Parrot 0.6.3 "Beautiful Parrot" Released | parrotcode.org/ | 5/649/88 new/open/stalled tix | logged in irclog.perlgeek.de/parrot/today
Set by moderator on 26 June 2008.
00:04 kid51 joined 00:10 AndyA joined
nopaste "bacek" at 211.29.157.151 pasted "Patch for #56516 (for moritz)" (15 lines) at nopaste.snit.ch/13443 00:10
bacek_ moritz: forget this patch :) 00:11
purl bacek_, I didn't have anything matching this patch :)
00:34 TiMBuS joined 00:40 cjfields joined 00:53 petdance joined 00:57 cjfields joined
nopaste "bacek" at 211.29.157.151 pasted "Another patch for #56516 (for moritz)" (23 lines) at nopaste.snit.ch/13444 00:57
bacek_ moritz: it's fixing parrot's repeat. And I'm not sure that this is 'right way'. But there is too many calls to 'repeat' in rakudo... 00:58
01:00 rdice joined
pmichaud in general we don't fix rakudo bugs by changing Parrot unless the Parrot behavior is obviously a bug 01:13
since throwing an exception is a reasonable behavior for repeat, that means Rakudo needs to do something different.
01:14 acalhoon joined
cotto_work rt.perl.org/rt3/Search/Results.html...20'__NEWLI 01:16
NE__'%2C%0A%20%20%20''%2C%20%0A%20%20%20'%3Csmall%3E__Requestors__%3C%2Fsmall%3E'%2C%0A%20%20%20'%3Csmall%3E__CreatedRelative__%3C%2Fsmall%3E'%2C%0A%20%20%20'%3Csmall%3E__ToldRelative__%3C%2Fsmall%3E'%2C%0A%20%20%20'%3Csmall%3E__LastUpdatedRelative__%3C%2Fsmall%3E'%2C%0A%20%20%20'%3Csmall%3E__TimeLeft__%3C%2Fsmall%3E'&Query=Queue%20%3D%20'parrot'%20AND%20Status%20%3D%20'open'&Page=7&Order=ASC&OrderBy=id
bacek_ pmichaud: got it. 01:17
nopaste "bacek" at 211.29.157.151 pasted "Yet another patch for #56516 (for moritz/pmichaud)" (16 lines) at nopaste.snit.ch/13446 01:20
bacek_ pmichaud: this one fixes 'infix:x' only. 01:21
And actually all other 'infix:x' just call this one.
pmichaud need to initialize $S0
bacek_ pmichaud: If the count is less than 1, returns the null string. 01:22
This from S03...
I've made mistake once again? 01:23
pmichaud only in assuming that $S0 is always initialized to a null string.
bacek_ ouch.
And how initialize S0 to 'null string'? 01:24
pmichaud $S0 = ''
but you could also just do .return ('')
bacek_ I didn't understand. "" is empty string for me. Not null... 01:25
pmichaud also, there's no 'n' symbol declared :-) :-)
bacek_ yak.
nopaste "bacek" at 211.29.157.151 pasted "'guess what?' for #56516 (for moritz/pmichaud)" (23 lines) at nopaste.snit.ch/13447 01:28
01:29 Theory joined
bacek_ pmichaud: is there is difference between 'empty string' and 'null string'? 01:29
afk # 01:30
pmichaud depends on the context. Sometimes the terms are used interchangeably. Sometimes people use "null string" when they really mean "null pointer". 01:35
in this case, I'm fairly certain that S29 means "empty string"
dalek r28928 | pmichaud++ | trunk: 01:37
: [rakudo]:
: Make sure the expression in a for statement is interpreted in
: list context. cjfields++ for finding and reporting this.
diff: www.parrotvm.org/svn/parrot/revision?rev=28928
01:45 magnachef joined
chromatic found the closure bug, if not the fix 01:53
dalek r28929 | coke++ | no_builtin_methods: 01:54
: remove builtin methods
diff: www.parrotvm.org/svn/parrot/revision?rev=28929
chromatic Delete line 134 in src/pmc/closure.pmc -- some tests fail, but it fixes the double-:outer bug. 01:55
RT #56184
So far, only one test in t/op/lexicals.t fails though. 01:56
DietCoke One reason not to commit json is that now we have compilers/json and languages/json, a tad confusing. (is there a readme in languages to clarify that?) 01:58
02:00 magnachef_ joined
DietCoke I have a double outie. It's kind of kleinish. 02:00
pmichaud ... I'm wondering why line 134 would make a difference 02:04
it seems to me that (for the example code I posted in RT #56184) we would never get to line 134 02:06
oh, wait, never mind -- wronge xample.
02:07 workbench joined
DietCoke chromatic: you've added ops. Is there magic other than running the special .mak target? 02:10
hurm. perhaps reconfigure.
nope. 02:11
ah. looks like you have to add them manually. lame. =-) 02:29
kid51: with the current system you have to manually add an op to ops.num when you add an op. who knew! 02:33
msg kid51: with the current system you have to manually add an op to ops.num when you add an op. who knew! 02:34
purl Message for kid51 stored.
02:43 cjfields joined 02:50 Psyche^ joined, cjfields joined 02:52 cjfields_ joined
dalek r28930 | jkeenan++ | opsrenum: 02:53
: Eliminate %seen from inside renum_op_map_file(). By using it, the program was
: just endlessly rewriting the ops already seen in the *old* file to the new
: file -- but failing to pick up new ops.
diff: www.parrotvm.org/svn/parrot/revision?rev=28930
kid51 DietCoke: Just now looking at IRC for first time in hours.
