www.parrot.org/ | 292 RTs left | Next release: 2009-07-21
Set by moderator on 13 July 2009.
jdv79 i think the article on sdtimes.com pushes the envelope of truth. 00:16
Infinoid "expands dynamic integrations" 00:17
somehow, I like the sound of that
I'd buy anything that has the word "dynamic" in it
dalek rrot: r40071 | pmichaud++ | trunk/compilers/nqp/src/Grammar.pg:
[nqp]: Eliminate deprecated <fail> subrule (TT #460).
 00:24
rrot: r40072 | pmichaud++ | trunk/compilers/pge/PGE (2 files):
[pge]: Remove some deprecated subrules and methods (TT #460)
pmichaud whoppix: pong
woops
whiteknight left.... tab completion fail
whoppix happens.
00:29 TiMBuS joined 00:46 japhb joined
kid51_at_dinner jdv79: Which article on sdtimes.com? 00:53
00:56 MoC joined
cotto kid51, sdtimes.com/PARROT_1_4_EXPANDS_DYNA...RROT/33606 01:01
alternately, sdtimes.com/PARROT_1_4_MAKES_PUPPIE...PIES/33606 01:03
Apparently their SEO guys are a little extreme. 01:04
kid51 Yes, indeed. Thanks. 01:11
01:23 chromatic joined
dalek rdinal: 23a0f62 | (Ted Reed)++ | src/classes/Array.pir:
Implement Array.concat
 01:33
rdinal: fee49aa | (Ted Reed)++ | src/classes/Array.pir:
Implement the * operator for Array.
01:36 patspam joined 01:46 Zak joined
dalek cnum-dynpmcs: r108 | darbelo++ | trunk/ (4 files):
A small step towards a cleaner build infrastructure:

   Moved intermediate object files to the 'build' directory
   Removed some now-unneeded targets.		 01:51
02:03 darbelo left 02:07 Andy joined 02:08 Khisanth joined 02:09 patspam joined 02:31 Andy joined 02:42 janus joined
Andy ping jettero 02:47
purl I can't find jettero in the DNS.
dalek rdinal: 6ac0177 | (Ted Reed)++ | src/classes/Array.pir:
Implement + operator for Array.
 02:57
rrot: r40073 | jkeenan++ | trunk/t/pmc/exception.t:
Un-SKIP test #30 (which actually should have been TODO-ed rather than SKIP-ped. Cf.: trac.parrot.org/parrot/ticket/40.
 02:58
moderator www.parrot.org/ | 291 RTs left | Next release: 2009-07-21 03:04
kid51 must sleep 03:04
purl $kid51->sleep(8 * 3600);
cotto bacek_at_work, poing 03:19
dalek kudo: 3e1952f | pmichaud++ | src/parser/ (2 files):
Update compiler to avoid deprecated regexes and Match methods.
 03:28
kudo: 147b3d7 | pmichaud++ | build/PARROT_REVISION:
Bump build/PARROT_REVISION to get latest PGE changes.
 03:29
rrot: r40074 | pmichaud++ | trunk/DEPRECATED.pod:
[pge]: Remove deprecated items from PGE (TT #460, #461).
 03:33
TT #460 closed by pmichaud++: [TODO] remove deprecated built-in PGE rules 03:36
bacek_at_work cotto: poing 03:40
cotto bacek_at_work, Is your intent for opsc that we add an extra stage to the compiler that runs the oplib code and the runcore-specific stuff, then spits it out to the appropriate files? 03:43
bacek_at_work cotto: I have no strong opinion on either approaches. 03:45
but I slightly prefer to have compiler and emitter separated. 03:46
So we can parse everything once and spit various runcores from single PAST tree
cotto That's more efficient. 03:47
bacek_at_work Yes. 03:48
And it can be fit into current Makefiles more easily.
cotto So should BUILD take care of the runcore-independent stuff, i.e. turning the past into something the runcores can work on? 03:49
bacek_at_work For now - yes.
But keep in mind that we'll probably will use OpLib for L1 too. 03:50
cotto That's an interesting concern. 03:51
bacek_at_work ETOOMANYWILL
cotto I suppose it wouldn't be too hard to stick the proper compiler in there somewhere.
?
bacek_at_work cotto: parse_ops_file is good candidate.
Actually we can prefix L1 based ops with "L1" to select PCT rules for op_body. 03:52
afk # $dayjob 03:53
cotto The thing that confuses me is that parse_ops_file is producing a PAST, which duplicates opsc_core.pir
bacek_at_work, ^ (whenever you get nafk) 03:57
bacek_at_work cotto: erm... What do you mean? 04:10
parse_ops_file _uses_ opsc_core to parse ops.
cotto right, but my assumption is that we'll run opsc.pbc to generate the runcores 04:11
i.e. ./parrot compilers/opsc/opsc.pbc src/ops/*.ops 04:13
I don't see how parse_ops fits in with that.
bacek_at_work opsc.pbc can mimic ops2pl. 04:14
e.g. not use "traditional" PCT compiler workflow.
"do not use" means "We'll create own flow". 04:15
cotto So what'll that flow look like?
bacek_at_work Just because currently we are not targeting parrot. We targeting multiply "C compiler" targets
cotto I was thinking we'd just add an extra "generate_runcore" stage to the HLLCompiler. 04:16
bacek_at_work 1. Read command line args. 2. Create few objects. 3. Emit runcore1. ... 7. emit runcore4
we have to generate multiple runcores. I don't now how it fits into HLLCompiler. 04:17
cotto We could generate them all at once.
We just need to make sure the past doesn't get messed up. 04:18
bacek_at_work And it's challenge.
Similar to "implement refcounting of Contexts"
cotto eew
All we'd need to do is stick the opcodes into something that can be cloned. 04:19
It'd be a bit of annoying boilerplate code, but I don't think it'd cause any problems.
bacek_at_work OpLib::ops is good candidate :)
It's Hash. And it deeply cloned by default.
cotto I thought cloning was shallow by default. 04:20
bacek_at_work not for hashes
cotto It copies strings, but it also just uses the default clone for any nested types 04:22
chromatic Did someone make a Git SVN clone? 04:23
bacek_at_work cotto: hm... May be.
cotto If we have an RPA or RSA in there, we'll get new PMCs whose values are old PMCs. 04:24
bacek_at_work chromatic: Jonathan Leto did.
chromatic I saw his GitHub repo.
dukeleto what did I do :) 04:25
github.com/leto/parrot/tree/upstream
chromatic I have commit access to Parrot SVN though.
