight, nm 01:11
theorbtwo G'night, nm. 01:14
rafl mugwump: What about the Debian package? I maintain it. 6.2.9 is in sid. 01:18
geoffb: No matter. If you intend to do some work I can also add you as a co-maintainer with the next release. 01:19
mugwump rafl: I was just looking for something to play with my backporting system 01:20
geoffb rafl: Yeah, let's aim for the next release.
mugwump I have 11 x86 Debian distributions on my desktop with which to do this :)
rafl mugwump: A backport would be hard. You need ghc 6.4. 01:21
mugwump yes, I'd have to backport that, too
hmm, no parrot dependency?
geoffb mugwump, why do you have 11 x86 Debians on your desk? Oh, do you mean VMs? 01:22
rafl geoffb: As I get my @d.o account the next week I'll do some experimental releases using svn snapshots. So there will be more (debian) releases soon.
mugwump well, there's Debian, and Ubuntu, and three distributions of each of those, and then 64 vs 32 bit
rafl mugwump: No. No parrot in Debian yet. I'm working on that, though.
geoffb rafl, good, gives me a chance to get $work beaten back 01:23
rafl :-)
mugwump geoffb: they're not in VMs, just chroots with Linux-Vserver.org 01:25
rafl mugwump: btw, ask igloo. I remember he did a backport of ghc6 once. 01:26
He's a regular on #haskell.
geoffb mugwump, ah, interesting. So why did you do this? Variant distro compat a job, or just a hobby? 01:27
autrijus rafl: heads-up: we no longer install any of the src/*
rafl: does Debian package Cabal separately from GHC as gentoo does? 01:28
mugwump I just got sick of various packages I wanted not being in the particular flavour of debian I have to work with...
rafl autrijus: Yay! 01:30
geoffb By "have to work with", I'm assuming you do a lot of Debian $work or $contracts . . .
rafl autrijus: But you still install perl5/ as it is, do you?
autrijus: No, cabal is in the ghc6 package.
mugwump geoffb: the OS development company I work for are a debian shop (with an openbsd garnish) 01:31
geoffb cool 01:32
what company?
mugwump catalyst.net.nz
rafl autrijus: I really hate the fact that the Makefiles.PLs and t/ests are installed for each perl5/ module. 01:33
autrijus: Especially for those who a) aren't perl5 modules and b) aren't useful at all when being installed (org.perl6.metamodel). 01:34
autrijus rafl: I really hate them too.
rafl: better suggestions welcome.
rafl autrijus: Put it into lib/?
autrijus rafl: can't; they are not really perl5 libs as such
rafl autrijus: And the tests into t/.
autrijus: I think they are. MM2 for example could also be used without pugs. Others as well. 01:35
autrijus hrm.
rafl autrijus: Alternatively you could put them into /usr/share/pugs/perl5.
autrijus rafl: currently they are installed under pugs sitelib
rafl (the built version of course)
autrijus rafl: I guess we can reclassify them so only the ones really used during -BJS and -BPugs 01:36
geoffb mugwump, looks decent, from the weblies at least
autrijus gets installed
rafl Yes, please.
But not under /usr/share/perl6/5.8/perl5. Think of 'use perl5::Perl6::MetaModel, etc.
autrijus I agree. 01:37
but chaining to other Makefile.PL for perl5 libs did not work so well last time 01:38
and to release them all separately to CPAN is... probably not for this release, as discussed with Stevan.
so the lightweight approach would be make out a manifest of what to copy.
rafl Why do their makefiles need to be changed?
autrijus or what to left out, etc. it's in util/src_to_blib.pl
rafl: many of them does not have makefiles. 01:39
rafl Pleas *DO* *NOT* *DO* the same crap as parrot does. It has a MANIFEST which says what needs to be installed in what place.
QtPlatypus rafl: What is wrong with that? 01:40
autrijus QtPlatypus: it makes adding new stuff a bit annoying. 01:41
yay I fixed the symbol table bug.
rafl What about let the "main makefile" go though each perl5/* directory, do perl Makefile.PL; make; and make install DESTDIR="pugsblib or something else that'll be installed into $PREFIX/share/pugs/perl5/" 01:42
autrijus proceeds to cut build time in half
rafl QtPlatypus: Try to make a Debian package out of it and you'll see.
autrijus rafl: sure, go ahead with that. I'd just like to point out that the normal "subdirs" approach does not work
so you'd need to make up a rule
the reason for that is MakeMaker, in its infinite wisdom
put the subdir blib into the parent blib.
geoffb nice 01:44
rafl autrijus: Yes. Unfortunately I don't have that much free time to code on pugs. If I find some tomorrow, which I should, where to start hacking? I don't know exactly what the current makefile does to get everything built.
geoffb autrijus, "symbol table bug" whose fix "cut build time in half"?
autrijus rafl: the symtable bug was preventing us from linking cabal into main pugs 01:45
er
geoffb: the symtable bug was preventing us from linking cabal into main pugs
so I had to build each once
now it should be much simpler
rafl: nod... as I said, util/src_to_blib.pl is probably the place to hack.
geoffb ah, cool
autrijus "You may wish to check for new settings. 01:46
"
hm, not really helpful :)
autrijus wonders if it should silently merge new keys.
{ %$new_template, %$old_value }
probably yes. what you think?
rafl autrijus: It relies on MM to do the installation. I don't know if MM can handle it to install files somewhere outside of {SITE,PRIV,VENDOR,WHATEVER}LIB. 01:47
geoffb It was intended to be a quick hack . . . I only could spare enough time to hack and test what I did. :-(
And yeah, that would be a nice thing.
autrijus rafl: it could; see install target in MAkefile.PL
geoffb But you should still tell people that new keys appeared.
rafl autrijus: Looks cool. I think I can hack something together tomorrow. 01:48
geoffb Otherwise a new key may come with a default they didn't like.
svnbot6 r7368 | autrijus++ | * Unicode.C no longer linked with GHC 6.4.1. 01:49
r7368 | autrijus++ | * Fix the \\n tokenization bug (probably ghc bug) that
r7368 | autrijus++ | renders primitive disappear when linked with cabal.
autrijus geoffb: no worries, I fixed 01:50
stevan hey autrijus :)
autrijus yo stevan.
rafl: cool!
geoffb geoffb-- # net 0 added value, sigh
stevan I managed to convince robkinyon that eigenclasses with implicit inheritance was the way to go 01:51
rafl btw, I talked to someone today who wants to write a PIL-Run for Ruby.
stevan however, he made some good points about not allowing class method inheritance
the idea that since class data itself is not inheritable,.. then you end up with methods without the corresponding state 01:52
autrijus rafl: ooh, ruby _backend_?
rafl: who?
geoffb: nope, that was very useful.
rafl autrijus: Timo Schneider. A fellow student. 01:53
autrijus excellent :) get him on irc?
stevan give that boy a committer bit!
geoffb & # Mmmm, dinner calls . . . .
rafl I'll push him to get an account on openfoundry tomorrow. 01:54
autrijus rafl: you can send an invitation to him :)
rafl autrijus: How?
autrijus (you are now metacommitter.)
rafl: log in to openfoundry first
then
rt.openfoundry.org/Foundry/Project/...?Queue=270
fill in his email on "Invite External Staff"
and voila.
svnbot6 r7369 | autrijus++ | * Merge old settings with new ones when a new config.yml template 01:55
r7369 | autrijus++ | occurs, but still inform the user. geoffb++
autrijus geoffb: you too :)
rafl autrijus: I'm not an admin and can't visit that page therefor. 01:56
autrijus rafl: have you logged in?
geoffb autrijus, thx. :-)
& # No, really
rafl I am an admin.. how comes? 01:57
Does the invite also work for users without an openfoundry account? 01:58
autrijus rafl: yes, that's the idea.
you fill in email, they get spammed, and go thru one-click registration.
instant gratification and all that.
I think we should lift all channel regulars into metacommitters 01:59
but I'm a bit too lazy to do that now. helps welcome :)
rafl openfoundry++
autrijus (you can promote one to admin by click on their name on that page and choose "Admin" in the "Found people" panel.
okay, pugs successfully and entirely cabalized.whew. 02:00
svnbot6 r7370 | autrijus++ | * ...successfully cut build time in half _and_ did away all the
r7370 | autrijus++ | random .o files in src/.
autrijus still not tested on win32
and parrot/perl embedding needs work
but it's 5am and I need to sleep.
rafl tries that for 'Juerd' because he also made me an 'Admin' on feather ;-)
autrijus rafl: oh, you are feature sudo?
rafl sudo grep rafl /etc/sudoers 02:01
rafl ALL=(ALL) ALL
;-)
autrijus rafl: can you "sudo passwd autrijus" and /msg me whatever password you set for me? I opted to go passwordless and rely on pubkey, but then Juerd made me sudo, and I don't have a passwd anymore ;)
eos hi all 02:02
autrijus (and the sudo as not NOPASSWD.)
greetings eos.
eos anyone ever work with cgi scripts that load in local video files per web requests?
I know that's general.. I do have a very specific question, and it's just blowing my mind.
rafl autrijus: OK, I'll do that.
autrijus eos: that... is probably for #perl 02:03
eos oops, didn't notice the 6
good luck with parrot!
=)
autrijus thanks. :) 02:06
svnbot6 r7371 | autrijus++ | * no need to via-C for Prim.hs anymore.
rafl autrijus: What does descend into perl5/ and ext/ to build all the modules currently?
autrijus rafl: perl5 isn't built
it's merely copied by srctoblib
rafl: ext is governed by regular MM "build_subdirs"
rafl autrijus: Ah, fine. 02:07
autrijus: You need to do a make register now to install pugs, do you?
autrijus rafl: no... "make install" does it for you
at least I hope so. 02:08
rafl autrijus: Please don't do that. That's contra-productive for packaging people. At least make me able to disable it.
autrijus: I need to install pugs into a temporary directory, build an archive from that and run the register stuff in a maintainer script when it really gets installed into the real system.
autrijus rafl: you'd rather me have a separate "register" target? 02:09
rafl autrijus: Yes. And to have a *BIG* *FAT* *HINT* that's displayed after perl Makefile.PL and make. :-D
autrijus: Why isn't perl5/ handled using build_subdirs? 02:10
autrijus rafl: because most of them lacks a Makefile.PL, and some of them does not really make sense to install into perl5 sitelib, and even more of them relies on a particular dircetory layout between them 02:11
rafl: which should of course all be fixed.
rafl OK. My approach would be to 1st) put the stuff needed for -CJS, -CPerl5, etc. into another directory. or move the unneeded stuff to somewhere else (suggestions on where what should go?) 2nd) Write proper makefile.PLs for the needed stuff 3rd) hook into MM to make it descend into that dirs and put it into pugsblib/ or perl5/$module/pugsblib which will be installed into $PREFIX/share/pugs/perl5. 02:14
Opinions?
autrijus rafl: unneeded stuff into somewhere else, definitely 02:16
I wouldn't mind misc/
and move modules/* over to misc/
otherwise it makes much sane.
s/sane/sense/
rafl What differs modules/Perl-MataModel from perl5/Perl6-MetaModel*? 02:17
autrijus rafl: modules/P-MM is a prototype for perl6-metamodel-with-perl6 02:19
which is at this moment not worked on
and perl5/P6-MM is perl6-metamodel-with-perl5
rafl: "make register" is now separated. thanks for the hint
svnbot6 r7372 | autrijus++ | * "make install" no longer implies "make register", thanks 02:21
r7372 | autrijus++ | to rafl's prompting.
r7373 | rafl++ | * Created misc/ as a place where modules that should not be installed live.
r7373 | rafl++ | * Moved modules/*, perl5/org.perl6.metamodel and perl5/Blondie into misc/.
rafl Ok, what parts of perl5/ are actually needed? 02:23
autrijus rafl: T2 and org.perl6.metamodel is not
Blondie is not
I think you can rm Code-Perl directly 02:24
rafl Only Perl6.MetaModel and Perl6-MetaModel I guess. Maybe Perl6-Value and Perl6-Container.
autrijus rafl: everything else stays.
er, Perl6-MetaModel 1.0 is also gone 02:25
as in no longer referenced
but talk to stevan before deleting it; miscify it should be fine
miscify it into Perl6-MetaModel-1.0 and rename 2.0 into Perl6-MetaModel perhaps.
if so, fix the ref in PIL-Run. 02:26
PIL2JS wants Perl6.MetaModel/.
eof.
rafl I think the MM/PIL-Run guys should rename 2.0. I don't even know how to run the test suite correctly. :-) 02:29
svnbot6 r7374 | rafl++ | * Moved T2-Perl6 and MM1.0 to misc/ (renamed the latter one). 02:30
r7374 | rafl++ | * Deleted Code-Perl.
autrijus rafl: I think you should just go ahead and rename for them.
autrijus praises anarchism
svnbot6 r7375 | rafl++ | * Moved perl5/Perl6-MetaModel2.0 to perl5/Perl6-MetaModel. 02:32
r7375 | rafl++ | * Updated references to Perl6-MetaModel2.0 in PIL-Run.
r7375 | rafl++ | * Updated PIL-Run's README.
r7375 | rafl++ | * Praised anarchism.
autrijus lol
rafl -Ofun. :-) 02:33
OK, 1st step done.