DietCoke pokes dalek. 03:06
dalek r28931 | coke++ | no_builtin_methods:
: first pass, delete say builtin support and create say
: opcodes instead.
diff: www.parrotvm.org/svn/parrot/revision?rev=28931
03:27 contingencyplan joined 03:32 cjfields joined
cjfields perl6: perl6: for [1..3],[4..6] { say $_.WHAT} 03:35
polyglotbot OUTPUT[Statement not terminated properly at line 1, near ": for [1.."ā¤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)ā¤called from Sub 'parrot;Perl6::Grammar;eat_terminator' pc 20609 (src/gen_grammar.pir:2813)ā¤called from Sub
..'parrot;Perl6::Grammar;statementlist' pc 19622 (src/gen_grammar.pir:2450)ā¤called fr...
cjfields perl6: for [1..3],[4..6] { say $_.WHAT}
polyglotbot OUTPUT[Arrayā¤Arrayā¤]
cjfields perl6: for [1..3] { say $_.WHAT} 03:36
polyglotbot OUTPUT[Intā¤Intā¤Intā¤]
pmichaud polyglotbot is out of date.
cjfields yes, but I updated and get the same thing
pmichaud > for [1..3] { say $_.WHAT }
Array
cjfields hmm... 03:37
pmichaud perl6: for [1..3,4..6] { say $_.WHAT }
cjfields rebuilding
polyglotbot OUTPUT[Intā¤Intā¤Intā¤Intā¤Intā¤Intā¤]
pmichaud > for [1..3,4..6] { say $_.WHAT } 03:38
Array
cjfields what does #perl6 give?
pmichaud 03:38 <pmichaud> rakudo: for [1..3,4..6] { say $_.WHAT } 03:39
03:38 <p6eval> rakudo 28931: OUTPUT[Arrayā¤]
cjfields Never mind. The rebuild works. pmichaud++ 03:40
I'm adding a few tests for this (t/statements/for.t). Does any on this need to go into t/spec? 03:41
pmichaud it would be better to just add the tests directly to t/spec 03:42
we're trying to migrate things out of the non-spec directory, I think (unless they're very pugs-specific)
oh, I see, there's no S04-statements/for.t
cjfields ok, will do.
pmichaud so, what we really want to have happen is for t/statements/for.t to become t/spec/S04-statements/for.t 03:43
03:43 shamu joined, shamu left 03:45 shamu joined
cjfields I'll try to run a conversion over for for.t (I have already added in tests for this particular case). 03:46
03:47 cjfields left 03:54 Ademan joined 04:26 masak joined 04:27 Psyche^ joined 04:28 tetragon joined 05:07 Theory joined 05:39 Theory joined 05:47 Ademan joined 05:50 Psyche^ joined 06:11 magnachef joined 06:17 base joined
dalek r28932 | moritz++ | trunk: 06:55
: [rakudo] fix infix:<x> for negative repetition count, patch curtesy of
: bacek++. Closes RT #56516
diff: www.parrotvm.org/svn/parrot/revision?rev=28932
r28933 | moritz++ | trunk: 06:57
: [rakudo] adding two more files to spectest_regression
: Auzon++ for fudging and reporting
diff: www.parrotvm.org/svn/parrot/revision?rev=28933
07:07 Psyche^ joined 07:15 Ademan joined
dalek r28934 | moritz++ | trunk: 07:17
: [rakudo] added S29-str/lc.t to spectest_regression
diff: www.parrotvm.org/svn/parrot/revision?rev=28934
07:26 nopaste joined 07:36 particle1 joined
dalek r28935 | bernhard++ | trunk: 08:11
: [languages/json]
: The content of json.pir and say.pir was duplicated.
diff: www.parrotvm.org/svn/parrot/revision?rev=28935
08:13 barney joined
dalek r28936 | bernhard++ | trunk: 08:13
: [languages/json]
: Set SVN properties for new files.
diff: www.parrotvm.org/svn/parrot/revision?rev=28936
r28937 | bernhard++ | trunk: 08:24
: [languages/json]
: Add copyright statements.
diff: www.parrotvm.org/svn/parrot/revision?rev=28937
r28938 | bernhard++ | trunk: 08:26
: [Plumhead(Pipp)]
: svn merge -r 28889:28934 svn.perl.org/parrot/branches/libs4...s/plumhead
: Disable PCRE until there is a check whether PCRE is available
diff: www.parrotvm.org/svn/parrot/revision?rev=28938
chromatic And there's the lexical fix.
dalek r28939 | chromatic++ | trunk: 08:33
: [PMC] Made Closure PMC respect enclosing runtime-created Closures when
: initializing lexpads (see RT #56184). In short, the simple pointer address 08:34
: comparison to find an enclosing outer Sub didn't take into account the fact
: that Closures created with the newclosure op, for example, have different
: pointer addresses than the constant Closures frozen into the bytecode constant
: tables.
: A better long-term fix is to give every :outer sub a unique lexid and compare
: those directly, but this will work until someone gets around to that bigger
: refactoring.
: Also added a test from the ticket (Patrick Michaud).
diff: www.parrotvm.org/svn/parrot/revision?rev=28939
moritz chromatic++ 08:35
bacek_ wow 08:36
chromatic++
dalek r28940 | bernhard++ | trunk:
: [languages/json]
: Tell config/gen/languages.pm and config/gen/makefiles/languages.in about json
: The content in languages/json/config/makefiles/root.in was duplicated.
diff: www.parrotvm.org/svn/parrot/revision?rev=28940
chromatic Just don't think of me as "The guy who fixes lexicals."