Then I decided to check out only the last ~2000 revisions. 04:26
cotto s/values/elements/
bacek_at_work chromatic: git svn init -s svn.parrot.org/parrot; git svn fetch -r 38000 04:29
chromatic Yep, just did that. 04:30
Now to figure out how to work on an SVN branch. 04:31
dukeleto the full git svn clone took about a week
cotto protip: avoid using avian carriers 04:32
dukeleto cotto: like svn over http?
cotto bacek_at_work, do you have any objections to trying a "traditional" HLLCompiler workflow that generates all runcores at once? 04:36
bacek_at_work cotto: no. go ahead if you can make it I'll be happy 04:42
cotto Awesome.
04:43 mokurai joined
cotto ENONOPASTE 04:45
bacek_at_work, is there any reason not to run load_op_map_files from BUILD? 04:53
04:53 nopaste joined
bacek_at_work cotto: for simplify testing during development. It should be called from BUILD. 04:58
cotto ok. I'll move it there with the next commit (which should be later tonight). 05:01
pmichaud, ping 05:12
msg pmichaud In HLLCompiler, why is $?FILES a space-separated String instead of an RPA? 05:25
purl Message for pmichaud stored.
06:01 uniejo joined
dalek rrot: r40075 | chromatic++ | branches/pluggable_runcore:
Created branch to make runcores pluggable
 06:02
cotto woohoo! next stop, world domination through profiling! 06:06
06:07 hiroyuk__ joined 06:09 mokurai joined
chromatic Eventually. 06:19
purl eventually we'll all be dead
06:20 iblechbot joined
bacek_at_work chromatic: ping 06:21
06:22 TonyC joined
chromatic pong 06:38
bacek_at_work chromatic: My English is too weak to understand your mail about my GC patch. Is't thumb up for patch for thumb down? 06:45
lists.parrot.org/pipermail/parrot-d...02503.html this one 06:46
chromatic Thumbs up. 06:57
purl it has been said that thumbs up is at www.friedmanarchives.com/China/Web/...%20dpi.jpg
chromatic We should keep the constant STRING pool for now tgough.
though
bacek_at_work Ok. 07:00
I didn't touch pools actually. I just enforced "contract" with constant pools.
That's main reason for hate constant pools per se..
07:09 barney joined 07:11 szabgab joined 07:27 Zak joined
dalek rrot: r40076 | cotto++ | branches/ops_pct/compilers/opsc (3 files):
[opsc] start working on a generate_runcores stage to transform the past into C runcore code
 07:48
cotto Would it be a Bad Idea to use the svn id string as part of a grammar? 07:49
nm. It feels hacky. I'll find something else. 07:50
07:55 Zak joined 07:57 mikehh joined
nopaste "chromatic" at 72.90.115.31 pasted "Runcore struct draft (for cotto)" (13 lines) at nopaste.snit.ch/17256 08:07
08:12 Gerd joined
cotto which runcores would necessitate the flags? 08:23
would that just be C?
nm. I need to be more coherent to ask that question. 08:25
chromatic See src/call/pcc.c:3020 08:28
cotto chromatic, could I get your thoughts on the following nopaste? 08:31
08:31 masak joined
nopaste "cotto" at 74.61.2.46 pasted "proposed change to .ops file format" (266 lines) at nopaste.snit.ch/17257 08:31
cotto basically remove VERSION = PARROT_VERSION and make it explicit what's part of a ops file's preamble 08:32
08:33 mokurai left
bacek_at_work cotto: why do you want explicit BEGIN_OPS_PREAMBLE? 08:34
cotto to make it easier to parse all the ops files when they're catted together 08:35
bacek_at_work afaik currently everything before VERSION = ... is preamble.
Ah.
Got the point
cotto also, VERSION = ... is useless and obsolete 08:36
bacek_at_work But why do you need it?
I agree that VERSION is obsolete.
cotto it makes opsc easier to write
also, it's actually everything before the first pod that's preamble. 08:37
there's one file that exercises that
clock? 08:40
purl cotto: LAX: Tue 1:40am PDT / CHI: Tue 3:40am CDT / NYC: Tue 4:40am EDT / LON: Tue 9:40am BST / BER: Tue 10:40am CEST / IND: Tue 2:10pm IST / TOK: Tue 5:40pm JST / SYD: Tue 6:40pm EST /
bacek_at_work time to go home.
cotto time to go to sleep 08:42
good night
bacek_at_work good night, cotto
08:55 Casan joined
Casan jonathan: heya. Unfortunately I must skip yapc:eu this year. I have been accepted to masters studies in France starting in September, but my funding is still not in the clear, so I'll have to work on that instead. 08:58
09:34 donaldh joined 10:06 iblechbot joined
Infinoid msg chromatic Your comment on TT #678 didn't get through. We still need to get email2trac working... 10:18
purl Message for chromatic stored.
dalek kudo: 0c2cb6f | masak++ | src/ (2 files):
moved C<.split> POD from src/builtins to src/setting
 10:19
Infinoid Gerd: TT #509 looks like a communication failure to me. Allison mentions that symlinks will fail on platforms that don't support symlinking, but the patch explicitly checks for symlink support before using them. On the other side, it's very unclear what the benefit of this patch is 10:22
Gerd Infinoid: The benefit is only for building rpms 10:25
Infinoid: On the system library there are only symlink from noversionnumber library to version librarys 10:26
If the symlink is done with make from Parrot there is no need to take for it in the spec file of the rpm 10:28
Infinoid I agree with that assessment, but I think putting it in the specfile is cleaner overall 10:29
I know we're working on reorganizing the temporary blib tree structure to more accurately match the install path. But I'm not sure what the status of that is
I would note that your patch does not check whether an existing symlink is relative or absolute, so if we ever use them for that blib organization stuff, you will end up with an rpm containing symlinks to the build tree 10:30
But that's a strawman. I don't have a good example of that
(and I'm pretty unfamiliar with the installation effort in general) 10:31
Gerd Infinoid: The symlink is create first and get lost with install_files script 10:32
Infinoid Yes. Preserving that symlink probably wouldn't hurt. My only concern is that there's no guarantee that it will remain the *only* symlink 10:33
anyway, my main point of bringing this up was, explaining in further detail in comments to the ticket would increase its chances of being applied 10:34
Gerd Okay, so I will let the patch in the specfile
mikehh All tests PASS (pre/post config, smolder, fulltest) at r40076 - Ubuntu 9.04 amd64
Infinoid Gerd: I'll post a followup to the ticket, thanks 10:41
Gerd Infinoid: thanks 10:42
mikehh rakudo does FAILs to run even make test - t/00-parrot/01-literals.t .......... too many arguments passed (2) - 1 param expected 10:43
this happens for all tests
Infinoid too many arguments to what? 10:44
mikehh dunno where the does comes from - I don't remember typng that