Why doesn't MM2 use proper namespaces? 02:36
autrijus because stevan did not have proper tuits to rename them into Perl6::MetaModel::Pneuma etc. 02:37
he agreed to it, just had not got a round tuit.
(trivia: he gave the round tuit to nothingmuch at toronto.)
svnbot6 r7376 | autrijus++ | * dons noticed that it was the CPP extension that was
r7376 | autrijus++ | wrecking all havoc with Prim's symbol table. Turning it
r7376 | autrijus++ | off solved all the problems. dons++.
autrijus rafl: so again, feel free to go ahead :) 02:39
rafl autrijus: Hm. How about you doing that while I'll write some of perl/*/Makefile.PL? 02:40
:-)
autrijus I'm journaling... but why not ;)
rafl Yeah, then you can journal about it. :-) 02:41
autrijus ;)
rafl Hm, inc::Module::Install/MM doesn't allow '::' in a distributions name, does it? 02:43
svnbot6 r7377 | autrijus++ | * move MetaModel into lib/Perl5/MetaModel/
autrijus ...done. 02:45
rafl: it is distname, so use dash
but I remember I added compat
rafl autrijus: It gives me an error when doing make if there's a '::'. 02:46
A dash works, of course.
autrijus k
svnbot6 r7378 | autrijus++ | * finish the Odyssey. 02:49
r7379 | rafl++ | P6-Value:
r7379 | rafl++ | * New Makefile.PL using inc::Module::Install.
r7379 | rafl++ | * Updated MANIFEST.
r7380 | rafl++ | * Nice identing vor P6-Value's Makefile.PL.
r7381 | rafl++ | * Fix typo in fglock's mail address in P6-Values Makefile.PL. 02:52
r7381 | rafl++ | * New Makefile.PL for Perl6-Container using inc::Module::Install.
r7381 | rafl++ | * Updated MANIFEST for P6-Container.
rafl OK, what's left in perl5/ gets more difficult. They have lib, lib6 and PIL2JS even has libjs that needs to be installed somewhere.. 02:54
autrijus hence, my suggestion of keeping the structure for now.
but that'd not be the Right Thing. :) 02:55
rafl Well, the makefile would look similar to the pugs one i guess.
How should the Makefile var containing the directory for private, arch-indep files be called? 03:03
PUGSPREFIX? 03:05
No, that wouldn't be good.
PUGSSITELIB maybe?
eric256_ hey....just saying hi so you don't feel like you are rambling to yourself ;) no real input though. What does perl5 use? that might be a good way to go sense it will automaticaly make sense to more people 03:07
rafl eric256_: Hi. We talked about perl5/ before you joined. It's used for stuff that needs to be installed to make the backends usable. 03:09
autrijus rafl: PERL6_SITELIB or PUGS_SITELIB
eric256_ what about PUGS_PRIVLIB or PUGS_ARCHLIB (/me has the feeling he should probably go find the logs before saying much more) 03:10
rafl autrijus: OK. PUGSSITELIB. I'll drop the underscore for consistency. 03:11
Hm. Any way to access the pugs prefix in util/PugsConfig.pm? 03:14
Hm. I don't want that anyway. Forget it. :-) 03:17
eric256_ PugsBuild::Config->lookup("prefix"); in case you change your mind 03:19
rafl Thanks. 03:21
eric256_ juerd.ping 03:25
rafl Hm.. get_pugs_config->{sourcedir} gives me the directory of the current module, not the pugs sourcedir. 03:27
What's going wrong here?
(I want to execute a util/ script during the build of a module.
Hrm.. I'm too tired to continue working on that. But luqui can because he just showed up! 03:40
eric256_ would it make sense to let people set things in config.yml that are currently set as env variables? 03:41
luqui cannot work
work on what, btw?
rafl, I'm starting up the autoupdating evalbot again
eric256_ config.yml makes sense to me, but ENV variables dont. lol
luqui so you can kill yours if you like
eric256_, I think that's a good idea 03:42
rafl luqui: On writing Makefile.PLs for the perl5/ stuff. 03:43
eric256_: Yes, but the env hooks should still be available with a higher priority then the config.yml.
luqui hahaha
luqui doesn't know crap about that stuff
rafl luqui: OK, I'll do that.
luqui I just take the default whenever I make a module
rafl luqui: Me as well. That's the problem I guess. ;-) 03:44
luqui: OK, killing evalbot now.
luqui: What was the problem some time ago?
luqui with evalbot?
eric256_ i wonder if PugsBuild::Config could be setup to check ENV then config.yml...and replace all direct references to $ENV{} ...downside to that? 03:45
rafl luqui: Yes.
luqui feather: 05:45:07 up 6 days, 17:04, 9 users, load average: 1.35, 0.81, 0.51
feather rebooted
rafl Ah, OK.
Was that only 6 days ago?
luqui looks like it 03:46
rafl autrijus: What does ./setup install actually do?
autrijus: Does it honor PREFIX and stuff that was set in the Makefile?
autrijus: It should!
eric256_ night all. 03:50
luqui ?eval "this is only a test" 03:54
evalbot_7381 'this is only a test'
luqui okay, gtg before I get distracted by you guys :-) 03:55
dduncan fyi, I've just done an upgrade to ghc 6.4.1 and a pugs make-realclean ... now rebuilding Pugs under the new version ... currently in the middle of 'make' and no issues so far 04:47
there was a failure at the end of 'make', however ... here's part of the output: 05:19
src/Main.hs:18:0:
Failed to load interface for `Pugs':
Could not find module `Pugs': use -v to see a list of the files searched for
Triggering rebuild... done.
Generating precompiled Prelude... Can't exec "./pugs": No such file or directory at /Volumes/Programming160/Perl/lib/perl5/5.8.7/IPC/Open3.pm line 244.
brentdax XML is the new COBOL. 07:12
QtPlatypus brentdax: Thats stament seems wrong on a number of levals. So please explain yourself. 07:14
brentdax It's incredibly painful and suits love it. 07:15
QtPlatypus would say that Java is the new COBOL (though java isn't painfull enought) 07:16
QtPlatypus doesn't think that XML is that painfull. "Infact its not so bad realy, there are infinitly worse ways to serizile infomation" 07:17
brentdax I mean, it's fine for expressing some data in (although it's a bit heavy). But when you try to use it as a meta-language--schemas, XSLT, etc.--it totally breaks down.
QtPlatypus brentdax: I totally aggry with you there. 07:18
its been pushed outside its neache but a rather evangilical croud that wishes to make everthing XML. 07:19
brentdax It makes me want to implement a YAML parser for $gui_project.
QtPlatypus chuckels.
geoffb brentdax: <peer_pressure>do it, do it, Do It, Do It, DO IT!</peer_pressure> 07:33
brentdax It's not a Pugs thing... 07:34
geoffb brentdax, who cares? It's still the right thing to do. :-)
With the metacommitter bit comes a responsibility to push people to -Ofun . . . 07:35
Dangit, my spidey sense is tingling, and I don't know why. 07:36
Bah, I'm going to head off to bed and hope it resolves itself 07:38
gaal wow, i notice autrijus was pushing metameta committer bits. 08:25
sadly, the build is still broken on windows :-( 08:26
fpstring.o src\cbits\fpstring.c
NMAKE : fatal error U1073: don't know how to make 'src\UnicodeC.o'
(ghc 6.4.1 prerelease) 08:27
i'll upgrade to the official release and see.
..as i see some special Makefile moosery that has to do with 6.4(.0)? 08:30
problem still exists in ghc-6.4.1 release. 08:53
joao hello 09:06
scook0 hello!
joao :) 09:07
gaal heya joao 09:13
et al.
scook0 uh, is anyone else getting build failures? 09:18
"Failed to load interface for `Pugs'"
gaal scook0: yes.
scook0 is there a last-known-good revision? 09:19
gaal you're using ghc-6.4 yes?
scook0 oh, yeah
did we upgrade?
gaal i'm getting that with 6.4 and different errors with 6.4.1
scook0 yuck
gaal cabalization, see autrijus' journal
no, the difference in errors is explainable, we have some version-dependednt makefile moosage. 09:20
GeJ r7368 talks about Unicode.C with some Makefile.PL changes. Maybe something's missing. 09:22
gaal looking... 09:25
heh, no, that's just the branch in the makefile that i already saw :( 09:27
scook0, what OS?
scook0 Ubuntu/x86
gaal hmm. well, if you need a pugs bad, bo back a few dozen Rs, before the cabalization attempts 09:28
r7343 i think should be ok. 09:30
QtPlatypus gaal: What is cabalization?
scook0 QtPlatypus: short answer: look at autrijus' journal
it's some kind of Haskell build-related system, I think 09:31
Juerd eric256: pong 09:35
GeJ has a feeling that eric256 and Juerd will never be able to be online at the same time 09:36
Juerd That's not a problem if eric256 starts asking/saying things instead of just pinging :) 09:38
This is asynchronous communication, but some people do wish to synch before commencing
GeJ ah... a protocol issue then... I see :)
Juerd Yep
svnbot6 r7382 | gaal++ | use canonpath 10:06
xinming_Beijing is environment variable ($GHC) which can be used for pugs removed? 10:07
I just do a svn update, and can't compile anymore
scook0 xinming_Beijing: others are also having build problems 10:08
xinming_Beijing scook0: :'( 10:09
scook0 gaal suggested that if you really need to build, try reverting to around r7343 10:10
xinming_Beijing what I do is just `svn update && make ` for everyday life. :-)
I can wait. 10:11
joao nothingmuch, welcome 10:20
nothingmuch thanks
morning
joao :) 10:21
what time is it there?
nothingmuch 13:21
joao gmt+2
nothingmuch aye 10:22
but it varies
joao how?
nothingmuch well, we have the penguins
they would like to pray
but get the easiest time possible
they need to get up at sun-up to pray
so this was formalized to be a certain hour 10:23
and now they lobby each year to make the daylight savings time last as long as possible so that they won't have to get up so early
isn't that wonderful?
scook0 wow
nothingmuch however, starting next year it sould stop,
this minister from the opposing party passed a new law 10:24
and the dates should be fixed, and no longer varied according to the jewish holiday
and decided ad-hoc on a year-by-year basis
joao that's weird 10:27
autrijus I'm workin on buildsys 10:34
but lunch first
will bbl
I got broken build on w32 too.
luqui++ # journal
sorry for breaking the build for 3/4 of configs. :-/ 10:35
&
theorbtwo wonders why they don't go for a comprimise position of making the dates fixed on the hebrew calendar. 11:14
Or figuring the min and max possible dates of the high holy days, and base it on that.
rafl_ autrijus: Ok, did I break something yesterday? I was too tired to continue, sorry. 11:30
nothingmuch TaPL arrived 11:33
customs charged me $20
because $50 and $60 + $15 for shipping were not enough
*sigh*
nothingmuch thinks it's sick to charge customs for literature 11:34
scook0 nothingmuch: there's no customs-free threshold? 11:36
nothingmuch there is, but it's very low
if something is worth more than $40 or so they start charging
scook0 ouch
some people I know did a bulk ThinkGeek order recently 11:37
and got slugged ~AU$200 on a ~AU$500 order
nothingmuch holy crap
scook0 turns out if they'd split it into chunks of less than $250, they'd have payed nothing (except extra postage) 11:38
nothingmuch aye
scook0 once you go over the threshold, they really bite you 11:39
first you pay the processing fee (~$50-100 IIRC)
then you pay the actual customs charges 11:40
then you pay tax on everything
nothingmuch aye =( 11:41
scook0 I suppose I should buy myself a copy of TaPL -- I've got a library copy atm 11:45
nothingmuch should have made his mum do an interlibrary loan instead of buying 11:46
even that wouldn't have taken that long ;-)
scook0 (my uni library)++ 11:48
nothingmuch has no uni yet
i really ought to fix that soon
joao I have a copy here at home
from my univ library
nothingmuch, my univ let other people to loan books. They just have to register 11:49
(and pay a fee)
nothingmuch oh, getting books from the library is no problem
my parents are both employees
(www.math.bgu.ac.il/~kojman/, profiler.bgu.ac.il/site/public_site...r_id=1045) 11:50
scook0 but the library doesn't have its own copy?
nothingmuch hmm, that page is out of date.. mum is a prof now
anyway, the library doesn't have TaPL
but the tel aviv university does
so we could do an interlibrary loan
but that doesn't matter, I still need to start thinking about studying ;-)
scook0 thinks that after all this extra-curricular activity, nothingmuch will find real uni boring ;) 11:53
nothingmuch will probably not study compsci, if that's what you mean
scook0 ah
nothingmuch i want math & linguistics
with maybe a bit of compsci on the side for fun 11:54
Juerd There's a reply to a p6l posting that I wish to respond to, but it's not arrived yet and I accidentally deleted it already.