I already get that with GC and IMCC.
moritz ;-) 08:37
bacek_ :) summon jonathan
moritz what about "the guy who fixes ugly low-level guts"? ;-)
chromatic I made this one a little uglier, I fear.
bacek_ 4 lines 'if'... 08:40
dalek r28941 | bernhard++ | trunk: 08:48
: [languages/json]
: Update CREDITS and languages/LANGUAGES_STATUS.pod.
: Add languages/json/MAINTAINER.
diff: www.parrotvm.org/svn/parrot/revision?rev=28941
r28942 | chromatic++ | trunk: 08:59
: [src] Tidied code slightly; no functional changes.
diff: www.parrotvm.org/svn/parrot/revision?rev=28942
09:08 Whiteknight joined 09:35 rdice joined
dalek r28943 | kjs++ | trunk: 09:58
: [languages/json] actions and grammar files contents were duplicated, this is now fixed.
diff: www.parrotvm.org/svn/parrot/revision?rev=28943
r28944 | moritz++ | trunk: 10:35
: [rakudo] add tools/autounfudge.pl
diff: www.parrotvm.org/svn/parrot/revision?rev=28944
10:46 tetragon joined 10:53 kid51 joined 11:21 nopaste joined 11:50 tetragon joined 12:39 barney joined
dalek r28945 | bernhard++ | trunk: 12:48
: [languages/json]
: Tell SVN about generated files.
diff: www.parrotvm.org/svn/parrot/revision?rev=28945
12:51 iblechbot joined
dalek r28946 | bernhard++ | trunk: 12:52
: [languages/json]
: Tidy some more files with duplicated content.
diff: www.parrotvm.org/svn/parrot/revision?rev=28946
barney is starting with renaming Plumhead to Pipp 12:53
12:55 kj joined
dalek r28947 | bernhard++ | trunk: 12:57
: Let SVN ignore the generated file parrot_test_run.tar.gz
diff: www.parrotvm.org/svn/parrot/revision?rev=28947
barney kj: I'm still getting failures with languages/json under Linux 12:58
kj barney: hi
purl privet, kj.
kj well the only reason I didn't commit languages/json (the patch) myself was that I wasn't sure whether it would be useful; I just didn't want to loose the source 12:59
but I saw it was committed probably to reduce the # of open tickets :-)
I'll have a look at it
barney Yes, In a way it makes sense, in other ways it is redundant 13:00
It would be interesting to have a benchmark between PCT and handrolled parsing 13:01
moritz t/00-sanity.t....end of file expected at line 1, near "# This jus" 13:02
kj barney: you mean make test fails?
that's because I didn't include any tests...
dalek r28948 | bernhard++ | trunk:
: Fix json.realclean in languages/Makefile
diff: www.parrotvm.org/svn/parrot/revision?rev=28948
kj I'll fix it with a simple change; other tests can be added later
mm problem is that json is not a language with statements 13:04
so no print functionality
13:04 petdance joined
particle1 kj: didn't you use tools/dev/mk_language_shell.pl to create json? 13:04
kj yes
yes i did
particle ok, but you didn't update the tests
kj that's right
barney Ah t/00-santity.t is a stub generated by tools/dev/mk_language_shell.pl
particle sorry, i meant to respond to that ticket that the patch file was duplicated 13:05
kj I didn't expect the patch to be applied
particle oh, yuck
kj the duplication in patches is something I experienced before. Not sure why that happens.
particle i don't think that was clear in the ticket... not expected to be applied
kj well, I sent it in for a review 13:06
dalek r28949 | bernhard++ | trunk:
: Clean up parrot_test_run.tar.gz on 'make clean'
diff: www.parrotvm.org/svn/parrot/revision?rev=28949
kj and whether it could be used to replace the hand-written one
because that one contains A LOT of code
and this one is minimal
but you're right. The ticket wasn't clear.
particle whoever applied it should have done a better job of analysis 13:07
kj well I don't want to blame anyone; I
particle because barney has had to do quite a bit of cleanup
kj I'm sure the person who did meant well
particle sure, understood 13:08
most people here mean well :)
kj I would think s/most/all/
particle but we can all learn something, too :)
kj true. So my intention was to put this up for review, and maybe people would be interested in replacing the manually written one with this, at some point. 13:09
I'd sent it to parrot-porters, but somehow it never arrived, AFAICT
particle i got it 13:10
that's why i said "meant to respond"
barney That's still a good plan. Let it mature in 'languages' and eventually move it over to 'json'
s/'json'/'compilers'/ 13:11
kj there's currently only a json->parrot translator, not a parrot->json translator. 13:12
where 'parrot' means array/hash built-in pmcs 13:13
13:14 Ademan joined
particle yes, right 13:17
wait, didn't i start that? 13:18
13:18 gryphon joined
particle runtime/parrot/library/JSON.pir 13:18
kj oh I guess that's the one used by compilers/json 13:19
particle oh, right, coke did that
i did yaml
kj so array/hash->json is not actually in compilers/json. that's good
particle ...four letter acronyms... all look alike...
we need a yaml parser 13:20
parser/dumper pair, actually
kj anyway, i don't mean to throw away compilers/json just like that. I just was bored (without internet at home)
particle so we can serialize past, post, etc
NotFound Did we have a pmc or a class for json objects?
kj json objects are just arrays and hashes 13:21
[] -> array, {} -> hash
NotFound Yeah, but hashes actually can contain other things. Json objects are supposed to contains only other json objects. 13:22
kj so you're implying that hashes are not failsafe in the to-json operation? 13:23
not proper json-able
13:24 paco joined
DietCoke there's a question as to whether compilers/json should even be in here. =-) 13:24
I did _that_ because I was bored one day, it wasn't a requirement for anything. 13:25
NotFound And that we can't take any pmc and have some way granted that it only contains json data.