typing 10:45
i did make test in rakudo andthat was the output o the first nd subsequent tests 10:46
wireless keyboard :-{
Infinoid what happens when you run one of those tests with "prove -v"? 10:47
Oh, wait. 10:49
Gerd: If a symlink was detected, won't this code run both symlink() and copy() on the files? An if/else chain might be more appropriate
mikehh mhh@mhh-desktop:~/rakudo$ perl t/harness t/00-parrot/01-literals.t 10:51
t/00-parrot/01-literals.t .. too many arguments passed (2) - 1 param expected 10:52
I built and tested parrot in ~/parrot and then switched to ~/rakudo 10:55
Gerd Infinoid: I do not remember the code, but I think not because of an next statemant
mikehh make realclean, git pull and then perl Configure.pl --parrot-config=../parrot/parrot_config, make, make test 10:56
Gerd Infinoid: Why be operating system can be checked, if it is Linux and only to the protect the syslink on linux 10:57
Thanks Infinoid, I have to leave irc, because I have to look for another thing, bye 11:05
Infinoid oh, there is a next statement, I missed that. 11:09
11:21 donaldh joined 12:04 Whiteknight joined 12:06 MoC joined
Whiteknight good morning #parrot 12:21
jdv79 are there any examples of walking the arenas? 12:32
Whiteknight jdv79: what do you mean by that? 12:43
Arenas form a linked list 12:45
jdv79 i'm looking into generating an object graph like Coke hinted at
walking through the arenas and poking around seemed like the simplest way 12:46
the gc doesn't seem to encapsulate well enough to get the parent/child rel easily
12:57 szabgab joined
Whiteknight yeah, I was thinking about that the other day. We do the marking through Parrot_gc_mark_PObj_alive(), but that function doesn't take a reference to the parent so you cant make a graph from it 12:58
there is some arena walking code in src/gc/system.c, it traverses the linked list to determine whether a given pointer is located in a given pool 12:59
there is more stack walking code in src/gc/gc_ms.c in the sweep code. It sweeps over the arenas linearly and frees items that are not alive 13:00
jdv79 yeah, there are many invocation sites for Parrot_gc_mark_PObj_alive(). once i noticed that i started looking for another way. 13:01
Whiteknight you could look into the blocking macros in include/parrot/gc_api.c 13:02
GC_WRITE_BARRIER 13:03
that macro is supposed to be used whenever we add or remove one PMC into an aggregate
it's supposed to be used for GC cores like a generational core and a train algorithm so we can keep track of inter-generational pointers
but since Parrot doesn't have a generational GC, those macros are basically useless, and they haven't been used in the codebase consistently 13:04
if they were used more consistently, you could define them to call a custom function to build the graph 13:05
"if they were used consistently" 13:06
jdv79 well, i'll poke around some more. thanks for the pointers.
Whiteknight no problem, happy to help whenever I can
jdv79 is off to work
Whiteknight I'm sorry things are so messy that there's no help to be had
jdv79 it is what it is
pmichaud Good morning, #parrot 13:17
Util Good morning
purl And good moroning to you, Util.
13:32 Andy joined
dalek kudo: 8037412 | masak++ | src/setting/Any-str.pm:
fixed Perl5-centric POD in Perl 6 code
 13:40
13:40 szabgab joined
dalek kudo: eaa3ea8 | duff++ | src/setting/Any-str.pm:
make POD consistent
 13:51
nopaste "pmichaud" at 72.181.176.220 pasted "exit opcode doesn't set return value" (7 lines) at nopaste.snit.ch/17260
13:52 mokurai joined
dalek TT #834 created by pmichaud++: exit opcode doesn't set process return value 13:57
13:57 ruoso joined 14:05 bacek joined
bacek good localtime() 14:06
14:07 skids joined
Coke msg bacek code.google.com/p/partcl/wiki/TestingPartcl - t/cmd_lsort.t passes all individual tests but reports failures because of RT#57088 ; the namespace failure is NOT expected, however. 14:08
purl Message for bacek stored.
bacek Coke: hey! No need to msg me. I'm here :)
Coke that article has at least one factual inaccuracy. (sdtimes.com/PARROT_1_4_MAKES_PUPPIE...IES/33606) 14:10
bacek Coke: at least about scheduled release? 14:13
Coke bacek: sorry about the /msg, was several hundred lines back. 14:14
bacek: "Since the version 1.0 release in May, outside contributors have begun adding languages, such as APL, Lua and SNOBOL"
bacek Coke: no worries :)
Coke the APL work started a /wee/ bit before 1.0
bacek It's just typo. It should be 0.1 I think 14:15
Coke you're too generous. 14:16
Don't get me wrong, I'm glad to see some press.
pmichaud actually, the point was "outside contributors" 14:19
i.e., people who have played with parrot who aren't on the parrot core team.
dalek rrot: r40077 | bacek++ | trunk/src/pmc/pmcproxy.pmc:
[cage] Remove PMCProxy.mark and .destroy.

PMCProxy.mark and .destroy should be exactly the same as Class'.		 14:22
rrot: r40078 | bacek++ | trunk/src/pmc (2 files):
[cage] Don't use constant Hash for parent_overrides in Class and PMCProxy.
rrot: r40079 | bacek++ | trunk/src/hll.c:
[cage] Don't use constant OrderedHash for HLL_info.

failures.
rrot: r40080 | bacek++ | trunk/src/pmc_freeze.c:
[cage] Clear constant flag during thawing PMC from PBC.
rrot: r40081 | bacek++ | trunk/src/hash.c:
[cage] Put big assert to parrot_hash_put to prevent using non-constant
PerlJam gets the feeling that bacek uses git-svn 14:24
bacek PerlJam: indeed.
I can't use plain svn. It's way to restrictive and inflexible. 14:25
14:26 davidius joined
Coke pmichaud: ok. that's not clear, and ... really? who contributed to APL ? 14:27
or Lua. I thought lua was pretty much feature complete before parrot 1.0 shipped. 14:28
pmichaud I don't know about Lua. For APL, we had that person who inquired about doing things with APL (although that might've been pre-1.0 as well) 14:30
I think the original interview question was "Are you aware of anyone other than the core parrot team who are looking at or playing with building languages on Parrot?" 14:31
Coke I don't know anyone who is taking it seriously, no. =-)
pmichaud But I don't think there was any mention of "post-1.0" in the question.
Coke aside from you. =-)
(that was a joke, btw.) 14:32
dalek rrot: r40082 | whiteknight++ | trunk/src/ops/core.ops:
[ops] make sure to set the exit_code of the new exception object in the exit opcode. Previously we were only setting the 'type', which doesn't translate to exit_code. This should resolve TT #834
pmichaud #ps in 240 14:33
Coke So, I'm still left happy that there is some publicity, but grumbly regarding the reporting.