Frustrating!
nothingmuch Juerd: do you remember the poster? i could fwd as attachment, and you can append to your mailbox
Juerd No
nothingmuch thread?
Juerd It isn't on p6l yet
I guess he's stuck in moderation
nothingmuch oh, i get it 11:55
i thought you meant it hassn't reached the archives yet
hmm
Juerd The post suggested allowing declaring named roles in expressions
nothingmuch i like !Int better than Int(!) 11:56
but that's a nice syntactic idea
Juerd $foo but role is_default {}
nothingmuch wonders what that means ;-)
Juerd nothingmuch: Yes, I had ! first, but that doesn't play nice with the RHS
It's:
role is_default { }
$foo but is_default
nothingmuch oh
Juerd Likewise: my $subref = sub foo { ... }
nothingmuch why name the role if it's declared inline?
Juerd To refer to it later
nothingmuch i think that's confusing wrt runtime/compile time and the whole aspect of when and at what scope it's declared 11:57
in that case, can't you already do that?
wolverian Juerd, wouldn't that require you to add a ; to every named sub declaration? 11:58
nothingmuch wolverian: nuh uh 11:59
see t/syntax/something_or_other
eval {
expr;
}
+ 1; # this is a second expression - prefix + on 1 in void context
it's in one of the apocalypses
stmt_or_expr.t
wolverian oh.
I thought we didn't need that now that we have statement and expression types for constructs 12:00
but maybe it's useful anyway.
nothingmuch A02
wolverian yes, I remember now.
I just thought it didn't apply anymore.
nothingmuch how come? 12:01
wolverian see my line above
Juerd wolverian: No, there's this weird syntax rule for semicolons that I used to hate but can certainly appreciate in this new light
wolverian Juerd, yeah :)
I feel the same.
besides, after coding Haskell I don't hate whitespace rules so much anymore.
nothingmuch hardly ever notices whitespace rules in haskell 12:02
Juerd I don't hate whitespace rules
wolverian do you use a smart editor? :)
Juerd I hate indenting rules.
nothingmuch wolverian: nay, it's just that normall you indent it the way it likes anyway
wolverian nothingmuch, true enough.
Juerd nothingmuch: I tend to refactor a lot. Cutting and pasting code shouldn't break it. 12:03
nothingmuch and 90% of the functions won't really require newlines in them
Juerd I want to refactor first, test, then format nicely.
nothingmuch Juerd: in haskell?
Juerd No, I haven't used haskell yet
wolverian Juerd, I just = in vim.
nothingmuch ah
haskell tends to make you write prefactored code ;-)
Juerd wolverian: That doesn't do exactly what I want.
wolverian Juerd, oh.
Juerd My indenting rules are very simple, consistent and readable 12:04
nothingmuch = seems to get haskell wrong in my vim
Juerd Yet no editor has built in support for exactly what I mean.
integral it's nice not having to use whitespace too 12:05
wolverian Juerd, hmm, what does perl.vim do differently?
Juerd Is indentation vimfile dependent? 12:06
I thought it was a global option.
What I want is any of (, [, { and < at the end of a line to start indentation 12:07
And a ), ], }, > at the beginning of a line to end indentation, the line containing the character already no longer being indented 12:08
Obviously, the bracket that ends indentation has to be of the same type.
wolverian Juerd, it is indent file dependent 12:09
(syntax file)
Juerd And there must be no special case for words like "for", "do", ... EVER 12:10
joao nothingmuch, still related with uni studies (i was having lunch :P ) , comp. sci. + math sounds cooler :) 12:12
nothingmuch is interested in compsci, but not on a global scale
every small thing interests me for a while
but it doesn't remain exciting long enough 12:13
so maybe later i'll expand into it, but i'd like to start off with things that seem to have a better long term interest for me
joao (unbelievable. 97% of Portugal is in severe or extreme drought!)
wolverian nothingmuch, my major is CS, but I'm free to study whatever I want - really useful! 12:14
nothingmuch wolverian: how come?
wolverian nothingmuch, how come it's useful or the other point? 12:15
nothingmuch no, how did you get them to let you do it?
wolverian nothingmuch, oh, it's standard here. of course you don't get credit for the extra studies unless they form a minor (or another major)
nothingmuch ah 12:16
wolverian (or rather, you don't get credit that would count towards your diploma)
nothingmuch, lectures are also free for anyone, even non-students
nothingmuch that is also true here
wolverian nice.
nothingmuch however, IIRC the excercises and tests are for paying students only
wolverian paying? ;) 12:17
nothingmuch errm
you know
real students
paying == scholarship too
wolverian right, we don't have either.
nothingmuch people the university is getting money from
eu?
wolverian right. finland.
nothingmuch ah
joao here you pay a small fee to extra courses. but it's hard to make extra courses, because time is not enough
nothingmuch ought to study in eu
i have a citizenship
wolverian joao, the latter point is true here as well
and you have to maintain a certain rate of credit, otherwise you don't get student benefits 12:18
joao and sometimes, classes are brain killing stuff
it's more useful to learn thing by ourselves, sometimes
wolverian (but there the credit can be anything, not just that in your major/minor(s)) 12:19
joao, do you have to attend the lectures? 12:20
joao wolverian, unfortunately, in some cases, yes. 12:22
usually, students in portugal don't want to learn. they just want to get a job, and that's it. 12:23
wolverian joao, ew. :)
nothingmuch morning kolibrie 12:24
kolibrie morning nothingmuch
wolverian afternoon
kolibrie it's finally raining here - been weeks/months
nothingmuch huraah for rain 12:25
wolverian I am so waiting for winter. 12:26
kolibrie what does winter bring
wolverian snow, different (more) clothes, coldness 12:27
QtPlatypus Sad wolverian in snow?
wolverian no, happy. 12:29
kolibrie wolverian: I just realized your nick is not wolverine 12:30
which syllable does the stress go on
joao wolverian, note that I talk about my experience (some may contest...)
wolverian I have no idea. I have only spoken it in finnish and there the stress is always on the first syllable. it also doesn't mean anything.
nothingmuch drat and fsck
joao I hope it rains here, too 12:31
misc hi 12:42
nothingmuch hi misc 12:43
gaal remoose
kolibrie so, I missed the 'moose' thing, and now I am beginning to feel uneducated 12:45
nothingmuch kolibrie: moose is gaal's word 12:46
misc is the smart match operator described in exegis 4 implanted in the latest version of pugs ? i was trying to use it on 6.2.9, but it doesn't seems to work like it should..
nothingmuch it's just moose
misc: do you have any specific behhavior you'd like to be fixed?
it's at least partly implemented
kolibrie nothingmuch: okay 12:47
nothingmuch if you like, you can edit t/operators/smartmatch.t to add new failing tests
and I can give you coommit bit
kolibrie gaal: you're weird :)
misc nothingmuch: if ( @multi =~ $value ) {, to check if "$value" is in @multi
nothingmuch hmm
i think that's spelled (if ($value ~~ @many) { }
?eval 1 ~~ [1, 2, 3]
evalbot_7382 bool::false
nothingmuch ?eval (1 ~~ [1, 2, 3])
evalbot_7382 bool::false
nothingmuch ?eval (1 ~~ (1, 2, 3)) 12:48
evalbot_7382 bool::false
nothingmuch ?eval ([1 , 2, 3 ] ~~ 1 ) 12:49
evalbot_7382 bool::false
nothingmuch in s04 i see: Array Num array contains number match if any($_) == $x 12:50
{ #L<<S04/"Smart matching" /\QArray\s+Num\s+array\s+contains\s+number\s+match\s+if\s+any(\$_)\s+==\s+\$x/>>
ok(((1, 2) ~~ 1), "(1, 2) contains 1", :todo);
ok(!((3, 4, 5) ~~ 2), "(3, 4, 5) doesn't contain 2");
};
misc: you can try to fix it if you like
misc nothingmuch: i doubt, but at least, i can try to understand 12:51
nothingmuch misc: first off, can you coimmit to the pugs repo? 12:52
misc nothingmuch: no
nothingmuch: i do not even know haskell or anything :)
nothingmuch in that case, if you give me an email i can invite you
we can walk you through
gaal remoose again
kolibrie: forum2.org/moose/
misc i do not fear i will have enough time to spare for this :/
nothingmuch: if i have something to commit, i will come back to you 12:53
nothingmuch misc: okay 12:54
gaal: that moose is 1/2 of the cat from alice, right? 12:55
gaal ) 12:56
or rather, }) 12:57
nothingmuch ?
dudley_ antlers and a smile...
gaal well, to answer your question, yes, it is
nothingmuch dudley_: aye, but that didn't answer my question =)
gaal it's just the smile (and the antlers)
nothingmuch =)
some rules.t tests are *so* broken 12:58
it's hard to convert a linearization back to multidimensional group spec 12:59
bah 13:12
is(("b" ~~ /(a)|(b)/ && $/[1].from), 0, 're_tests 360/2 (#470)'); 13:13
is that test correct?
wait
yes it is
svnbot6 r7383 | rafl++ | * Proper Makefile.PL for PIL-Run. 13:22
r7383 | rafl++ | * Deleted empty lib6/disabled.
kolibrie gaal: thanks for the link }) 13:25
13:32 rafl_ is now known as rafl
svnbot6 r7384 | rafl++ | * New Makefile.PL for Perl6-MetaModel. 13:33
rafl Only the PIL2JS stuff is missing now. Then we'll be able to install perl5/ properly. 13:34
nothingmuch yummy 13:36
rafl Are there some policies on how to install JS libraries system-wide? I can't find a good example.. 13:40
nothingmuch hmm
what libraries are we installing?
rafl Perl6.MetaModel needed by PIL2JS. 13:41
nothingmuch ah
rafl It could be into PIL2JS/libjs as well, if you ask me.
nothingmuch i think that makes sense 13:42
maybe you should ask on #javascript or #json first though
rafl Hrm, why?
I'm pretty sure it'll work with the MM in PIL2JS/libjs.
nothingmuch there might be a standard we dunno about?
rafl I'll need to ask there anyway because if it's in PIL2JS I'll still need to write a Makefile to install it. 13:43
Alright. Let's ask #javascript.