DietCoke we can just do our best, ja. good luck putting a closure into JSON.
NotFound DietCoke: not a requirement, but is nice to have a json way available when one needs a way to store/retrieve structured data. 13:26
kj DietCoke: it seems like a reasonable thing to do, converting stuff into JSON
so next is YAML and maybe XML :-)
particle somebody's gotta do xml, but it won't be me
NotFound I will vote for one way in core, and the others in modules.
barney kj: you forgot S-Expressions 13:27
kj that's the stuff with parentheses, right?
(key value) 13:28
barney yes 13:29
13:30 rurban joined
kj sounds to me there should be some common class for to_***() functionality, that can take parameters for the specific output; json taking [] and {,:,}; s-expressions only (), etc. 13:31
rurban Hi :) I'm just making up the new parrot packages for cygwin. I wonder how others named the rakudo or perl6 package? I want to name it parrot-perl6.
kj and xml using <>, </>
rurban I found nothing useful at debian, gentoo and freebsd ports.
particle how about parrot-languages-perl6?
or parrot-languages-rakudo
NotFound What about just 'rakudo'? 13:32
particle then we can distinguish parrot-languages from parrot-library etc
rurban Isn't "languages-" superfluous?
particle i'd like to see a rakudo package that depends on a parrot package
rurban I have parrot, libparrot0, libparrt-devel, and then parrot-perl6, parrot-lua, ...
NotFound rakudo is supposed to be perl6 in parrot, so the parrot part is redundant. 13:33
particle there can be more than one perl6 on parrot
rurban For pugs I need first ghc and this is a major work todo, so the simple parrot languages first.
NotFound particle: but we won't to call it rakudo.
rurban But I want pugs as reference also
particle NotFound: correct
rurban So far I build --languages="perl6 parrot_compiler punie lua" 13:34
particle languages/parrot_compiler isn't something that should be packaged
rurban ok.
particle it should probably be removed from the repo
kj i was just wondering, is parrot_compiler still used somewhere? 13:35
moritz rurban: pugs is currently packaged via cabal, the haskell package maanger
rurban just removed it from the ports file
what are the other most stable langauges that deserve a seperate package? 13:36
barney It's not used anywhere. Some code could be salvaged into an example.
NotFound Giving the interest recently seen, lolcode is a good candidate.
kj squaak is stable :-) (but not that useful I guess)
barney Pipp
Eclectus maybe 13:37
kj cardinal is improving I think?
barney m4 and hq9plus and urm defintily not
13:37 Whiteknight joined
dalek r28950 | fperrad++ | libs4php: 13:37
: [php] cache with PCRE
diff: www.parrotvm.org/svn/parrot/revision?rev=28950
13:38 TimToady joined
particle stable: abc cardinal eclectus lolcode lua perl6 punie pynie squaak 13:38
dalek r28951 | Whiteknight++ | gsoc_pdd09:
: a few diagnostic fixes and changes
diff: www.parrotvm.org/svn/parrot/revision?rev=28951
r28952 | bernhard++ | trunk: 13:39
: [Pipp]
: First round of renaming Plumhead to Pipp.
diff: www.parrotvm.org/svn/parrot/revision?rev=28952
rurban in the smokes there are only a few 100% ok for cygwin: unlambda, regex, lazy-k, bf. That's all
NotFound rurban: don't lose the opportunity, you can be the first person in the world to make a Pipp package.
13:40 vhold joined
rurban I definitely DO NOT want to be the first :) 13:40
NotFound Coward ;) 13:41
particle ok, now pipp is approaching stability, since it exists
a minute ago it was plumhead
NotFound plays 'A star is born'
rurban pipp: is that php? 13:42
particle rurban: if it's a hard requirement for all tests to pass in order for packages to be made/supported, then we have some work to do for sure
yes
pipp is parrot's php
rurban ok, pipp is a candidate also then.
particle lua should be a candidate 13:43
NotFound There will not be a new name for Lua in parrot? What about the winged- idea?
rurban lang_stable="abc cardinal eclectus lolcode lua perl6 punie pynie squaak pipp" 13:44
lang_alltests_ok="unlambda regex lazy-k bf"
particle i don't understand why perl6 tests are failing 13:45
rurban lang_alltests_ok="unlambda regex ook lazy-k bf"
particle make test on languages/perl6 is very simple
rurban oops, sorry. perl6 is not my languages-smoke. I forgot it
particle make sure all the stable langs are in languages-smoke 13:46
it may be, however, that parrot's languages-smoke can't deal with certain languages' test suites
rurban If I remember the perl6 tests run only with make fulltest 13:47
particle oh god no 13:48
that won't pass for a year or so
make test
or make spectest_regression
but fulltest includes all of the official perl6 spec tests
rurban fulltest is not so bad:
moritz particle: fulltest is fine 13:49
particle we're only passing ~25% of the 20% that's been converted
moritz particle: fulltest is spectest_regression + codetests + sanity-tests
particle: 'spectest' is bad
particle i must have an older makefile then
rurban t/codingstd/c_function_docs (Wstat: 256 Tests: 1 Failed: 1)
t/codingstd/c_macro_args (Wstat: 256 Tests: 1 Failed: 1)
t/codingstd/fixme (Wstat: 256 Tests: 1 Failed: 1)
t/codingstd/pod_todo (Wstat: 256 Tests: 1 Failed: 1)
and a few other examples failures 13:50
moritz rurban: in the last release? or svn HEAD?
rurban svn HEAD 13:51
from my dailiy smokes
dalek r28953 | Whiteknight++ | gsoc_pdd09:
: [gsoc_pdd09] updating to trunk 28949
diff: www.parrotvm.org/svn/parrot/revision?rev=28953
rurban Now I only have to try out building in a seperate builddir without lndirs 13:52
particle ah, yes, fulltest is reasonable now
moritz rurban: the condingstd tests in languages/perl6/ should be the same as in the main parrot dir -do they fail there as well?