NotFound pmichaud: I remember an article, maybe 2-3 years ago, about someone that used parrot to port business code written in some old and discontinued 4GL. 14:35
Coke NotFound: ... that was dan. 14:36
Whiteknight pmichaud: can you see if r40082 fixes TT #834 for you?
NotFound Ah, doesn't count as "other than the core", then.
Coke and that didn't end well, as i recall, as parrot changed too much for that project to keep up with.
(which, as a HLL author, I can totally relate to.)
pmichaud NotFound: I suspect it will work, although it's weird to be setting the $1 value in two places. 14:37
NotFound: I'll try it here.
Coke NotFound,pmichaud: we need to add a test for that.
NotFound Ehhhh... What are you talking about? 14:38
pmichaud I'm not sure how to test exit codes in the current test harness.
NotFound: line 957
14:38 Andy joined
NotFound Line 957 of what? 14:40
pmichaud core.ops
the file you just patched
we end up putting the value of $1 in two places.
oh, sorry 14:41
that was Whiteknight
Nick misread
NotFound NM
Whiteknight pmichaud: I think that the value $1 is really just the exit code and not the "type"
pmichaud Whiteknight: so what should the "type" be? 14:42
14:42 Theory joined
Whiteknight i have no idea, but it works as $1 for now and I'm not changing it till I know what it should be 14:42
pmichaud okay. :-)
bacek Coke: afaiu cmd_namespace failing because of hash keys order changed. 14:44
NotFound The keys branck has landed? 14:45
pmichaud Whiteknight: your patch seems to resolve the build issue for Rakudo, many thanks!
bacek NotFound: not yet.
pmichaud now I just need to run a spectest and bump PARROT_REVISION :-)
Whiteknight pmichaud: no problem. I'll look into the issue a bit more and clean things up a little in there
bacek NotFound: I hope r40078 will fix latest issue with packfilerawsegment.t occasional failures... 14:47
NotFound bacek: good!
Coke bacek: checking... 14:48
bacek NotFound: It is not...
Coke bacek: no failures on that one for me. which test? 14:49
bacek Coke: 9.
# got : '::audreyt ::bob ::Bob ::tcl'
# expected : '::audreyt ::Bob ::bob ::tcl'
Coke bacek: I'm calling RPA's .'sort' before returning that list. 14:50
dalek rrot: r40083 | bacek++ | branches/tt761_keys_revamp (38 files):
Bring branch up-to-date with trunk.
bacek Coke: ahem... RPA sort... Sounds familiar. 14:51
Coke using children_cmp sub from runtime/builtin/namespace.pir
... which should be sorting things case insenstive, which could matter on original order of elements, given Bob and bob. 14:52
hurm.
www.tcl.tk/man/tcl8.5/TclCmd/namespace.htm#M5 doesn't indicate that they are sorted, tclsh8.5 passes that file... 14:53
do you have a copy of tclsh8.5 lying about? 14:54
bacek Bah! PackfileRawSegment failure is trivial. Just because someone (me) was too lazy to properly implement test...
Coke bacek: ah. it's not documented, and we shouldn't be relying on it. will explicitly sort them in the test. 14:56
bacek Coke: ok. Glad to hear that not everything is my fault :) 14:57
Coke no, good catch. 15:00
dalek rrot: r40084 | bacek++ | branches/tt761_keys_revamp/t/pmc (2 files):
[t] Add _find_segment_by_prefix helper function and use it in t/packfilerawsegment.t to avoid relying on Hash keys order
 15:04
rdinal: 4865c28 | (Ted Reed)++ | src/classes/Array.pir:
Switch use of Array.append to Array.concat to work around Parrot bug.
 15:05
rdinal: 611c3c2 | (Ted Reed)++ | src/classes/Array.pir:
Correct Array's + operator to be non-destructive. Whoops.
rdinal: 8457f68 | (Ted Reed)++ | src/classes/Array.pir:
Add & operator to Array.
bacek cardinal: Switch use of Array.append to Array.concat to work around Parrot bug. 15:06
polyglotbot OUTPUT[Parrot VM: Can't stat languages/cardinal/cardinal.pbc, code 2.ā¤main: Packfile loading failedā¤]
bacek hmm...
Ted Reed? 15:12
15:12 HG` joined
moritz cardinal: Array.new 15:16
polyglotbot OUTPUT[Parrot VM: Can't stat languages/cardinal/cardinal.pbc, code 2.ā¤main: Packfile loading failedā¤]
Coke bacek: fixed in partcl r535 15:18
bacek Coke: thanks!
Coke were you testing against parrot in your branch? 15:19
bacek Yes. 15:20
15:21 donaldh joined
Coke woot. 15:21
dalek rtcl: r535 | coke++ | trunk/ (2 files):
The output of [namespace children] is not guaranteed to be sorted.

Found thanks to bug report from bacek++
bacek Hooray! Rakudo's make spectest passed without new failures comparing to trunk.
Coke me rants again about the runloop bug. 15:22
bacek Coke: it will be very helpful if I can test partcl against non-installed parrot. Just save ~30 seconds to make install.
Coke: which one?
Coke bacek the one that causes t/cmd_lsort.t to have a bad exit value. 15:24
bacek: (non-installed parrot) - yah - I'm hoping that's going into 1.4 15:25
... but I'm not hacking on it myself.
bacek Coke: fair enough.
clock? 15:30
purl bacek: LAX: Tue 8:30am PDT / CHI: Tue 10:30am CDT / NYC: Tue 11:30am EDT / LON: Tue 4:30pm BST / BER: Tue 5:30pm CEST / IND: Tue 9:00pm IST / TOK: Wed 12:30am JST / SYD: Wed 1:30am EST /
bacek Sleep time. Good night everyone.
bacek departing to bed dreaming about never touch GC bugs again.
Coke stares at moritz's last blog entry. 15:41
moritz somebody reads my blog?
Coke it's agregated into PP6. 15:42
and the last one seems to be empty.
moritz ah, that explains a lot
15:45 Theory joined 15:53 darbelo joined 16:02 Psyche^ joined
dalek kudo: 2fb60ee | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION so that we get appropriate exit status codes (Whiteknight++)
 16:22
pmichaud #ps in 112 16:38
don't forget to pre-paste reports
dalek kudo: b7643e8 | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 415 files, 11784 passing, 0 failing
 16:41
16:45 hercynium joined
cotto hi 17:03
17:06 mikehh joined
Whiteknight hello Cotto! 17:08
cotto hi Whiteknight 17:12
dalek rrot: r40085 | Util++ | trunk/src/io/api.c:
[io] Corrected typo in api.c
 17:17
Coke trolls again for volunteers to put together a version of partcl that uses PCT. 17:24
moritz I thought that wasn't going to be feasible?
due to TCL not having the traditional parse-compile-execute phases 17:25
Coke it does.
purl stays quiet
Coke it's just that the tcl parse is very limited compared to, say, the perl6 parse.