They don't seem to be lucky with my question. Where's that #json you talked about? 13:46
nothingmuch err, jsan 13:52
not json, sorry
svnbot6 r7385 | nothingmuch++ | minor fixes 13:57
r7386 | rafl++ | * Moved the JS MetaModel into PIL2JS. 14:02
r7386 | rafl++ | * Adjusted paths in the tests so they run again.
r7386 | rafl++ | * Removed empty perl5/Perl6.MetaModel.
r7386 | rafl++ | * Praised Anarchism, again.
rafl OK, one Makefile missing. 14:03
svnbot6 r7387 | rafl++ | * Makefile.PL for PIL2JS. 14:14
rafl Can someone confirm that the cabalized pugs build is broken currently? 14:15
"Could not find module `readline'" seem's to be the first error.
eric256 i can.if you tell me how to test it ;) 14:16
rafl perl Makefile.PL; make :-)
nothingmuch rafl: worked for me about an hour go 14:17
ago
eric256 thought so but wanted to check
;)
rafl Hrm. Where are the lamda hackers? I don't really get those ghc errors. 14:18
eric256 hmmm....make finished...but way to fast. says "Failed to load interface for 'Pugs':
nothingmuch rafl: nopaste it all?
module readline is ghc-pkg related
try ghc-pkg list readline 14:19
rafl ghc-pkg describe readline works.
nothingmuch hmm
rafl list not, of course. It doesn't take an argument.
svnbot6 r7388 | iblech++ | * Usual svn props, also updated svn:ignore of / to contain "Setup" etc.
r7388 | iblech++ | * PIL2JS: PIL2JS.pm: Unbreak PIL2JS by adjusting the path to the JS metamodel.
r7388 | iblech++ | rafl++ for cleaning the perl5/ mess!
nothingmuch it works for me
what ghc are you using? 14:20
eric256 hmmm make is failing on feather
rafl feather.perl6.nl/~rafl/pugs_build.log
eric256: My assumption is that this problem is Debian related somehow. 14:21
pasteling "eric256" at 66.102.136.70 pasted "Failing Make" (13 lines, 939B) at sial.org/pbot/13517
eric256 your trying to make on feather too?
rafl No, on a local debian installation. 14:22
nothingmuch is Unicode.c autogenerated?
rafl: this may sound dumb, but does 'ld -lreadline' work?
rafl No. ld: warning: cannot find entry symbol _start; not setting start address - but I don't thin this is supposed to work. libreadline is installed nevertheless. 14:24
nothingmuch hmm
stevan wow,.. lots of files have moved around :)
morning all 14:25
rafl nothingmuch: Yes, UnicodeC.c is autogenerated.
Morning stevan
stevan rafl++ # doing the job no one else wanted to do :)
eric256 hmm wasn't there somethine where you had to create a fake pugs file to convince it to compile?
stevan rafl++ # you deserve a couple ++ for this actually
nothingmuch what about me? i've been doing rules.t for three days
nothingmuch never gets enough credit around here 14:26
i want a divorce, stevan
stevan nothingmuch++ # for doing something... uhm.. yeah that stuff
rafl nothingmuch: Fix my build failure first!
stevan nothingmuch: you already have the house,. the kids,.. the car,.. WTF,..
nothingmuch rafl: i doubt I can
damnit
rafl nothingmuch: Any further ideas where to start debuging?
nothingmuch i hate it when tests don't check that failure is failure for the right reasons 14:27
ok(not(....))
beh
rafl: nope =(
rafl Duh.. I can't do the perl5-installing stuff if pugs doesn't build :-(
eric256 its looking for a Unicode.c file, but all i have are UnicodeC.c files 14:28
andi have Unicode.hs ....just in case this meeans anything to someone else ;)
rafl eric256: I think I found it. 14:29
Commited. 14:30
svnbot6 r7389 | rafl++ | * Fixed typo that made the build fail.
eric256 rafl++ for having all of the most recent commits. lol
rafl Yeah, stupidly moving files around gives you a lot of commits. :-) 14:31
eric256 one step closer
pasteling "eric256" at 66.102.136.70 pasted "Pugs Build Fail" (17 lines, 1.1K) at sial.org/pbot/13519 14:32
rafl It should be syck_parse(p), IMHO. 14:35
nothingmuch bah! 14:36
for the record i think that rules.t is not good enough to be a test
the tests rely too much on bad behavior
and are not accurate enough 14:37
rafl Err, no.
nothingmuch but i've got stuff almost passing 100% via Parrot::Test::PGE
rafl eric256: I'm not too sure. Maybe including gram.c helps. 14:38
eric256 i wouldn't know. ;) 14:39
i can try something localy if you have an idea though
rafl Yes. Try #include "gram.c" in syck.c. I'm not to optimistic, though. Including .c files isn't a good Idea normally. 14:41
eric256 i put it right below #include "syck.h"
rafl Sounds good. 14:42
pasteling "eric256" at 66.102.136.70 pasted "Same error i think" (18 lines, 1.2K) at sial.org/pbot/13520 14:43
eric256 could it be that the make clean isn't cleaning well enough?
rafl I don't know. Sorry. 14:44
eric256 thats okay.
Limbic_Region is getting a different Win32 build error
rafl Well, it should AFAIK. You could try to make realclean and then remove all object files and auto-generated stuff by-hand. 14:45
eric256 hmmm....irc client didn't like me 14:52
could "/usr/bin/ar: dist/build/src/Data/Yaml/Syck_stub.o: No such file or directory" be the real heart of th matter?
in that directory i have Syck.o but no Syck_stub.o
svnbot6 r7390 | nothingmuch++ | All failures are due to PGE now 14:54
pasteling "Limbic_Region" at 129.33.119.12 pasted "Win32 Compile failure" (62 lines, 2.8K) at sial.org/pbot/13521
Limbic_Region and don't let the "Cygwin" in the path fool you, that is Active State Perl 14:55
eric256 hey thats the same error i'm getting L~R ;) (on feather) now i don't feel so alone
Limbic_Region ok - but this is on a winbugs box 14:56
eric256 yea. so its definitly/maybe something to do with Syck_stub.o ?
Limbic_Region dunno - I don't do much with Pugs these days other than verify that it compiles and runs the code that I have previously written 14:57
pasteling "nothingmuch" at 82.81.247.180 pasted "converted rules.t output" (23 lines, 1.3K) at sial.org/pbot/13522 14:58
nothingmuch does anybody know whither pmichaud? 14:59
jabbot: seen pmichaud?
jabbot nothingmuch: pmichaud was seen 99 days 19 hours 41 minutes 28 seconds ago
eric256 line 81 of build_pugs.pl seems of interest..
nothingmuch blech
Limbic_Region wanders off to a meeting but will try again later
15:00 ods15__ is now known as ods15
rafl eric256: Can you pase a full build log? Maybe the error is above. 15:03
eric256 easiest way to do that?
rafl copy&paste? make >pugs_build.log 2>&1 ? 15:04
eric256 thanks
nothingmuch blah 15:12
enough unfun for one day 15:14
svnbot6 r7391 | nothingmuch++ | remove pointless diagnosis 15:18
rafl What's in the "Chasing modules from:"-line when building pugs for the people who can currently build pugs? 15:20
svnbot6 r7392 | rafl++ | * Moved ext/XML-SAX to misc/. 15:24
r7393 | rafl++ | * Clean Setup{,.hi} and Pugs.cabal properly.
kolibrie my build from yesterday or so says: Chasing modules from: src/Main.hs
rafl kolibrie: Could you please rebuild HEAD? 15:25
kolibrie rafl: yikes! # want me to break everything :) 15:26
rafl: sure
rafl make a copy of a working version! :-)
kolibrie waits while svk pulls in yesterdays changes 15:27
rafl There were a lot, yes. 15:28
kolibrie make realclean first?
rafl Why not? :-) 15:29
Hurry.. I'll leave in a few minutes. :-) 15:30
kolibrie I'm trying first without realclean... that's faster 15:31
so far so good 15:32
the Chasing modules line had LOTS of files
gaal failsforme on win32
w ghc-6.4 15:33
pasteling "iblech" at 62.216.197.14 pasted "rafl: "Chasing modules from"" (1 line, 1.1K) at sial.org/pbot/13523
iblech (successful build)
gaal still the Failed to load interface for `Pugs' error here.
iblech Here too. 15:34
iblech cleans and tries again, just to be sure 15:35
Limbic_Region is in the middle of a realclean now 15:36
running nmake 15:38
pasteling "Limbic_Region" at 129.33.119.12 pasted "Total output of nmake up to and including failure" (221 lines, 14.3K) at sial.org/pbot/13524 15:40
Limbic_Region preceeding nmake was nmake realclean; svn up; perl Makefile.PL 15:42
iblech rafl: Now the build failed, but same "Chasing modules from" line. Also it couldn't find dist/build/src/Data/Yaml/Syck_stub.o 15:43
autrijus back 16:15
sorry, the fix is in
finally got network.
so, please rebuild; it should Work Now (tm) 16:17
svnbot6 r7394 | autrijus++ | * build fixes.
autrijus however, not yet tested with ghc 6.4 16:19
testing that now
Limbic_Region tries 16:21
autrijus - same failure - ghc 6.4 btw 16:26
autrijus ok, I still need 8min to download 6.4
have you rerun makefile.pl btw?
Limbic_Region yes - every time 16:27
nmake realclean; svn up; perl Makefile.pl; nmake
fwiw - the error isn't exactly the same as before but it is in the same place
it suggests I run setup configure 16:28
autrijus I got the fix
Limbic_Region ok
autrijus r7395 16:29
Limbic_Region tries 16:30
pasteling "iblech" at 62.216.197.14 pasted "build failure with r7395" (17 lines, 1.1K) at sial.org/pbot/13525 16:32
Limbic_Region running nmkae now
nmake even
autrijus iblech: 16:33
/usr/bin/ar: dist/build/src/Data/Yaml/Syck_stub.o: Datei oder Verzeichnis nicht gefunden
iblech: where is Syck_stub.o then?
(a critical stage in util/build_pugs.pl is to add that stub back to .a)
iblech ./dist/build/src/Syck_stub.o
autrijus in line 98
Limbic_Region same error autrijus
autrijus Limbic_Region: ok, win32 ghc 6.4? 16:34
svnbot6 r7395 | autrijus++ | * GHC 6.4 has no 'Win32' package, so skipping those.
autrijus iblech: excellent, more variable locations. can you hack build_pugs?
Limbic_Region yes - just a sec
autrijus in line 91, accomodate for more candidate positions
alternatively, hunt for one using File::Find ;)
iblech autrijus: works! committing 16:35
pasteling "Limbic_Region" at 129.33.119.12 pasted "Win32, AS Perl, GHC 6.4, r7395" (73 lines, 3.8K) at sial.org/pbot/13526
autrijus nice
iblech++
Limbic_Region whines 16:36
so what about me 16:37
autrijus Limbic_Region: r7397
svnbot6 r7396 | iblech++ | * Usual svn props.
r7396 | iblech++ | * util/build_pugs.pl: cp dist/build/src/Syck_stub.o
r7396 | iblech++ | dist/build/src/Data/Yaml/Syck_stub.o, to fix the build (at least) on Gentoo
r7396 | iblech++ | Linux x86 GHC 6.4.
autrijus Limbic_Region++
Limbic_Region Win32 are people too you know
;-)
autrijus I'm on win32 at this moment :)
svnbot6 r7397 | autrijus++ | * ditto for the 'depends' line for ghc 6.4-win32
r7396 | iblech++ | * Usual svn props.
r7396 | iblech++ | * util/build_pugs.pl: cp dist/build/src/Syck_stub.o
r7396 | iblech++ | dist/build/src/Data/Yaml/Syck_stub.o, to fix the build (at least) on Gentoo
r7396 | iblech++ | Linux x86 GHC 6.4.
Limbic_Region ok, trying r7397 16:38
iblech food & 16:39
Limbic_Region autrijus - no dice
should I nopaste the entire nmake output? 16:40
pasteling "Limbic_Region" at 129.33.119.12 pasted "entire Win32 perl Makefile.PL and nmake output" (244 lines, 13.1K) at sial.org/pbot/13527 16:41
autrijus pouts 16:43
wait, I'm installing ghc now
misc the pod from Net/IRC.pm is wrong, it talk of $bot<add_command_handler> while this no longer exist ( all other examples are good, but i decided to read doc instead of source code, just to see ) 16:47
autrijus misc: please fix away. 16:48
Limbic_Region: I'm fixing
Limbic_Region I'm Joshua, nice to meet you fixing 16:49
autrijus rofl.
misc mhh, no, it was already correted in trunk, sorry 16:50
autrijus k
Limbic_Region try now?
oh, nevermind 16:51
you were addressing misc
autrijus Limbic_Region: erase .setup-config 16:53
it should be added to clean_files (if you got tuits please add it to Makefile.PL)
Limbic_Region ok - will attempt both 16:54
autrijus and I fixed the other bug... committing 16:55
Limbic_Region autrijus - there is no .setup-config ???
does that happen after perl Makefile.PL ? 16:56
autrijus no, after an attempted build
please just fix Makefile.PL
Limbic_Region but what I am saying is that there isn't now, nor do I have any reason to believe I have ever had a .setup-config file 16:57
I will add it to the Makefile.PL though for you
autrijus yes, because configure failed
and you should get much farther with r7398.