NotFound The codingstd tests checks for platform EOL in sources? 13:54
rurban moritz: I obviously never did make test -C languages/perl6 13:59
moritz rurban: ok 14:00
14:05 purl joined
tewk_ P{} 14:07
barney Yes languages-smoke only works for langs with Perl5 test scripts.
rurban My current languages-smoke on cygwin is not good: 280 OK from 2722 tests (10.29% ok) 14:15
perl6 fails with No rule to make target `/usr/src/perl/parrot/parrot-svn/runtime/parrot/library/PGE/Perl6Grammar.pbc', needed by `src/gen_grammar.pir'. Stop
moritz that's bad, that should be created during 'make' in top level parrot 14:16
rurban: care to submit a bug report?
parrotbug?
purl parrotbug is mailto:parrotbug@parrotcode.org or svn.perl.org/parrot/trunk/docs/submissions.pod or see also "rakudobug"
rurban first I fix my 0.6.3 release. then I look into the new bugs. 14:17
moritz ok
rurban the new glut bindings broke a lot of things
14:21 iblechbot joined
rurban Configuring languages... step gen::languages died during execution: Can't open languages/pipp/config/make files/root.in: No such file or directory at lib/Parrot/Configure/Compiler.pm line 266. 14:29
oops, sorry! I mixed up 0.6.3 with svn HEAD
DietCoke (what, no tcl on cygwin? :P) 14:35
barney rurban: For svn HEAD you might need to do a 'make realclean' first
DietCoke I am surprised that tcl is failing any tests. there might be one warning with the lsort test, but if you squint, all the test did pass. =-) 14:36
14:42 stupidbot joined
dalek r28954 | bernhard++ | trunk: 14:46
: [Pipp]
: Second round of renaming Plumhead to Pipp.
diff: www.parrotvm.org/svn/parrot/revision?rev=28954
barney BTW: There is now a simple languages-test-separate Makefile target
14:47 TonyC joined 14:52 gmansi_ joined 14:53 bacek joined
dalek r28955 | bernhard++ | trunk: 14:54
: [Pipp]
: Get rid of some TGE related code.
diff: www.parrotvm.org/svn/parrot/revision?rev=28955
14:55 peepsalot joined
dalek r28956 | bernhard++ | trunk: 14:58
: [Pipp]
: The files driver_nqp.pbc and pipp_pct.pbc are no longer generated. 14:59
diff: www.parrotvm.org/svn/parrot/revision?rev=28956
15:01 ambs joined
dalek r28957 | bernhard++ | trunk: 15:03
: [Pipp]
: Readd the *.java files generated for the Antlr3 variant,
: after the renaming from Plumhead to Pipp
diff: www.parrotvm.org/svn/parrot/revision?rev=28957
r28958 | bernhard++ | trunk: 15:08
: [Pipp]
: Clean up the files benerated during 'make maintain-antlr3'
diff: www.parrotvm.org/svn/parrot/revision?rev=28958
bernhard.schmalhofer@gmx.de | Plumhead: 15:10
link: www.perlfoundation.org/parrot/index.cgi?plumhead
bernhard.schmalhofer@gmx.de | Pipp:
link: www.perlfoundation.org/parrot/index.cgi?pipp
bernhard.schmalhofer@gmx.de | Plumhead: 15:13
link: www.perlfoundation.org/parrot/index.cgi?plumhead
bernhard.schmalhofer@gmx.de | Pipp:
link: www.perlfoundation.org/parrot/index.cgi?pipp
bernhard.schmalhofer@gmx.de | plumhead_renaming: 15:17
link: www.perlfoundation.org/parrot/index...d_renaming
15:18 Theory joined
cotto_home barney, I'd recommend that we not change names on the wiki (or in the code) until pipp.org is actually transferred. 15:26
There's no indication that anything will go wrong, but I don't like jumping the gun. 15:27
barney Sorry, it's already done. 15:28
cotto_home It won't hurt anything. I'd leave the code alone, for now
ambs kind of OT question: anybody with firefox3 and problems editing pages on socialtext? 15:29
cotto_home also, we'll have to coordinate with fperrad so his code doesn't get out of sync
DietCoke cotto_home: it would be nice if we had someone else on cygwin tell us they had a successful build.
cotto_home DietCoke, wrt which ticket
? 15:30
DietCoke the one you just closed. 42692
barney cotto_home: fperrards changes are merged into trunk, but there are some extra changes 15:31
cotto_home barney, ok. 15:34
DietCoke, I closed that one because of apparent lack of interest, but I plan on getting a cygwin build environment set up soon, at which point I can test it myself. 15:35
rurban firefox3 has definitely problems with socialtext where you mention it. 15:40
I tried the search box and nothing happened when I was in edit mode.