"command word word word" -> dispatch to command. 17:26
not being in own HLL was the last big blocker I recall.
(which is now gone.)
Coke also ponders re-writing some of the runtime in a HLL than pir. 17:28
dalek rrot: r40086 | NotFound++ | trunk/src/hash.c:
[cage] codingstd parens
moritz in TCL?
Coke the only problem with that is dealing properly when things are redefined out from under me. 17:29
does NQP handle exceptions yet? 17:30
moritz no idea
pmichaud NQP doesn't have any special exception handling structures in it, no. 17:35
But we could add some.
Coke pmichaud: just looking at what's in, say, partcl's runtime/builtin/*.pir and seeing if any of that is easier to write or maintain in NQP. 17:37
Whiteknight ponders rewriting partcl in Close 17:38
NotFound ponder writing Open X-) 17:39
Coke I thought partcl was more complete than close. =-) 17:44
cotto The Close spec can be redefined. 17:50
dalek TT #717 closed by Util++: parrot_config segfaults when invoked with the --dump option
darbelo Done := "Where we are now" 17:51
mikehh codetest FAIL All others PASS (pre/post config, smolder fulltest) at r40084 - Ubuntu 9.04 amd64 17:53
had a patch but notfound beat me to it at r40086 17:54
s/notfound/NotFound/ 17:55
NotFound I'm the faster cleaner on the west 17:57
mikehh hey I was doin' fulltest not just codetest :-} 17:58
17:59 allison joined 18:01 Theory joined 18:02 Theory_ joined, chromatic joined
dalek website: darbelo++ | This week on decnum-dynpmcs. 18:03
website: www.parrot.org/content/week-decnum-dynpmcs.
mikehh All tests PASS (pre/post config, smolder fulltest) at r40086 - Ubuntu 9.04 amd64 18:25
Whiteknight chromatic: ping 18:30
chromatic: unping 18:33
18:34 athomason joined
chromatic hooray 18:35
allison ah, chromatic is alive 18:36
Whiteknight chromatic: I'm really interested to see what you have planned for the pluggable_runcore branch
chromatic Not much. 18:37
purl same here, dude
chromatic 1) define a struct runcore_t which represents everything necessary to run a runcore
2) refactor the existing runcores to register themselves
3) refactor runcore invocation to use this struct
4) add a(n optional) loading feature to load a new runcore and register it when Parrot doesn't recognize foo in -R foo 18:38
Whiteknight oh, that all sounds very fun
chromatic Should be simple.
Whiteknight sounds it, and potentially very beneficial 18:39
I think there are probably about a dozen tickets in RT and Trac by various names that suffer from this same problem mechanism 18:51
18:53 mikehh joined 18:55 webnov8 joined
dalek kudo: d8d0640 | pmichaud++ | src/parser/grammar-oper.pg:
infix:ļæ½=>ļæ½ and related operators should be right associative.
 18:55
18:56 autark joined
Whiteknight The issue in TT #833 is the particular confluence of :init subs executing in the PIR compreg immediately and throwing an exception without providing a handler in the compiled code string 18:59
so there are several potential fixes for this:
18:59 rindolf joined
mikehh rakudo builds on parrot r40086 - PASSes make test/make spectest (t/spec/S03-operators/precedence.rakudo - TODO passed: 53-54) 18:59
Whiteknight 1) Don't allow exceptions to propagate out from :init subs when run at :init time (they can be called by name otherwise and should propagate exceptions like normal in those cases) 19:00
mikehh Ubuntu 9.04 amd64
pmichaud #1 isn't workable.
Whiteknight pmichaud: even from :init subs only?
pmichaud Whiteknight: even then.
I can't guarantee that the code I'm compiling won't have an error coming from an :init sub, and I *have* to be able to catch it.
webnov8 hello. i'm new to parrot and was going thru the list of languages on the site looking for a language that i can take for a test drive. i'm trying to get github.com/cardinal/cardinal/tree/master but i'm not familiar with the perl Configure.pl way. any pointers on how to get this going?
allison aye, stopping the propagation of exceptions just isn't an option, ever
Whiteknight 2) force :init subs from the PIR compreg to run in the parent runloop 19:01
(altough that obviously doesn't resolve the rest of the inferior runloop-related problems(
but then again, it might be a good start
pmichaud #2 also doesn't feel like it would be a general solution 19:02
moritz webnov8: without knowing I'd just guess you have to call 'perl Configure.pl' and see what happens - did you try that?
pmichaud there's nothing to say that the parent runloop isn't also inferior to a runloop with an exception handler
Tene webnov8: webnov8 perl Configure.pl --parrot-config=/path/to/parrot_config
Whiteknight pmichaud: not a general solution now, but if we eventually switch to L1 it _will_ be general for all cases
Tene webnov8: you need to tell it where your parrot_config utility is, presumably in your parrot build directory, unless you're installed Parrot
PerlJam webnov8: or perl Configure.pl --gen-parrot # for it to auto download and compile parrot 19:03
pmichaud Whiteknight: do you see L1 happening before 2.0? If not, we need another solution I think.
webnov8 i already have parrot installed using the package provided by my distro
Whiteknight ...no
Tene webnov8: what happens when you just run Configure.pl on its own, then?
chromatic Whiteknight, did you write up a problem analysis somewhere?
Whiteknight TT #833 is the most recent
PerlJam webnov8: what dist is that exactly? And what version of parrot? 19:04
webnov8 Tene, do i run make after the config?
Tene webnov8: yes
webnov8 PerlJam, archlinux with parrot 1.2
Tene, when i run make i get: No rule to make target `/parrot', needed by `cardinal.pbc'. Stop. 19:05
PerlJam webnov8: cool. archlinux++
19:05 MoC joined
moritz webnov8: on some systems there's a 'parrot' and a 'parrot-dev' or 'parrot-devel' package - you need both (don't know about archlinux) 19:06
19:08 treed joined
Whiteknight RT3842 contains a similar problem (with some analysis) where the exception is handled in the child runloop instead of propagating into the parent runloop, and still creates problems 19:08
treed webnov8: I understand you're having build trouble with cardinal?
webnov8 treed, yep
treed I've just started on the project, so I'm still familarizing myself with it, but I'll help if I can.
PerlJam treed++
webnov8 i was just trying to take ruby or python for at a test drive to see what parrot's like 19:09
treed Ah.
Cardinal is still very basic.
Tene treed is the guy I'm trying to con into working on cardinal for me. :)
treed I've been working on adding things to the stdlib.