(and at which time you would have one.)
Limbic_Region ok
and .setup-config should be in the root? 16:58
autrijus yup
Limbic_Region checking in now 17:00
svnbot6 r7398 | autrijus++ | * proper fix for GHC6.4+win32, for real!
Limbic_Region but I screwed up 17:01
re-checking in 17:02
hey - I just came across this JS link on closures jibbering.com/faq/faq_notes/closures.html 17:03
might be interesting to those working on the JS backend 17:04
autrijus Limbic_Region: so build works? 17:06
svnbot6 r7399 | Limbic_Region++ | Added .setup-config to clean_files in Makefile.PL per request from autrijus
r7400 | Limbic_Region++ | Added .setup-config properly to clean_files in Makefile.PL per request from autrijus
Limbic_Region autrijus - don't know if it works
but it certainly is making it further
it takes a LONG time to build on this machine
autrijus "nmake unoptimised"? 17:07
Limbic_Region but I don't ever run unoptimized so it really doesn't do "me" any good 17:08
autrijus k.
eric256 is the build for pugs fixed now? 17:12
autrijus eric256: I believe so. try it? 17:13
sorry for leaving it burning.
eric256 trying riight now
eric256 crosses fingers ;)
pasteling "eric256" at 66.102.136.70 pasted "Still gives these warnigns, but appears to be working" (4 lines, 220B) at sial.org/pbot/13528 17:15
iblech FYI, same warnings here 17:16
eric256 of course as far as i know it always gave those warnings. ;)
autrijus woot. 17:19
adding some prototype should fix those
but not priority for me now.
geoffb pulls to try a build also 17:23
I'm doing a no-embed build . . . but are embed builds a priority before 6.2.10? 17:25
brentdax Interesting...165 text messages/second = no voice service in Manhattan. 17:26
autrijus geoffb: it should not regress, but that's all
geoffb Yesterday, P5 embed was dead. After this build, I'll try again 17:27
Juerd eric256: pong, 2nd attempt :)
geoffb brentdax, huh?
brentdax If you can send 165 text messages a second to that area, it ties up the control frequencies so voice calls can't get through. 17:28
geoffb interesting!
Where did you learn that?
Juerd And... of course we can :)
brentdax Bruce Schneier linked to it from his blog. 17:29
geoffb I'm honestly a little surprised that doesn't happen all the time
geoffb goes to check BS blog
Juerd The single most important reason that they don't get ddossed all the time is that sending these messages is expensive :)
brentdax Not if you use the e-mail gateway, which conveniently enough is connected to the Internet, where automated attacks are easy... 17:30
Juerd Is that a free service?
brentdax Free to the sender. 17:31
Juerd We have email to sms thingies here too, but they're still expensive
Of course
theorbtwo Hm, those are rather unfortunate initials.
Juerd I don't think this service will even allow this many message :)
s
Limbic_Region autrijus - so far so good, which means I am guessing it is fixed
geoffb SMP++ 17:32
autrijus Limbic_Region: woot. 17:36
geoffb BTW, autrijus, can ghc compiles use SMP? Right now I'm just happy that my X is still responsive while I build 17:37
Looks like it built!
now to nuke a pile of stuff in /usr/local and see if the install target is sane 17:38
pasteling "gaal" at 192.115.25.249 pasted "still no luck, 6.4/win32" (16 lines, 941B) at sial.org/pbot/13530
gaal trying with 6.4.1... 17:39
in irssi, how do i set PgUp/PgDn to scroll a full screenful rather than half a page each time? this is giving me contextual anxiety. 17:42
iblech gaal: Adding an appropriate copy() call to util/build_pugs.pl:91 fixed the problem here (copy("where/ever/Syck_stub.o/is", "dist/build/src/Data/Yaml/Syck_stub.o")) 17:43
gaal iblech: thanks! so far so good on 6.4.1. 17:44
brentdax Heh...non-programmer friend of mine who follows Perl 6 stuff points out that 'Japanese "ofun" can be read as "honorable part (of something)".'
autrijus iblech: I think we need a File::Find hunt
can you take a stab at it? I need to reboot and bbiab.
gaal to hunt for what, autrijus?
autrijus geoffb: Syck_stub.o in its myriad of places in build_pugs 17:45
er
s/geoffb/gaal/
brentdax: good, we are having honorable fun
gaal iblech/autrijus, dynamically or is this a one time fix?
autrijus dynamically (I think) 17:46
gaal eeerk, hit a Strange Object ghc crash. trying again with optimizations on :)
geoffb Good lord, my /usr/local was littered. OK, time to try make install 17:47
Is "make register" now manditory, or just for Haskell hackers wanting to embed pugs? 17:48
autrijus geoffb: latter. 17:49
geoffb ok, then I'll leave that undone so that I can catch it becoming accidently mandatory at some point 17:50
make install still litters /usr/local, but not as badly, thanks rafl and autrijus
Still a lot of .t's though
autrijus the perl5 bit? 17:51
17:52 ods15_ is now known as ods15
geoffb Hmmm, I see this in the install logs: 17:52
Registering Pugs-6.2.10...
Is it both registering and warning that the user needs to?
autrijus, yeah, looks like the perl5 stuff. 17:53
Are the .hi file still needed in the install dir?
autrijus not at all 17:54
they are needed in GHC install
not in perl6
huh, how is it registering? where's the log? 17:55
geoffb /usr/local/lib/Pugs-6.2.10 is where I see all the .hi, along with HSPugs-6.2.10.o and libHSPugs-6.2.10.a 17:56
I thought the latter two files made the .hi unneeded? or just the other .o's?
autrijus just the o
.a is just an ar of o
geoffb I just saw that registering line when I scrolled back
Shillo Hullo, folks!
autrijus hi is the equiv of h
so still need to be around 17:57
geoffb autrijus, ah, OK, got it
It's the equiv of /usr/local/include
autrijus yeah. haskell does not have special .h form
its .hi is derived from the .hs
to put an interface on top of the otherwise nonselfdescript .o files
svnbot6 r7401 | autrijus++ | * do not install test files from perl5/*. geoffb++ for noticing it.
geoffb Hmmm, I built 7400, but /usr/bin/pugs -v shows 7362 still. 17:58
I wonder if the install didn't do so
geoffb rm's /usr/bin/pugs and reruns install
WTF? 17:59
autrijus hrmph?
svnbot6 r7402 | iblech++ | util/build_pugs.pl: First stab at trying to locate Syck_stub.o using
r7402 | iblech++ | File::Find to work around a Cabal bug -- works here fine.
geoffb Dammit, it's reinstalling 7362. 18:00
autrijus hm, wtf indeed
geoffb I made realclean before this latest run, so I'm mystified
bad realclean?
autrijus most likely 18:01
joao says hello
autrijus add dist/ to realclean?
greetings joao.
joao :)
geoffb autrijus, is that clean_files in Makefile.PL? 18:02
autrijus yup
geoffb Does it rm -rf? IOW, can I just add dist/ and it will recurse properly? 18:03
geoffb tries it 18:04
autrijus geoffb: it's rmrf
geoffb seems to nuked it
now trying pull and rebuild
autrijus geoffb: also, most of the existing clean_files can iirc go away
as there will no longer be any .o and .hi in src/ 18:05
geoffb One change per test. :-)
autrijus geoffb++ # commendable discipline
but "svk ci --interactive" will cure you of that :) 18:06
gaal win32 6.4.1 - r7400 works
autrijus gaal: excellent
6.4?
geoffb What is --interactive?
autrijus geoffb: aka darcs.
gaal in think not, let me see.
autrijus navigate chunks and choose which to commit.
gaal (takes a while because unoptimized barfs here)
geoffb only uses a few SVK features, and has not done a darcs push, only a darcs get
autrijus, ah, interesting
autrijus most of the time a bit annoying 18:07
sometimes very useful.
geoffb diff-chunk at a time, so multiple times per file even?
autrijus yes.
geoffb ah, yeah, I can see that being both annoying and really useful
autrijus you can choose "a" to apply all chunks.
so it's not that bad
but still a few more keystrokes.
gaal darcs has something like that too, no?
geoffb keystrokes--
autrijus gaal: svk only steals from the best 18:08
geoffb
.oO( Copied from darcs? )
autrijus ;)
gaal or is it that you just choose diffsets to apply there?
Shillo is *HAPPY* RL. Bought myself a new toy... :)
gaal hee :)
autrijus Shillo: ooh toy
geoffb Shillo, what is it?
Shillo Archos PMA 400
Linux/Qtopia PDA, with 30 gigs HD and WiFi. :> 18:09
autrijus oooh I want one
geoffb I saw someone playing with one of those at OSCON. Very schweet
joao sounds nice, how much? :)
svnbot6 r7403 | geoffb++ | Added dist dir to clean_files in Makefile.PL
Shillo Yep! :)
autrijus it's billed as "portable mp3 player", eh?
geoffb autrijus, from what the OSCON guy said, that's a massive understatement 18:10
Shillo Yeah. Need to install Emacs, Common LISP and Haskell on it, in that order. :)
autrijus indeed.
Shillo And yes, it is.
It's really a laptop. :>
joao Shillo, and pugs :)
Shillo Will it fit 64 megs RAM?
autrijus hm no keyboard?
Shillo Actually I know it will. Mmmm.... :>
geoffb Shillo, I forget, does it have 802.11 or bluetooth?
Shillo Can take USB keyboards. :>
802.11
geoffb gut 18:11
gaal www.archos.com/products/overview/pm...co2bfcjyy4
Shillo Lack of bluetooth is the only thing I have against it. :)
I can live with that, though.
geoffb hates bluetooth for security reasons
Not that 802.11 is a shining example of secure design . . . .
gaal lol
Shillo It does support WEP. :) 18:12
geoffb rolls his eyes painfully
Shillo And they toll about it so much that you just -feel- like turning it on. :)
But yeah. Crackable.
Still, BT is a *LOT* worse.
geoffb Anyone designing a security protocol to be rolled out to everybody worldwide should be forced to take it to at least Schneier before committing to it . . . . 18:13
Shillo, nodnod
svnbot6 r7404 | iblech++ | * Added ".installed-pkg-config" to /'s svn:ignore.
r7404 | iblech++ | * util/build_pugs.pl: Extremely minor refactor.
theorbtwo Reminds me, I need to get around to WEP-protecting my 802.11 setup. 18:16
Shillo geoffb: 'sides, there is the Ultimate Security Override with both 802.11 and BT:
geoffb: Walk away. :>
geoffb :-)
Shillo theorbtwo: *gasp*
theorbtwo I wish I could just ipsec it, but that'll probably be as much work as getting it working in the first place was.
Shillo, indeed. 18:17
Shillo Hmm... So many things to try with this toy... And what do I do? Irc. And MUD.
geoffb has purchased a bag of 802.11 hardware, but refuses to even plug it in until he knows how to lock it down like Alcatraz
Shillo (well, from my PC, not from the toy)
theorbtwo I just got it working a couple days ago, and the laptop and server haven't been in the same room since. 18:18
geoffb tries make install without doing any deletions first 18:19
Juerd geoffb: The best way is to have the network itself be relatively open, but as separate as possible. Then use an encrypted VPN to connect to your real network.
geoffb r7403. Much better
Juerd, more detail please . . . 18:20
svnbot6 r7405 | iblech++ | * src/Pugs.hs: Pass --pugs=$*EXECUTABLE_NAME to jspugs.pl.
r7405 | iblech++ | * PIL2JS: jspugs.pl: Ignore -BJS (needed to silence the warning the change to
r7405 | iblech++ | Pugs.hs causes).
r7405 | iblech++ | * This means pugs -BJS works now even if pugs had been installed and
r7405 | iblech++ | $sourcedir/perl5 has been rm-rfed.