ambs rurban, good its not just me
rurban I am on Win XP sp2
ambs starts safari to edit TPF website
rurban, macos
rurban SeaMonkey 1.1.9 also has pronblems with socialtext 15:41
cotto_home barney, it looks like pipp.org will be transferred later today 15:44
15:49 rdice joined
barney cool 15:49
use.perl.org/~Bernhard/journal/36838 16:02
dalek r28959 | cotto++ | trunk: 16:03
: [pipp] fix some PMC-related targets
diff: www.parrotvm.org/svn/parrot/revision?rev=28959
16:03 sjansen joined 16:08 ruoso joined
barney cotto: clean-pmc is already in clean-common 16:10
cotto_home barney, my mistake 16:11
I'll take it out. 16:12
(out of clean)
dalek r28960 | pmichaud++ | trunk:
: [rakudo]:
: * Update spectest-progress.csv: 80 files, 1290 passing tests
diff: www.parrotvm.org/svn/parrot/revision?rev=28960
r28961 | cotto++ | trunk: 16:15
: [pipp] backing out an unnecessary dependency in the makefile
diff: www.parrotvm.org/svn/parrot/revision?rev=28961
pmichaud www.pmichaud.com/perl6/rakudo-tests...-07-02.png 16:16
16:18 mj41 joined
cotto_home barney, my code is all up-to-date. Thanks for taking care of the name change. 16:18
barney++
Auzon pmichaud: Does that mean that we added 150 tests were added since the last data point? 16:20
That's quite impressive if so :)
pmichaud on 07-01 there were five new test files added to spectest_regression, and a total of 183 new tests 16:21
Auzon I see the file now 16:22
pmichaud on 07-01 we also passed 118 more tests
because of chromatic++ 's lexical bug fix, I'm hoping I can get $_ and .method working today
(but kids keep interrupting around here ) 16:23
kj can't you learn the kids to program? :-)
s/learn/teach/
cotto_home "Daddy, what's an incremental garbage collector?" 16:24
pmichaud "Our dog is a good example."
kj ha ha 16:25
cotto_home pmichaud++
pmichaud "or your rooms."
(I'll avoid the obvious comments about "Daddy's desk" :-)
kj i think there's a leak in there too :-)
NotFound kj: let them play with a good old line-numbered Basic Interpreter.
kj is it running on parrot? 16:26
NotFound Not sure, but I think I readed that the basic in the repo accepts line numbers.
kj it seems thecompiler works, but the generated code is not 16:27
s/is/does/ 16:28
NotFound Anyway, IMO a true old-style basic must be a pure interpreter. 16:30
kj does it matter what the interpreter does internally?
that is, it needs to parse the input anyway, and it might as well generate PIR 16:31
NotFound kj: yes, if you want to play old games like auto-modifiable code.
kj mmm ok. I don't have experience playing old games
.. in basic
NotFound kj: if you want to try, here is my interpreter: blassic.org 16:32
kj nice! don't you feel the urge for a parrot port? :-) 16:33
Infinoid *urge* *urge*
NotFound Maybe someday I try with a subset. 16:34
Or maybe I try to embed parrot inside it. 16:35
kj mmm. ecmascript seems to fail tests :-( 16:53
it was working last time I touched it, iirc
16:55 magnachef joined 16:57 magnachef_ joined
NotFound kj: fails also for me. 16:59
kj the grammar seems fine. 17:00
17:06 rurban_ joined
NotFound Pipp build with C++ pass all tests. Good work, guys. 17:07
17:18 cjfields joined
NotFound A doubt about the imcc_init function: several places call it, and docs/embed.pod claims it is required to be able to use the compiler. However, it is called from make_interpreter, so it must be always already initialized, it isn't? 17:24
kj there seems to be something wrong with \\N in ecmascript's grammar. Anybody knows about any changes in PGE w.r.t. \\N? 17:33
pmichaud havent been any changes for \\N that I'm aware of. 17:35
kj ok
NotFound Dropped all calls to imcc_init in t/src/compiler.t and all test still passes.
DietCoke had fun with his eldest (catching up on scrollback) yesterday giving him a puzzler from car talk, letting him figure it out, then showing him how to do it in perl, then showing him how much faster it was in C. 17:37
japhb How old is your eldest? 17:38
DietCoke 11.5 17:40
NotFound DietCoke: are you premature optimizing your childs?
dalek r28962 | pmichaud++ | trunk: 17:45
: [pct]:
: * PCT now generates lexical scopes via :lexid.
diff: www.parrotvm.org/svn/parrot/revision?rev=28962
r28963 | kjs++ | trunk:
: [ecmascript] change throw and return into tokens using \\N; tests pass again.
diff: www.parrotvm.org/svn/parrot/revision?rev=28963
NotFound kj: pass here also. 17:48
kj NotFound: thanks for checking 17:50
I suspect it has something to do with the <ws> rule. 17:51
18:03 rdice joined
Whiteknight DietCoke still here? 18:21
DietCoke yes 18:25
and still 18:32
Whiteknight I want to create a branch to work on that :vtable has "Self" problem. you dig? 18:34
dalek r28964 | pmichaud++ | trunk: 18:35
: [pct]:
: PAST::Compiler no longer calls newclosure for nested blocks --
: it relies on the PAST tree to make newclosure op nodes
: as appropriate. (This part of the design may change yet
: again, but we'll go with this for now as it appears to work.)
diff: www.parrotvm.org/svn/parrot/revision?rev=28964
Whiteknight the problems are sufficiently far-reaching that I think a separate branch is in order
pmichaud chromatic++ # fixing RT #56184 18:36
stupidbot Error calling said() for rt: Malformed RT response received from rt.perl.org/rt3/
cjfields stupid stupidbot (is that redundant?)
pmichaud purl, stupidbot? 18:37
purl no idea, pmichaud
pmichaud stupidbot is still stupid.
Whiteknight stupidbot? 18:38
purl stupidbot is still stupid.
Whiteknight purl++
cjfields stupidbot is also purl's red-headed stepchild 18:42
purl okay, cjfields.