19:09 Theory joined
treed Starting with Array, and also some with Hash as needed. (Just added a bunch of stuff for includes? etc.) 19:09
Where's the build failing? 19:10
"trying to con" is an interesting description for "started working on it on his own" :-P
webnov8 treed, right at the start: make: *** No rule to make target `/parrot', needed by `cardinal.pbc'. Stop. not sure if i forgot to do something. i did run Configure.pl with the path to my parrot_config like Tene suggested 19:11
treed Hm.
And that's what happens when you just type "make"?
webnov8 yep
PerlJam webnov8: what was the exact Configure line you used? 19:12
webnov8 PerlJam, perl Configure.pl --parrot-config=/usr/bin/parrot_config
treed makes me think $(BUILD_DIR) is empty
grep "^BUILD_DIR" Makefile 19:13
?
webnov8 treed, i had set it to a dir in my home directory cuz home/sergej/community.cvs/devel/parrot/src/parrot-1.2.0 is what was provided 19:15
treed I see.
And there's no parrot binary in there?
(Also, that may get weird if you're using a parrot_config differently from the parrot you're using. 19:16
Coke your parrot_config should very likely be in the same dir as your parrot binary
pmichaud fresh checkout of tt761_keys_revamp branch fails tests on my system (x86 Kubuntu 9.04)
treed I would definitely try building parrot in your build dir and using its parrot_config.
webnov8 ok 19:17
nopaste "pmichaud" at 72.181.176.220 pasted ""make test" failures for tt761_keys_revamp branch (r40086, Kubuntu 9.04 x86)" (8 lines) at nopaste.snit.ch/17265 19:19
19:20 donaldh joined
moritz lots of fallout in cardinal's 'make test' 19:23
treed Oh, god yes.
Blame tene for that. 19:24
Tene blame me? I didn't commit those failing tests.
Maybe.
treed I'll need to look into fixing the test harness soon.
Tene: I can't even get "make test" to run.
Tene >.>
treed I've tried running some of the tests manually. 19:25
Some work, some don't.
But before I can start really working on that, I need the test suite to be easily runnable. 19:26
Tene yeah, I haven't had them all passing since someone sent me a bunch of additional tests
dmknopp, iirc
moritz treed: what's the error message you get?
Tene many of them were failing at the time they were committed due to NYI
treed moritz: No error message. 19:27
Just no output.
nopaste "moritz" at 91.10.186.183 pasted "cardinal 'make test' summary (amd64/linux)" (64 lines) at nopaste.snit.ch/17266
treed NYI?
purl hmmm... NYI is not yet implemented
treed Ah.
treed pets the bot.
I know the failing tests in array/uniq.t aren't due to NYI, at least on the part of array.uniq!
webnov8 installed pynie, no luck there either :S. any language on the language list that i can get up and running just to print hello world :) ? 19:30
treed Looks like mostly just failed plan numbers.
Should be easily fixable.
Whiteknight the problem with learning software lessons in hindisght is that when you've finally learned the lesson you have a mountain of code that does things the wrong way
treed Whiteknight: My maxim has always been "Expect to code everything at least twice; you only really understand the problem after you've solved it once." 19:31
(and more more than once)
Whiteknight treed, yeah I've heard that before
allison webnov8: what trouble did you have with pynie?
Whiteknight That's fine and good for small programs, but we can't just up and throw Parrot away
webnov8 allison, it can't run a simple function that computes the fib sequence. seems like parser issues? 19:32
NotFound writes helloworld 2.0
treed Whiteknight: Well, you recode bits at a time.
Util We can throw Parrot away; let's just do it in small chunks, and call it Refactoring :)
treed +1
purl 1
allison webnov8: well, pynie isn't a complete implementation of python yet, you're likely using features it doesn't support
webnov8 allison, ah i see
treed Is # a comment in PIR?
Whiteknight treed: yes, and it's harder to recode something in place then it is to start from a clean slate and a good idea 19:33
allison webnov8: hello world it'll do fine
webnov8 allison, yea i didn't mean hello world in the literal sense :)
treed Yeah.
webnov8 goes to try Pipp next
treed Whiteknight: That's true.
Util treed: comment - yes
treed That's also life. :-)
allison Whiteknight: hmmm... well it often *seems* simpler to start over, but the devil is in the details
treed Util: Thanks, realized it shortly thereafter.
Util Unquoted # is command-to-end-of-line
s/command/comment/ 19:34
particle notes that there is no news entry on parrot.org pointing to www.sdtimes.com/PARROT_1_4_EXPANDS_...RROT/33606
Whiteknight allison: even easier then starting over is speculating idly on IRC about starting over
19:34 eternaleye joined
PerlJam Whiteknight: is that what L1 is? :) 19:34
Whiteknight is a pro speculator
allison Whiteknight: true, true 19:35
19:35 mikehh joined
allison particle: are you adding one? if not I will 19:35
cotto btw, you can change and/or shorten that URL 19:36
particle i am not.
rats, missed #ps again
pmichaud "Impeachment!"
particle: off hiking in the Appalachians? 19:37
particle instead, i'm working on getting a better donor management system set up
no, but i'll be cycling in eastern washington this weekend :)
Whiteknight treed: can you post a code example to TT #835 that demonstrates the problem?
allison particle: excellent
treed Whiteknight: I did. 19:38
Should be attached.
Whiteknight oh, haha, I didn't see it
sorry
treed No problem.
dalek TT #835 created by treed++: Crash when appending to subclasses of ResizablePMCArray 19:39
treed Came up while I was working on CardinalArray.
I worked around it by just implementing Array.concat and using that.
(Which needed to be implemented anyway.)
Whiteknight that's still obnoxious 19:42
Whiteknight hates all the stupid little failures with subclassing
Tene I found some clues indicating that I'd run into this a few times in the past. 19:43
pmichaud particle: there's an important release in seven days and you're going to be off cycling in eastern washington?!? :-)
Tene I'm starting to remember that I actually tracked down exactly what the problem was, but didn't know how to fix it.
Of course, it's not recorded anywhere...
pmichaud I guess I can't complain too much -- I'm thinking of heading to the west coast this weekend also. :-)
Tene I'm going out to Portland the week of July 27 19:44
particle no, just for the weekend. i'll be ignoring the release from home :P
parrot releases are boring anyway.