Juerd geoffb: Run a linux box that is connected to both the wifi and the real network
geoffb: Run openvpn server on it
geoffb I might be able to get away with having only linux systems connected wireless, but I also might be forced to allow windows, so supporting that would be nice
Juerd geoffb: Use openvpn clients on the wifi clients
Generate and spread around some keys 18:21
geoffb What is the best way to test that you did not screw up and leave a gaping hole in the security?
Juerd An open network doesn't attract crackers
And a VPN is much more secure than WEP or WPA
geoffb knows how to secure a wired firewall . . . a wireless access point, not so much 18:22
Juerd The best way is to hire an experienced cracker.
geoffb Juerd, makes sense
Juerd An access point (without router) is nothing more than a wireless hub
Or, well, switch, even though you can still sniff everything :) 18:24
Many access points have built in routers. It's easiest to disable that.
Now, if you have a GOOD router/access point, you're in luck
Because good ones let you run openwrt
And openwrt solves the need for a separate vpn server box
geoffb sorry, AP/switch/router
netgear, IIRC
theorbtwo I don't think that's quite true.
An access point also advertises itself differently.
But possibly from the standpoint of network security...
Shillo is now shopping for a suitable keyboard... 18:25
Juerd theorbtwo: What exactly? (not true)
Shillo :)
geoffb I learned about OpenWRT a week after buying the hardware. :-)
gaal whee! r7403 builds fine here with 6.4 too.
Shillo geoffb: BTW, exactly how crackable is WEP? 18:26
Juerd Shillo: Very.
geoffb Shillo, as Juerd said
Juerd I do use WEP, though
theorbtwo Juerd, not quite sure -- you probably know more then I do.
Juerd In fact, at my office, currently the network is entirely open
geoffb Like "click button, watch progress bar, wait a rather short while"
Shillo :) :) :)
Juerd theorbtwo: What exactly? (you don't think is true) 18:27
theorbtwo hides.
Juerd theorbtwo: What's wrong with you?
autrijus rolls eyes
... I rolled a 11!
Juerd theorbtwo: You say you don't think something is quite true, and then you won't say what you're referring to, when asked.
theorbtwo: That's weird. 18:28
geoffb autrijus hits and does 432 points of damage
theorbtwo Juerd, I'm rather depressed and unsure of my knowladge in this subject.
gaal autrijus hits!--more--
Juerd theorbtwo: This leaves very little room for me to confirm your thinking...
Your knowledge won't grow this way...
theorbtwo An access point is distingished from a normal 802.11 NIC not by what it does with the packets once it gets them, but by how it advertises itself. 18:29
Thus some NICs having infrastructure, ad-hoc, and HostAP modes.
Juerd Oh, certainly 18:30
But that's at an entirely different level
Eventually, the access point just bridges it with the cabled ethernet network
It could do so in ad-hoc mode if it (and its creators) wanted. 18:31
theorbtwo That's actually what I'm doing now -- I've got my NIC in ad-hoc mode, bridged with the ethernet.
Juerd hostap isn't a mode, it's the name for a project that enables a "normal computer" (whatever that is) to be an access point
infrastructure uses the modes Master and Managed 18:32
ad-hoc uses Ad-Hoc
theorbtwo Ah.
Juerd an access point or hostap interface is Master
clients are Managed
theorbtwo OK.
Juerd The names are weird because hostap came later. Otherwise it'd probably have been Master/Slave 18:33
Note that this is all linux lingo
It may or may not be applicable to other systems
theorbtwo In that case, with the new terminology, APs are in Master mode; some desktop-style NICs can only be in Managed or Ad-Hoc and not Master.
Juerd Also, it's wireless-tools, not wlan-ng
theorbtwo: Yes, except when the desktop-style nic is actually capable of being a Master too. 18:34
theorbtwo: Some use the hostap driver for this.
autrijus gaal: did I miss your ghc6.4/win32 build failuare?
theorbtwo Ah. Now I begin to understand.
gaal autrijus: you missed my success ;) 18:36
autrijus gaal: woot :) 18:37
all clear then
gaal smoke running.
autrijus good. I'll go back to embedding.
are you still up to cleanup ext/ fest tails?
gaal just what i was about to suggest. :)
autrijus I mean, test fails. mmm dyslexia.
Shillo Going. Later, folks! :) 18:38
gaal "Our Lord is a shoving leopard" -- Rev. Spooner
geoffb OK, deleted all install files (I hope . . . I at least cleaned /usr/local/... and /usr/bin/pugs*) 18:48
Now to try a build with P5 embed
Go boom 18:59
autrijus nopaste? 19:00
geoffb Looks like it compiled the src/*Perl5* files OK (delta a couple warnings that don't look too bad)
perlbot, nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
pasteling "geoffb" at 69.110.115.185 pasted "Last lines of Perl 5 Embed death" (359 lines, 27.7K) at sial.org/pbot/13531 19:01
geoffb Sorry, forgot to increase scrollback buffer size, so that's just the last few hundred lines
Nice, Gnome bug . . . can't change the scrollback settings without some config thingie exploding 19:04
autrijus okay, I fixed the death
retesting
geoffb do I need to realclean first? 19:05
oh, not committed yet
I normally just 'svk co <remote_url>' . . . what should I do to get the same thing, but in a new local depot (I'd like to have the parrot mirror depot separate from my other mirrors, which share a depot)? 19:11
autrijus svk co <remote_url> too 19:13
but when asked, instead of saying
//mirror/pugs
say
/foo/pugs
where "foo" is the depot name
you you may wish to establish it first with "svk depot" 19:14
geoffb Will I need to do some depotmap magic first?
Ah, I guess so
autrijus I don't know if my autovivivicatoion code is that clever
try and see... patches welcome 19:15
geoffb OK, will try. :-)
nope, autoviv fails -- in fact, it just loops back and gives the whole depot path blurb and question again 19:17
gaal new smoke up (win32 snapshot, '03)
there's some trouble over at ext/libwww-perl/t/base/headers-util, it crashes but the pos is wrong 19:21
autrijus I'm still hair deep in -lperl, sorry 19:22
gaal k
eric256 begins the process of installing all his tools on a new computer after his old one died..(realy died, badly) 19:23
anyone have a favorite windows editor for perl?
autrijus eric256: eclipse? 19:24
scite?
*gasp* visual perl?
nah. s/visual perl/komodo/ if you feel compelled that way 19:25
eric256 i like free tools...i'm a mucher that way. ;)
opps 19:26
Ash eric256: XEmacs on windows is grate 19:27
gaal eric256: i use vim from cygnus. 19:28
Ash there's a gvim for win32 as well 19:29
autrijus most of the time I use that too.
more used to it than gvim.
but sometimes eclipse is nice.
gaal to make the terminal not suck completely, i use a clever hacked version of putty that connects to cygnus.
autrijus (with epic)
gaal: oooh. url?
gaal autrijus: www.chiark.greenend.org.uk/~sgtatham/putty/ 19:30
only problem is that it's not up to date with mainstream putty
eric256 hehe. well this is the place to ask for ideas... excluding vim and emac....ecplipse is good?
autrijus it's good if you have the prerequisite amount of ram. 19:31
geoffb & # lunch -- autrijus, ping me if you get perl 5 embed in a testable state again, and I'll try when I'm bak
gaal so with my real putty i get real BiDi -- i mean, wow, i never wrote a Hebrew email on linux in my life, and now i can, without HTML mail in mutt.
autrijus geoffb: I got it, but I want to make sure it works in both threaded and non-
geoffb: enjoy yer lunch
geoffb k
theorbtwo Cool, gaal. 19:32
gaal ooh, they updated their stuff to putty's version. neat.
oh wait no, did i send you to the main site? oops :)
ah, here we go. gecko.gc.maricopa.edu/~medgar/puttycyg/ 19:33
oh, and!
for a totally sicksessful experience, you want to run that with GNU screen too 19:34
which you get here
gecko.gc.maricopa.edu/~medgar/screen.html
but be warned, no detach/resume 19:35
dduncan does this error during 'make' of pugs look more like a pugs error, or problem with my system? (GHC 6.4.1 final):
Compiling Data.Yaml.Syck ( src/Data/Yaml/Syck.hs, dist/build/src/Data/Yaml/Syck.o ) 19:36
Setup: waitForProcess: interrupted (Interrupted system call)
Build failed: 256 at util/build_pugs.pl line 84.
make: *** [pugs] Error 2
it happened on 2 attempts
svnbot6 r7406 | iblech++ | New t/pugsbugs/t/pugsbugs/passing_the_pair_class_to_a_sub.t (:todo<bug>eds): 19:37
r7406 | iblech++ | Passing ::Pair to a sub fails (this is the reason for
r7406 | iblech++ | ext/Perl6-Container-Array/t/array-operations.t and
r7406 | iblech++ | ext/Perl6-Container-Array/t/array-fetch-store.t not passing).
eric256 is there a standard putty site? searching for it returns quite a few sites
autrijus dduncan: so it always stop on the same Syck?
dduncan yes 19:38
gaal eric256: top google hit
autrijus dduncan: weird...
eric256 www.chiark.greenend.org.uk/~sgtatham/putty/ ??
gaal putty++, and (its author)++ also for writing that nice piece about coroutines in c.
dduncan more pertinent question is whether this happens to anyone else
gaal eric256: yes 19:39
dduncan who here has upgraded to GHC 6.4.1, and who hasn't?
gaal dduncan: as 7403, pugs builds on my win32 with either version of ghc.
eric256 gaal thanks...it didnt' realy look offical.;) i could jsut see someone posting a hacked version that collected server addressses and passwords too.
dduncan I was on 7405 ... try that one 19:40
gaal eric256: that apparently happened! but that is the site :)
eric256 tries to remember what other tools he is missing....hmm how about perl...;) 19:41
bigh23 Is there a release date for perl 6? 19:42
eric256 later than today
early than never
bigh23 eric256: Anything more specific?
I googled for it, but with no success. 19:43
QtPlatypus bigh23: A release date for perl6 has not been set. 19:44
bigh23 Is there any estimate? half a year? year?
I want to know if I should get "lerning perl" which discusses perl 5.8.
eric256 yes. 6 will not replace 5.8 in the near future ;) 19:46
autrijus but 5.10 will, for some value of near.
eric256 5.x and 6.x will probably coexist for a very long time. ;) 19:47
gaal is the "Expected:" string in verbose output officially part of TAP?
bigh23 eric256: OK, thanks, I will give it a try.
pjcj gaal: I believe so, shich 19:48
which is why Got can't be changed to something better 19:49
or the order swapped
iblech [FYI] New JS smoke up, 90.89% ok
eric256 wow
gaal pjcj: this is too bad, because it means that isnt(1, 2) 19:50
no, isnt(1, 1)
proclaims "Expected: 1, Got: 1"
which can be confusing.
pjcj yes, it confuses me 19:51
don't take my word as gospel in thus matter though 19:52
dduncan well, I got the compilation to continue on past the YAML
gaal it's extremely tempting to fix, but i don't know what i'd break if i do :(
dduncan now there's a different problem 19:53
more like what I reported yesterday
gaal hahahaah! if you look at a smoke you'll notice red on ext/Set/t/overload.t 19:54
dduncan when/following 'Skipping Pugs', it gives a warning in gram.c about syckparse
gaal but looking at the failing tests you'll see they are TODOed
eric256 whats the feather servers domain name?
pjcj gaal: why do you get that output from isnt(1, 1) ?
iblech gaal: I fixed that. The test descriptions contained "#", causing the harness to not recognize the TODO bug 19:55
gaal the reason: the desc contains a #, and the code that picks the # TODO up is confused :)
yes
iblech d'oh, but I forgot to ci
gaal pjcj: you get it because the test fails
iblech: i'm working on TODOing ext myself, so let's divide and conquer. 19:56
pjcj doh! it's been a long week
iblech gaal: ext/Perl6-* will be fixed as sonn as t/pugsbugs/passing_the_pair_class_to_a_sub.t is fixed 19:57
dduncan the last few make failure lines say:
ld: archive: dist/build/libHSPugs-6.2.10.a has no table of contents, add one with ranlib(1) (can't load from it)
Build failed: 256 at util/build_pugs.pl line 135.
make: *** [pugs] Error 2
svnbot6 r7407 | iblech++ | ext/Set: t/overload.t: Fixed test description to not include #, so the test 19:58
r7407 | iblech++ | harness doesn't get confused.
r7408 | iblech++ | util/version_h.pl: The location of the .hi and .o files changed because of the
r7408 | iblech++ | Cabalization, thus Version.hs didn't get rebuilt. Fixed.
rafl iblech: For me the first item is readline. 20:01
geoffb: Thought /usr/local is the default prefix.
wolverian eric256, feather.perl6.nl 20:03
eric256 thanks...it was among the numerous other things lost.../me mourns the loss of his old friend
Juerd eric256: You pinged me?
eric256 way long ago...now i don't have time but it was mentioned to me that you were looking for someone to run the automated scripts on feather 20:04
iblech rafl: Hm? You mean the readline bug from a few days ago?
eric256 just wanted to volunteer if you needed anything
Juerd eric256: That is true
eric256: Good; thanks
eric256: I'll copy the stuff over to your homedir soon
eric256: You can then do with it what you think is best
rafl iblech: Dunno. I'll try current HEAD now. 20:05
eric256 alright
svnbot6 r7409 | iblech++ | Fixed ext/FindBin/lib/FindBin.pm, ext/Config-Tiny/t/basic.t, and 20:07
r7409 | iblech++ | ext/Config-Tiny/lib/Config/Tiny.pm WRT "use fatal" being in effect by default
r7409 | iblech++ | (at least in current Pugs).
rafl iblech: Same error.. 20:08
iblech gaal: FYI, r7409 will be my last fix in ext/ for today
Juerd eric256: done
rafl iblech: Working on Config-Tiny? Please give me a method to read from strings and fhs. And maybe use proper OO.. :-)
iblech rafl: Could you give me some failing code?
rafl iblech: Yes.