Infinoid opbots, names 18:46
dalek r28965 | fperrad++ | libs4php: 19:02
: [php] move var_dump() tests
diff: www.parrotvm.org/svn/parrot/revision?rev=28965
r28966 | fperrad++ | libs4php:
: [php] implement count()
diff: www.parrotvm.org/svn/parrot/revision?rev=28966
19:04 ruoso joined
nopaste "pmichaud" at 76.183.97.54 pasted "patch results in infinite loop in imcc" (58 lines) at nopaste.snit.ch/13451 19:29
pmichaud can someone apply the nopaste patch to trunk and see if they can get perl6.pbc to build? 19:30
on my system it hangs at the ../../parrot -o perl6.pbc perl6.pir
if I comment out the "push_eh" instruction it compiles cleanly 19:31
pmichaud tries a fresh checkout 19:32
dalek r28967 | fperrad++ | trunk: 19:40
: [php] move var_dump() tests
diff: www.parrotvm.org/svn/parrot/revision?rev=28967
pmichaud okay, I get the infinite loop in a fresh checkout, also.
if anyone else can confirm the problem I'd really appreciate it.
Whiteknight what's the problem? stall on make perl6? 19:42
pmichaud stalls on compiling perl6.pbc
Whiteknight okay, let me fire it up 19:43
dalek r28968 | fperrad++ | trunk:
: [Pipp] implement count()
diff: www.parrotvm.org/svn/parrot/revision?rev=28968
pmichaud purl, 422/943? 19:50
purl pmichaud: i don't know
pmichaud 422 / 943
purl 0.447507953340403
Whiteknight mine built 19:53
update, clean, build, worked
pmichaud Whiteknight: platform?
Whiteknight: you applied the patch I nopasted?
Whiteknight no, no patch
pmichaud oh, I know it builds from clean.
Whiteknight Debian, ubuntu 8.04
oh, so we were testing the patch?
pmichaud yes, I'm trying to see if that patch causes anyone else's imcc to hang
Whiteknight ...I'll try that then 19:54
pmichaud thanks
I greatly appreciate it
dalek r28969 | fperrad++ | trunk: 19:57
: [Pipp] cache with PCRE
diff: www.parrotvm.org/svn/parrot/revision?rev=28969
NotFound pmichaud: if you are working with imcc issues, please take a look at the patch I've sended now, RT#56538 19:58
stupidbot Error calling said() for rt: Malformed RT response received from rt.perl.org/rt3/
pmichaud NotFound: I'm not working actively with imcc internals -- I'm just very surprised to come across something where I can make it "hang" 19:59
it's just compiling a (very lengthy) pir file, but it really should never hang
NotFound This is more an external than an internal ;)
pmichaud phone 20:00
dalek r28970 | fperrad++ | trunk: 20:01
: [Pipp] try to work without PCRE
diff: www.parrotvm.org/svn/parrot/revision?rev=28970
Tene Should 'make fulltest' work? It seems to be looping. 20:10
I've seen t/dynpmc/foo fail five times so far, I think.
pmichaud Parrot's 'make fulltest' tests all of the runcores iirc
Whiteknight yes, i'm pretty sure that's what it does 20:11
it's a long process
dalek r28971 | fperrad++ | trunk: 20:14
: [Pipp] MANIFEST
diff: www.parrotvm.org/svn/parrot/revision?rev=28971
Tene ahh 20:17
dalek r28972 | fperrad++ | trunk: 20:21
: [Pipp] svn properties
diff: www.parrotvm.org/svn/parrot/revision?rev=28972
Infinoid opbots, trust NotFound 20:29
slavorg Ok
clunker3 But I do not trust you Infinoid
slavorgn Ok
Infinoid (seems human.)
opbots, believe dalek
slavorg Ok
clunker3 But I do not trust you Infinoid
slavorgn Ok
Infinoid (doesn't seem human.)
NotFound Are you playing robot games? 20:30
Infinoid <dalek> exterminate all humans.
who belongs to clunker3? 20:31
NotFound All your robots are belong to us! 20:41
There is an announce of pipp in planet parrot. The plan was not to delay announces until domain transfer? 20:57
21:01 shamu left 21:04 gryphon joined
dalek r28973 | chromatic++ | trunk: 21:29
: [IMCC] Removed unnecessary imcc_init() calls; interpreter initialization
: already does this (NotFound, RT #56538).
diff: www.parrotvm.org/svn/parrot/revision?rev=28973
stupidbot Error calling said() for rt: Malformed RT response received from rt.perl.org/rt3/
21:36 slightlyoff joined 21:37 slightlyoff left 21:45 avar left
rurban just comitted a report about sane installation defaults 21:51
with the default options it will pollute the system 21:52
(with patch of course)
disassembler should be renamed to pbc_disassembler 21:53
sorry: disassemble => pbc_disassemble 21:54
The generated Makefile has also an insane LIB_DIR=/usr/lib default 21:55
Should be /usr/lib/parrot 21:56
japhb rurban: Note that disassemble is used by tools/util/dump_pbc.pl
rurban That's why I have no patch for this yet :)
japhb rurban: Ah ... you said "with patch of course"
rurban patch just for tools/dev/install_files.pl for now. 21:57
NotFound rurban: The lib dir thing is debatable.
rurban For the Make system and names I want to hear opinions
make reallyinstall intalls lots of mist into /usr/lib 21:58
Tene pmichaud: ping
NotFound rurban: but at leaste the main dynamic library is preferably that be installed in a standard location. 21:59
rurban /usr/lib/config, /usr/lib/compiler, ...
japhb NotFound: my thinking is that .so files should go in /usr/lib, but runtime/... et al. needs to go under /usr/lib/parrot/
purl Hmm. No matches for that, japhb.
rurban and the worst is the default main/library dest to prefix only. This should go to some /usr/lib/parrot analog to perl5 22:00
NotFound Using other place can cause problems, like having to touch ld config files.
japhb rurban: nod
rurban japhb: true
22:00 teknomunk joined
rurban of course the main lib is in the standard location. 22:01
NotFound Fine, then.