Tene as they should be
particle :)
pmichaud opens his (early) birthday present
Whiteknight what about the clever names we give them? aren't you on the edge of your seat to find out what 1.4 will be called?
treed lives on the west coast. 19:45
particle wonders if it's a blue laser
cotto Whiteknight, I know I am. ;)
moritz Whiteknight: parrot "you'll be seeing the dumb release namem for 6 month" 1.4
Whiteknight who is doing the release this month?
pmichaud oooh, blue laser would be cool.
particle pmichaud has a lot of toys
pmichaud I'd even go for the blue-violet one.
but for this year's present I got a new SSD for my laptop :-) 19:46
cotto By the time the release happens, I'll probably have spent more time thinking about that than working on the release.
btw, anyone have thoughts on this: nopaste.snit.ch/17257 ? It's an update to the ops file format to make parsing easier. 19:47
pmichaud Let's call the 1.4 release "Unstabel" :-P
cotto My current favorite is in that same vein.
chromatic cotto, that seems sensible to me. 19:48
cotto Cool. I'll file a tt and make a deprecation note. 19:49
Whiteknight cotto: that nopaste looks good to me
19:50 szabgab joined
cotto This is good: 19:51
nopaste "cotto" at 74.61.2.46 pasted "arraybug.pir backtrace" (27 lines) at nopaste.snit.ch/17267
cotto using the arraybug.pir from tt #835
treed Is it GCing too early? 19:52
Or just a GC failure? 19:53
chromatic It's during global destruction.
treed I was experiencing the same bug in the middle of a longer script, though. 19:54
Util cotto: ops file format change looks good to me. 19:56
19:58 webnov8 left
dalek TT #836 created by cotto++: [DEPRECATED] ops file preamble format 20:00
treed Does that affect cardinal in any way? 20:01
20:02 autark joined
szabgab in case it was not mentioned here earlier: www.sdtimes.com/PARROT_1_4_EXPANDS_...RROT/33606 20:03
cotto chromatic, Whiteknight, Util: thanks
dalek rrot: r40087 | cotto++ | trunk/DEPRECATED.pod:
[DEPRECATED] add deprecation notice for ops preamble and VERSION line
 20:07
Whiteknight cotto++ 20:16
Util Regarding treed's new tt #835 - in resizablepmcarray.pmc: 20:19
memmove(this_data + n, other_data, m * sizeof (PMC *));
Shouldn't that `+ n` be `+ n * sizeof (PMC *)` instead? 20:20
chromatic Depends what this_data is.
20:20 rindolf left
Util chromatic: PMC **this_data = PMC_array(SELF); 20:21
chromatic Then I believe that the math works out correctly as written.
This is an area of C that refuses to stay in my head if I think about it too long, however.
Util chromatic: Oh, I see. 20:23
chromatic It's not adding bytes, in that case. It's adding PMC pointers. 20:24
dalek rrot: r40088 | NotFound++ | trunk (3 files):
[DEPRECATED] add deprecation notices for Parrot_str_free, TT #826
 20:25
treed Except that a pointer is 4 bytes on 32-bit systems, right?
chromatic Sure, and 8 bytes on 64-bit systems.
treed So when it's adding PMC pointers, it's adding bytes. 20:26
chromatic It's adding bytes, if you want to look at it that way, but it knows the right number of bytes because it's not adding them as mere bytes.
You can say "C has no type system" and I mostly agree.
treed used to do C and ASM way back in the day. 20:27
(Which is funny, coming from a 25-year-old.)
chromatic In this case C at least pretends it has a type system.
treed But n is still an int, yeah? And "this_data" is presumably a point.
er
20:28 khisanth_ joined
chromatic As I understand it, and I could be wrong, the compiler treats that expression equivalent to an array access. 20:28
this_data[n] is equivalent to this_data + n
treed Hm.
NotFound Yeah
treed Sounds wrong to me, but I haven't done C in a long time.
Hm. 20:29
When you state it as that equivalency, it does make more sense.
chromatic Just remember that C doesn't have arrays, and you'll be fine.
pmichaud ooooh, ssd is fast. :-)
treed Yeah, C has pointers, and syntactic sugar. 20:30
NotFound Even funnier n[this_data] means the same.
Tene maybe someday they'll come out with a solid-state CPU.
chromatic I'm waiting for a solid-state society.
pmichaud I always thought of this_data[n] as being the same as *(this_data + n) 20:31
moritz Tene: in what ways are current CPUs not solid state?
NotFound Don't be fooled, C has arrays. It just decays them to pointer in almost any usage.
Tene moritz: none
moritz: It was a joke.
moritz ah
pmichaud moritz: I was wondering about it also, though.
moritz my joke-o-meter is already sleeping I feare 20:32
s/e$//
treed moritz: It can be awefully hard to tell when Tene's joking.
pmichaud moritz: I just didn't take the bait. :-)
treed Even in person.
moritz pmichaud: you don't work in solid state physics, you're excused ;-)
pmichaud notices that his cd-rom drive is very loud. 20:33
Tene treed: and in response to your question, isn't the best con one the mark thinks was his idea all along?
moritz pmichaud: mechanical components are always a mistake ;-) 20:34
Util pmichaud: it is to make up for the lack of HD noise.
pmichaud Util: yes, that's what I was just thinking.
Tene moritz: when are you getting rid of your mechanical components?
treed Tene: Heh.
Can anyone help me figure out why my PIR is failing to parse?
moritz Tene: I try, but sofar nothing could replace my keyboard 20:35
NotFound moritz: there are no mechanical keyboards
moritz NotFound: all keyboards I used so far had mechanical components 20:36
pmichaud if I could get a Kubuntu ISO image to boot from a usb or sd device, then I wouldn't need to use the cd-rom drive :-)
20:36 mokurai joined
Tene pmi\tlaptop doesn't usb boot? 20:36
pmichaud Tene: it's supposed to -- I've never had luck getting it to do so though.
NotFound moritz: there are thinks that project a keyboard on your table and read finger movements.
pmichaud but I suspect my preparation of the boot device more than the laptop.
NotFound s/thinks/things
pmichaud www.thinkgeek.com/computing/keyboards-mice/8193/ 20:37
moritz bah, no linux support 20:39
pmichaud I bet it works out-of-box with linux.
Linux bluetooth support is pretty good.
moritz assuming that it registers itself as a normal input device
and doesn't require some driver magic 20:40
pmichaud one would hope that it does, given the number of smartphone devices it works with
moritz that's something I'd try in a shop, but wouldn't order over the internet blindly
pmichaud (register itself as a normal input device, that is)
Tene My driver was written by pixies in the enchanted forest!
moritz (at least not until I have a real income ;-)
Tene moritz: you should put it on an INTERNETS WISHLIST!!! 20:41
moritz Tene: and hope that magic fairies find it, and order it for me?
(is that how the pcc rewiring is going to happen? /me ducks and runs) 20:42
pmichaud www.linuxquestions.org/questions/li...ty-657238/ 20:43
Tene are you calling allison a magic fairy?
moritz I never met her
20:43 Theory joined
moritz so as long as she writes awesome code she is one, kind of ;-) 20:43
(which applies to most other people here ;-)
treed But do the other people here *want* to be a magic fairy? :-P 20:44
chromatic No.