Juerd eric256: It's in ~/auto
rafl iblech: Hold on, it's being uploaded to feather.
Juerd eric256: See also crontab -e
rafl iblech: feather.perl6.nl/~rafl/pugs_build.log 20:09
eric256 so every 15 minutes it rebuilds pugs on feather? 20:10
iblech rafl: I've already ported Config::Tiny to real OO, but it's hosted on FreePAN, as Pugs was not nearly ready for OO at the time I ported it. Limbic_Region then ported it to fake OO. So we might be able to just copy my OO Config::Tiny port to ext/
eric256 and parrot it would seem
Juerd It's funny that so much time and effort is spent by so many programmers and system administrators all over the world, just to enable nice prompts...
eric256: Yes, but only if they have updates available
geoffb rafl, yes /usr/local is right. My complaints were that way too much was in /usr/local, and also in too many different subdirs of /usr/local.
It's getting better.
rafl iblech: Yeah, please!
iblech rafl: Ah, this error (I thought you were referring to P6 &readline). Sorry, no idea...
geoffb Thanks to you and autrijus. :-)
rafl geoffb: --verbose please. What directories don't you like?
autrijus: Please take a look: r7409 fails to build for me: 7409 20:11
autrijus: Please take a look: r7409 fails to build for me: feather.perl6.nl/~rafl/pugs_build.log
geoffb Can't give detail right now, because I deleted everything before starting my embed build/install test, which failed to build
I will post details next successful install.
rafl autrijus: libreadline and the haskell readline package are installed.
geoffb: Yeah, just highlight me. 20:12
eric256 was there anything specific that needs fixed? or am i just in charge of it incase something breakes?
geoffb rafl, will do (I assume using your nick is enough)
gaal iblech: ACK, thanks.
rafl geoffb: of course.
autrijus rafl: ack 20:13
rafl rafl: Not reproduceable on feather, which is also a debian box, though. 20:14
dduncan I'm going to try isolating whether the GHC upgrade broke anything ... 20:16
rafl Which ghc upgrade?
dduncan to 6.4.1 from 6.4.0
geoffb autrijus, how is p5 embed coming? 20:17
rafl feather still has 6.4.0.
autrijus geoffb: I ran into some Cabal-1.0 bugs.
rafl It works there.
autrijus I can commit what I have and let you test
geoffb autrijus, bleah
ok, do
dduncan I did both a GHC upgrade and applied some system security patches between when Pugs last compiled and when it failed 20:18
eric256 Juerd the script fails because its trying to cd to your directories
dduncan right now, I'm going to start off with trying to compile the Pugs 6.2.9 off CPAN, and a fresh checkout of the repository also, under the newer GHC and system
eric256 i'' change to cd to mine...just didn't know if there is something else you need to change 20:19
autrijus geoffb: done.
dduncan if those fail, I'll try reverting GHC to 6.4.0
Juerd eric256: I have no idea, really.
geoffb autrijus, OK, making
rafl dduncan: I also have 6.4.0.
svnbot6 r7410 | autrijus++ | * various weird voodoo to get perl5 embedding working. work in progress.
autrijus can you put one .a file into another?
pjcj gaal: See I think it may not be part of TAP per se, but see www.mail-archive.com/perl-qa@perl.o...3999.html, especially chromatic's reply 20:20
autrijus tries and see
eric256 Juerd: well you tell me...i don't even know what you are expecting me to do with it. ;)
hey you changed the file in my directory....spooky 20:21
dduncan I just discovered something else ... as of Pugs 6.2.9 ... having space characters in your current path causes 'make' to fail before it starts: 20:23
Darren-Duncans-Computer:/Volumes/Programming160/Parrot and Pugs distro archives/Perl6-Pugs-6.2.9 darrenduncan$ make
Makefile:449: *** target file `/Volumes/Programming160/Parrot' has both : and :: entries. Stop.
rafl Can't someone reproduce the 'readline'-problem I run into? 20:24
dduncan I can work around this easily, but it doesn't look like something a script should have a problem with ... it should escape as necessary
rafl autrijus: Maybe I can provide you some debugging info now? I'll be offline in about 5 to 10 minutes.
svnbot6 r7411 | gaal++ | avoid readlink on MSWin32 20:25
autrijus rafl: try commenting away line 130, among other lines, in build_pugs.pl 20:27
geoffb perlbot nopaste 20:29
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
pasteling "geoffb" at 69.110.115.185 pasted "p5 embed dead again" (215 lines, 14.3K) at sial.org/pbot/13532 20:30
rafl autrijus: 132? (the one mentioning readline) - I commented it out. Still doesn't work.
geoffb autrijus, see new paste -- seems to complain more now 20:31
autrijus geoffb: yeah, I think I nailed it 20:32
rafl: :-(
rafl: I can't reproduce.
rafl autrijus: Hrm.. no way. I'll need to go now unfortunately.
autrijus: I can give you shell access tomorrow, if you want. 20:33
Bye!
geoffb bye, rafl
autrijus rafl: ok. bye!
geoffb: committing 20:35
geoffb pulling and making after realclean
svnbot6 r7412 | autrijus++ | * link against perl5 embed. it would work for static perl5
geoffb And it's off . . . 20:36
(sounds of steeplechase . . . )
afk for a few while this builds 20:37
eric256 any svn gurus out there? since i'm starting fresh i was hoping to svn most everythign i work with ....is there a quick way to create an empty repository and ilnk a directory to it? seems like svn wants me to create, import, and then check out....which seems kind of silly 20:44
Juerd eric256: Root isn't spooky :) 20:45
autrijus eric256: svk import --to-checkout
eric256 autrijus do you have stock in svk or something? ;) 20:46
autrijus eric256: I happen to be its second committer :)
eric256 ahh ;) website? 20:47
svnbot6 r7413 | autrijus++ | * rename perl5.[ch] into p5embed.[ch] to avoid silly
r7413 | autrijus++ | case-insensitive filesystems.
autrijus svk.elixus.org
eric256 if i have many different projects (from random scripts to websites) are they better each in there own repository? or one repository for everything? 20:48
michael autrijus, so it's possible to use svk with a vanilla svn (remote) repository, right?
autrijus michael: sure yes
it's just a client.
eric256: I prefer one repo for everything, but immv. 20:49
s/immv/ymmv/
eric256 svk.elixus.org/?SVKWin32 :(
michael autrijus: do you happen to know whether it works with clients like TortoiseSVN or RapidSVN?
autrijus michael: it does. 20:50
michael very cool.
one more reason to convince $work to switch to subversion :-)
(from CVS, that is)
eric256 nm...page finaly came up after giving me errors several times
michael (I fear darcs is no option as they live and die with Eclipse's "Synchronise" feature, whatever that is)
gaal pjcj: i decided it's okay to change, and i made note of the lexical advice there. 20:51
pjcj gaal: will you change the order too?
gaal no, wasn't convinced that's important. :) 20:52
svnbot6 r7414 | gaal++ | when `isnt(1,1)' fails, don't say you expected 1 (you expected anything but 1)
pjcj as you wish, but I'd prefer the expected first ;-)
gaal oh oops 20:53
the expected already is first :)
pjcj aha! 20:57
gaal dduncan: for when your build works: two tests in LKT are failing, please have a look
i'm off to bed, i think we're pretty close to all green.
svnbot6 r7415 | putter++ | STATUS: p5 backend correction
autrijus gaal++ 20:58
gaal i barely changed anything today :)
putter briefy...
eric256 still odesn't have enough memory to compile pugs on windows... jeez
autrijus geoffb: how's the build?
eric256: hm? not at all, or just need to restart twice? 20:59
eric256 not at all. 750mb right now...gonna see if i can rearrange sticks and get a full gigi in here
gaal ext\libwww-perl\t\base\headers-util.t is weird, but i'm too tired to track down the failure now 21:00
good night! 21:01
autrijus eric256: this is very strange as I only have 768m.
eric256: what is the err?
gaal: night!
autrijus turns to parrot embed
leo__ autrijus: don't forget to make realclean after svn up
autrijus noted. 21:02
but I'm testing 0.3.0 first :) 21:03
svnbot6 r7416 | autrijus++ | * dynamic perlembed _should_ work now.
eric256 rebooting and reorganizing right now...i'll let you know when i get back to the error though 21:04
leo__ okie - but realclean applies to 0.3.0 as well as trunk head r9400 - and any witsching between these versions
autrijus ok
geoffb autrijus, sorry, afk for longer than expected. Looks like it failed again, though. 21:10
Pasting . . .
pasteling "geoffb" at 69.110.115.185 pasted "P5 embed failure #3" (29 lines, 6.9K) at sial.org/pbot/13536 21:11
autrijus geoffb: r7417 21:12
geoffb OK, realclean, pull, build . . . 21:13
autrijus hands geoffb "unoptimised"
geoffb Should I stop the opt build?
svnbot6 r7417 | autrijus++ | * first step toward parrotbedding.
geoffb trying to minimize variables during debugging 21:14
autrijus well, go ahead then, it's valuable
seeing that I did not make clean
geoffb doing 'make realclean; svk pull; perl Makefile.PL; make' each time
Putter, I saw you changed Perl 5 backend STATUS to <50% . . . but the smokeserve said like 52% last time I checked. What happened? 21:19
er, putter
autrijus geoffb: note the number of tests run.
geoffb yep, and that's why I'd said "of main tests (no ext/)" 21:20
:-)
Because otherwise it wouldn't match the way iblech counts PIL2JS quality
autrijus ah. well then
I think it's 52%.
geoffb & # brb, build still in progress 21:21
ooh, looks like success 21:31
svnbot6 r7418 | autrijus++ | * parrot embedding cabalized. I think.
geoffb installing
autrijus good!
whew.
geoffb Are you ready for me to try a dual-embed?
w00t: Version: 6.2.9 (r7417)
autrijus goahead
woot
eric256 the svk.elixus.org server doesn't seem very stable... 21:32
eric256 means the web server...as i keep getting proxy errors
and i'm not behind a proxy
Juerd Proxies can be used on both ends 21:33
geoffb oh, btw, looking at /usr/local . . . the tests for PIL2JS are getting installed, presumably because they are .t.html
Juerd It can be very useful to cache your own website.
autrijus geoffb: yeah, but I don't think we need to do much about it now
Juerd Or to balance the load using a proxy
geoffb nah, just fyi
autrijus geoffb: unless you feel strongly, in which case go hack util/src_to_blib.pl :) 21:34
Juerd Or to integrate several different kinds of web server software on a single hostname
geoffb Bah, it's been about 2000 revs since I built parrot. What's the standard thing?
ah Configure.pl
dduncan hello again 21:36
geoffb hears The Cars in his mind 21:37
dduncan ping gaal
geoffb er, that was really fast . . . does parrot normally build in like a minute or two?
dduncan I've known about those failures for a few days ... they are related to hashes-and-arrays/pairs.t not working
when pairs.t works, LKT should work
Juerd Fuck. 21:38
It's not good to type "halt" in a remote shell.