Maybe the disassembler can be dropped, by adding an option to do the work in pdb. 22:03
japhb NotFound: see my comment above -- it's needed by dump_pbc.pl 22:04
rurban I'll send a disassemble rename patch then. I don't want to package just a /usr/bin/disassemble
japhb agreed
NotFound Too generic name, yes. 22:05
(I hope the debate will be short than plumhead renaming) X-) 22:06
japhb NotFound: thankfully, we don't need to buy a domain for it ... 22:08
NotFound disassemblers_anonymous.org 22:09
rurban btw: how to submit a patch as attachment with parrotbug? 22:10
inline it :) yes
japhb Or just send it as a normal email, rather than using the tool
NotFound There is an attach button somewhere. 22:11
japhb NotFound: in RT, yes. I thought rurban was talking about the parrotbug command line tool
rurban yes, my joe has the ^KR button
japhb cringes
NotFound Ah, yes, I don't used the command line yet. 22:12
I have a copy of Wordstar for cp/m that has the same button.
japhb NotFound: developers rarely do. I believe it broke at one point and noone noticed until a user tried to parrotbug a release. I had problems with it once (that turned out to be a local MTA issue), and people went "Oh, does that work now?" 22:13
NotFound: yes, but do you have a box still running CP/M on which to run said Wordstar?
22:13 mire joined
NotFound japhb: something like that. 22:13
Several virtual boxes that can run it. 22:14
japhb ... one of which is running on a machine that can use 8" diskettes? :-)
NotFound And I have also an Amstrad CPC 6128, but no disks for it.
japhb: not, just 3" ones. 22:15
japhb still has old CP/M, CBASIC, SuperCalc, and SuperWriter 8" disks laying around, for historical value.
NotFound japhb: take a look at www.arrakis.es/~ninsesabe/aliados/ 22:16
rurban I have a box with seymonkey mail, which actually has such an Attach button :) 22:17
japhb NotFound: did you create that? 22:18
NotFound japhb: yes.
japhb cool! 22:19
NotFound The system only, the Z80 emulation core is adopted.
japhb coincidently has been reading about creating an assembler for Z80 recently 22:20
NotFound japhb: I already wrote one.
www.arrakis.es/~ninsesabe/pasmo/ 22:21
japhb "can also generate 8086 code from Z80 sources" -- now that's what I call cross-assembly! 22:23
NotFound japhb: yeah, and it can even run it in ms-dos.
japhb heh
Retro is just so gosh-durned fun. (says the forthwright) 22:24
NotFound The funny thing is that I started to write it as a tool to create test for the cp/m thing, and become much more popular than it.
japhb interesting 22:25
NotFound Well, for some value of 'popular' :D
In certain circles of people that still write games and demos for old machines. 22:26
rurban hmm. I wonder if /usr/bin/pdb is not also too generic? Isn't pdb already taken? 22:41
NotFound Python debugger? 22:42
rurban yup. pyton has it. oh my
Thanksfully this has no deps
NotFound Don't know if it's just a module or a standalone also.
rurban I just googled for /usr/bin/pdb and it's there for python 22:43
NotFound There is a chess Problem Database Server also.
rurban (debian etch)
pdb => parrot_pdb? 22:44
parrot_db sounds too much like a database
NotFound parrot_deb looks like a debian package. 22:45
parrot_debug?
rurban (why I am the first anyway? is there no debian package for the parrot utils yet?)
japhb rurban: Because we have not previously been in the phase in which installing needs to work. 22:46
NotFound There is even a lot of messages preventing people from installing. 22:47
japhb Basically, everyone would be happy if installing worked, but no one previously worked to hard on getting it right.
rurban maybe I should just wait for the weird debian guys to come up with a overarchitectured naming. or wait for a proper and stable fedora package
japhb There have been packages for a couple packaging systems, but I've never tried them.
And I suspect they are very old.
rurban yes, debian is very iold, had technical problems and wrong deps. And allison never answered to them. 22:51
Ah, I see. We wanted to change the debian mainainer to parrot-porters@. Good. So it will be the release manager. So it must be added to the release_manager_guide.pod 22:53
Tene rurban: the spec file seems to work just fine for me to build an rpm for fedora. I can probably build an rpm for you, depending on what arch/release you want it for. 22:55
rurban Tene: No thanks :) I rather make up my own specs. We call it ports, similar to freebsd and gentoo.
Actually the cygwin cygport is modelled after the gentoo ebuild / portage, which is modelled after the freebsd ports. 22:56
Is a debian-parrot already somewhere? I want to talk to them about packaging issues. The final proper layout, so to call. 22:58
fedora is also fine, even finer probably.
purl okay, rurban.
japhb purl, fedora? 22:59
purl rumour has it fedora is a style of hat, like the pulp-novel detectives used to wear or fine, even finer probably
japhb no, fedora is a style of hat, like the pulp-novel detectives used to wear
purl okay, japhb.
rurban rumor has it that debian packages are lawyers and managers and no programmers at all
:)
23:00 tetragon joined
rurban Hey; I just read that perl(Moose) is a fedora BuildRequires: I have no moose and parrot builds fine. 23:03
And I see that fedora also heavily patched the build process. But mainly the usual lib|lib64 issue. 23:04
And it's nice that they produce a pbc_exe binary for every single language 23:05
23:41 Coleoid left 23:59 TiMBuS joined

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!