Tene I do! 20:45
pmichaud Do I get wings? Will I be able to fly?
moritz pmichaud: you will, armored wings even. Like camelia ;-)
allison I want pixie dust, to sprinkle on difficult bugs ;) 20:46
NotFound treed: www.youtube.com/watch?v=kg5nN9S2qTk 20:47
moritz allison: oh you don't have that? now I'm disappointed
pmichaud we had some, but we ran out.
Tene pmichaud: I have a vague memory that I was supposed to do something with HLL namespace imports... maybe from japhb... did I do that yet?
pmichaud Tene: I think we're both supposed to be doing something with it. I'll definitely be reviewing the whole thing tomorrow and/or Thu 20:48
20:48 Lorn joined
allison we're working on a secret underground project to artificially manufacture more using only carrots and newts 20:48
treed NotFound: Um, wow. 20:49
NotFound likes Iginio Straffi woks
japhb Tene: There's the problem of 'use OpenGL::Math:from<parrot>;' and then trying to instantiate an 'OpenGL::Math::Vec4' and instead getting FAIL. Is that the problem to which you refer? 20:51
Tene Yeah, that sounds right.
pmichaud I'd work on it today, but I have a new toy to play with.
japhb Ping me if you need anything (more details, testing, whatever)
pmichaud: new toy?
purl new toy is G4?
pmichaud japhb: ssd drive for the notebook
japhb pmichaud: ah! Which one? 20:52
pmichaud intel 80GB
x18-m
Tene I'd work on it today, but I'm behind at work, and there's socialization scheduled for tonight.
japhb nice
20:52 ascent joined
pmichaud so far very fast. I'll have some timings here shortly -- have to reinstall operating systems and the like. 20:52
PerlJam pmichaud: we got some SSDs at work. /away 20:56
oops
20:58 mikehh_ joined
particle cotto: it may make sense to modify NEWS to list changes since 1.0, augmenting or superseding changes since 1.3. 20:59
PerlJam pm: anyway. The ones we got don't seem particularly low-power but they are about 50% faster based on my ad hoc timings
particle i moved my swapfile and virtual machines to a 48GB SSD on my laptop, noticeable difference in pep 21:00
21:02 iblechbot joined
dalek rdinal: c99d2b3 | (Ted Reed)++ | src/classes/Hash.pir:
Add methods to CardinalHash for checking whether a key exists/has a value.
 21:25
rdinal: 2cedd29 | (Ted Reed)++ | src/classes/Array.pir:
Implement - operator for CardinalArray.
21:28 Whiteknight joined 21:38 mokurai joined
treed Huh. Had no idea my commits were being echoed here. 21:46
cotto particle, good idea 21:48
except that it means more work for me ;)
particle yeah, you have to write an email requesting the thundering herd to help you write it :) 21:49
cotto allison, ping 21:50
allison cotto, pong
cotto allison, what happened with tewk? 21:51
allison cotto: he got a substantial grant from the university that left no time for GSOC
moritz it would have been nice to announce that, at least
allison cotto: I talked with the GSOC admins, so he's just been withdrawn from the program 21:52
cotto thanks
It's good to know that he didn't flake out. That'd be unlike someone who's contributed to Parrot already.
His grant doesn't have anything to do with Parrot, does it? 21:53
cotto hopes
22:17 mokurai joined
Whiteknight cotto: excellent email to the list! 22:20
22:34 kid51 joined 22:36 rg1 joined 22:43 bacek joined
cotto Whiteknight, thanks. 22:46
bacek Good morning 22:52
purl And good moroning to you, bacek.
23:07 skids joined
Whiteknight good morning bacek 23:09
bacek good morning Whiteknight
Whiteknight brb (switching to windows partition)
23:15 whiteknight joined
whiteknight is going to try to build Parrot again with 64-bit CL on windows 23:16
cotto living on the wild side
whiteknight I do it for the lulz 23:17
my friend and I were talking about starting our own server colocation hosting service, and we were going to name the company "TEH INTERNETZ" 23:18
dalek rrot: r40089 | jkeenan++ | branches/tt509_install_files:
Creating tt509_install_files in svn.parrot.org/parrot/branches
 23:20
23:20 donaldh joined
bacek created ticket for NQP and decommute to $dayjob 23:24
See you soon.
dalek TT #837 created by bacek++: [DEPRECATED] NQP quoting rules going to be close to Perl6. 23:25
23:30 whiteknight joined
kid51 Smolder down (again) :-( 23:30
whiteknight kid51: I have to manually edit the compiler flags without reconfiguring. Where do I do that? 23:31
treed export CPPFLAGS? 23:32
Or the like.
Assuming it's a C++ compiler.
whiteknight I'm on Windows, so who knows what will work
treed Oh.
Good luck, sir.
cotto whiteknight, you mean like editing the Makefile manually? 23:33
whiteknight I guess, is that where the build flags are stored?
I'm on windows, and I need to build with the flag "/GS-", but somewhere configure helpfully changed that to "\\GS-" and now it doesnt work 23:34
yay, it's building! 23:41
if everything miraculously goes my way for the rest of the night, we may have a "real" win64 smoke report soon 23:42
kid51 whiteknight: Can you clarify your request? 23:45
whiteknight oh nevermind, I figured it out 23:46
I had to edit the makefile directly
kid51 notes that his Linux disk is 99% full
Well, perhaps we should figure out where that originated. That might be a configuration bug.
whiteknight: And since you're on Windows ... could you test the tt509_install_files branch per the instructions in the latest post to TT #509? 23:47
Actually... anyone on Win32 who can test that, that would be appreciated. 23:48
ccache -C cleans up 23% of my drive!
whiteknight I'm not on Win32, I'm on Win64 and I can't even build trunk 23:49
kid51 has never encountered Win64
whiteknight that's because it doesn't work
kid51 whiteknight Don't you have enough on your plate that you shouldn't be attempting to get that OS working?
Which is more important: Win64 or Parrot GC? 23:50
Huh? huh?
whiteknight kid51: I have to bounce around to different projects occasionally, or I'll go insane 23:51
chromatic www.gamedev.net/community/forums/to..._id=496643
whiteknight plus we have the release coming up soon, and I would love to add a new entry to PLATFORMS
chromatic social.msdn.microsoft.com/forums/en...c6615b66a/ 23:53
23:55 mokurai joined
kid51 Well, I can't find any relevant string 'GS' in the config/ 23:56
So I don't know where it's coming from.
chromatic Probably an environment variable. 23:57
whiteknight kid51: I'm adding it
I specify it to Configure.pl
kid51 k 23:58
But then perhaps we need a way to preserve that OS-specific backslash.
*Not* that we need worry about that in the next 7 days.
whiteknight will file a bug, to be dealt with at our leisure 23:59

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!