Good thing this box is connected to an IP power switch.
autrijus Juerd: indeed
Juerd should learn to use the power button instead of commands.
dduncan or, I could be wrong ... will look into LKT again
autrijus geoffb: yes, what a refreshing change from pugs. 21:39
geoffb autrijus, wow, impressive
Juerd ACPI works very well. IFF you use it.
autrijus geoffb: but not if you throw ICU in
dduncan fyi, Pugs 6.2.9 seems to 'make' fine under my current system
autrijus which is considerably longer than pugs.
dduncan will try a fresh checkout next
geoffb autrijus, ah.
What should be in PARROT_PATH? I used the root of my parrot checkout, with trailing slash . . . but Makefile.PL spewed several 'Can't exec "parrot"' warnings 21:44
autrijus geoffb: have you built it?
geoffb nod, and installed
autrijus geoffb: please chase Makefile.PL
in parrot_config
that needs to be fixed 21:45
geoffb nodnod, I just wanted to make sure I had the expected dir in PARROT_PATH,
before assuming the code was wrong
dduncan fyi, I'm currently on 'make' of a fresh checkout 21:48
eric256 is there some reason that GHC wont use the pagefile while compiling? i mean if its gonig to be a memory hog...why isn't it nice about it? 21:50
autrijus i'm just about to start a new make, then i'll have the error for you 21:51
autrijus eric256: cool 21:53
dduncan on a side-track, I have a question about semantics concerning whether a Perl 6 module goes in /ext or /misc ...
geoffb smacks his forehead
Duh, it's for lack of ./ -- OK, can I add that and not break Win32? 21:54
dduncan if a module doesn't work "today" but it is expected to work by the time of the next Pugs release, can it go in /ext right away, or should it stay away until it works "today"?
autrijus dduncan: go into ext/ with failing tests.
eric256 hands geoffb a giant pillow (with a 2x4 hidden in the middle)....here this will hurt less
geoffb thanks, eric256 21:55
autrijus back... was hacking install logic in parrot's build system
leo__ autrijus: thanks 21:56
eric256 geoffb its always the little things that drive us nuts (spent 2 hours fighting with an upload script the other day only to find a typo in the actual HTML...;) )
geoffb ouch
eric256 perlbot nopaset 21:57
perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
pasteling "eric256" at 66.102.136.66 pasted "Memory Error" (7 lines, 393B) at sial.org/pbot/13537 21:58
eric256 ehh... i just realized that isn't the same error i was getting before
geoffb OK, committed (and some SVK DWIM), off for another build
svnbot6 r7419 | geoffb++ | Makefile.PL: SVK-friendly DWIM for parrot dir, plus exec ./parrot rather than just parrot
eric256 still don't know what it means though
autrijus geoffb: cool. so it built? 22:00
ah, still building
geoffb autrijus, yeah, still building 22:01
eric256 only now realizes that the evalbot includes the rev number in its now. ;) 22:03
autrijus any idea what causes that error?
autrijus sorry, a sec 22:06
geoffb OK, dual-embed failed. 22:07
Pasting 22:08
pasteling "geoffb" at 69.110.115.185 pasted "Dual-embed failure" (268 lines, 29.8K) at sial.org/pbot/13538 22:09
eric256 ?eval $?VERSION
evalbot_7419 Error: Undeclared variable: "$?VERSION"
geoffb ?eval $?PUGS_VERSION 22:10
evalbot_7419 \'Perl6 User\'s Golfing System, version 6.2.9, August 3, 2005 (r7419)'
eric256 ahhh
i was suprised it managed to get the new version compiled and restart itself in 4 minutes. pretty impressive
autrijus eric256: this is defintely not memory related 22:12
eric256: ghc 6.4.1 unoptimised?
eric256 yea
autrijus try optimised
geoffb: fixed 22:13
geoffb committed?
geoffb building as of 7420 22:14
svnbot6 r7420 | autrijus++ | * parrot embedding, try #142857.
geoffb autrijus, at least you don't have to do *everything* yourself. Just the hard part. :-) 22:15
autrijus lol.
dduncan fyi, the clean checkout had a failure under my current setup, whereas the clean 6.2.9 worked fine ... both at 'make' time 22:16
ld: archive: dist/build/libHSPugs-6.2.10.a has no table of contents, add one with ranlib(1) (can't load from it)
Build failed: 256 at util/build_pugs.pl line 165.
wilx You are missing Template Haskell support it seems.
Forgot to bootstrap GHC?
Or maybe not. 22:17
dduncan I didn't do anything special, just installed GHC
just a minute ...
SamB dduncan: were you the guy who came to #haskell with that problem? 22:18
dduncan I never went to #haskell at all
I only mentioned any issues on #perl6 22:19
if Template Haskell is a separate install from the normal GHC, then I don't have it
rep it's not 22:20
afaik, at least
SamB dduncan: I think someone mentioned a similar problem in #haskell. maybe ask there?
autrijus dduncan: how did you install ghc? 22:22
dduncan in a minute ...
I downloaded this: www.uni-graz.at/imawww/haskell/GHC-6.4.1.pkg.zip 22:23
it had an installer, which I ran
SamB dduncan: in fact, I believe they figured out what was causing it...
eric256 yea!!! it worked
dduncan my make failure does mention template haskell
eric256 thansk autrijus
dduncan before the error line I reported, it said: *** Building: /usr/local/bin/ghc -package stm -package network -package mtl -package template-haskell -package base -package unix -idist/build -Ldist/build -idist/build/src -Ldist/build/src -o pugs src/Main.hs -lHSPugs-6.2.10 -L. -Lsrc -Lsrc/syck -Lsrc/pcre 22:24
autrijus eric256: cool
dduncan but SamB, as you were saying ...
autrijus dduncan: hm, it seems OSX wants ranlib 22:25
geoffb autrijus, same error at 7420
eric256 any ideas why opimized failed? and should we add a note to that effect somwhere?
autrijus dduncan: what if you run ranlib with it?
geoffb: yow. sec
eric256: you mean "unoptimised" failed?
geoffb k
eric256 ehhh...yea
autrijus eric256: no, no idea why, must be perlbug
SamB I think I heard something about cabal not using libtool...
autrijus er I mean ghcpug
sigh I mean ghcbug. 22:26
(guess that shows how ingrained p5 is to me)
eric256: and gaal could reproduce it, so it's not your setup
somehow I can't though.
hm.
rafl OK, what causes my build failure (that readline thingy) is that ./Setup calls ghc with '-DPUGS_HAS_READLINE=1 readline' instead of '-DPUGS_HAS_READLINE=1 -package readline'. 22:27
autrijus eric256: it may worth documenting, but I don't know how -- in INSTALL perhaps
rafl I suspect it's a Cabal bug, isn't it?
autrijus rafl: ooh.
dduncan autrijus, there is a 'ranlib' on my system, but I don't immediately know what to do with it
rafl Cabal-1.0 as shiped by ghc 6.4.0.
dduncan I'm not an expert with the C side of things
autrijus dduncan: just run
ranlib dist/build/libHSblahblah.a
geoffb
.oO( Why, after half a century of programming, are build systems still such a bloody pain? )
autrijus where the blahblah is tab completed
rafl autrijus: "ooh", what? :-) 22:28
autrijus rafl: look at your Pugs.cabal
SamB geoffb: because they either don't abstract enough or the abstractions don't cover enough
autrijus see where readline occurs
dduncan I'm svn-up again ...
rafl autrijus: -package readline. 22:29
autrijus rafl: so something is fishy.
rafl: try removing the two occurences of readline, each at a time, ./Setup configure and ./Setup build 22:30
and see if things improve.
rafl The define as well?
autrijus the readline -any 22:31
in build-depends
rafl Removing the build-dependency doesn't improve things. Removing the -DFOO -package readline does. Not sure if it will work as it is finished, though. 22:33
dduncan okay, I did a: ranlib dist/build/libHSPugs-6.2.10.a 22:34
but a subsequent 'make' gave the same error
will try Makefile.PL again ...
autrijus rafl: I wonder if adding the -DFOO in will help 22:36
dduncan: a sec 22:37
dduncan btw, template haskell 1.0 is in my system, according to the makefile/make output
autrijus, I'll wait to see what you come up with ... my other tries failed 22:40
autrijus dduncan: up and try again
svnbot6 r7421 | autrijus++ | * do a ranlib for OSX, prompted by dduncan. 22:41
autrijus just straight "make"
dduncan seems to be doing something right ... 22:42
geoffb autrijus, ping me if you want me to try a dual-embed build again, my attention is mostly on $work right now 22:44
autrijus geoffb: yeah, I'm will fiddling with this on my laptop
geoffb k
dduncan still working ... 22:45
autrijus, your change got 'make' to continue on and generate the prelude ... now its on the second compile pass, working on Pugs.Run ... taking awhile
autrijus it would 22:46
but I think it works
dduncan I'll announce when something else happens
right, 'make' is done! 22:47
now trying a smoke 22:48
thanks for that ... good to have experts around! 22:49
rafl autrijus: OK, compilation succeeded. 23:03
autrijus: However, there's no pugs executable.
autrijus: So what's going wrong here? There's -package readline in Pugs.cabal but ./Setup omit's the -package when calling ghc? Seems to be a Cabal bug to me. 23:04
autrijus rafl: hm, "compilation"? 23:06
./Setup build
does not build executable
rafl: it may be a Cabal bug, but we don't need that extra -package anyway in the ghc cmdline 23:07
so maybe we can safely omit it.
rafl autrijus: OK, will you do the according changes? My development box doesn't have a connection ATM.
autrijus I'll do so
rafl Would be nice to put the diff somewhere on the net so I can do the same changes locally without needing to use websvn *brrr*.. 23:09
leo__ autrijus: parrot embed/install is ok? 23:10
autrijus leo__: yes
leo__ woot - super
autrijus geoffb: fixed (I think.)
geoffb r#? 23:11
autrijus still building here but got past the bogus point
7422
geoffb OK, making 7422 after realclean 23:12
autrijus gawd 23:13
one remaining problem
src/global_setup.c:34: undefined reference to `parrot_get_config_string'
leo__: what should I link against?
leo__ installed exec? 23:14
autrijus ah, need to put ./src/parrot_config.o
right?
hm weird, parrot_config.o is in the .a already 23:15
svnbot6 r7422 | autrijus++ | * remove the redundant "-package readline" from ghc-options at
r7422 | autrijus++ | rafl's suggestion to work around yet more Cabal-1.0 bug.
r7422 | autrijus++ | * also, normalize the -D flags to not use =1.
autrijus I don't get it. :-/
it should've been caught by build_pugs line 126 23:16
leo__ $(SRC_DIR)/install_config$(O) \
after 'make install' in parrote
autrijus leo__: right, it's just linking against parrot by its own is no problem
somehow parrot+perl5 fails
however, I need to find food and (maybe) nap a bit
leo__ parrot_config is the uninstalled parrot 23:17
autrijus yeah, I'm linking against the uninstalled one
for this build
leo__ install_config knows about --prefix, to find it's libs
autrijus yup
the thing is why doing p5+parrot fails
yet parrot and p5 by their own succeeds 23:18
geoffb: I'll bbiab
but please see if you run into the same bug.
geoffb autrijus, OK, I'll paste the result
autrijus geoffb++
geoffb (when it comes)
autrijus++ # persistence in the face of adversity
leo__ thinks that pugs install could just build install_config.o as the parrot Makefile does, then continue install 23:20
pasteling "geoffb" at 69.110.115.185 pasted "Latest dual-embed failure" (268 lines, 29.9K) at sial.org/pbot/13543 23:24
geoffb autrijus, latest failure pasted
(that was 7422, opt dual-embed, for those in the peanut gallery)

Powered by Cro and the Raku® Programming Language.

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

Thank you!