Parrot 1.1.0 Released | parrot.org/ | 322 RTs left | Weekly Priority: Remove Deprecated Items
Set by moderator on 6 May 2009.
darbelo I completely missed that function. Thanks! 00:00
Whiteknight PMC *da_new_one = pmc_new(interp, pmc_type(interp, CONST_STRING("DecContext")));
(spread that out over 2 lines for readability PLZKTHX) 00:01
My only idea is that contexts used to be more complicated, but parts of them got ripped out and now they're needlessly complicated for no reason 00:18
because otherwise I can't understand why they would be as complicated as they are for the task they perform
rg well if you can simplify code, by all means go ahead. 00:20
Whiteknight don't worry, I'm going to do it as soon as I figure it all out 00:24
00:55 tetragon joined 01:34 bacek joined 01:36 eternaleye joined 02:25 kid51 joined 02:35 janus joined
cotto perl wins: blog.doloreslabs.com/2009/05/the-pr...est-users/ (although perl6 is notably absent ;) 02:37
assuming that tweets constitute a meaningful sample
and a bunch of other questionable assumptions 02:39
02:40 Theory joined
cotto I do wish the implicit assumption that PHP doesn't exist were more valid. 02:41
Infinoid I guess PHP users hate themselves 02:56
02:57 japhb joined 03:20 donaldh joined
cotto pmichaud, ping 03:46
bacek, ping 04:04
04:05 Theory joined
bacek_ cotto: pong 04:15
cotto How do you eventually view pmcc being used? Something like: 04:18
pmcc foo/bar/baz.pmc #creates dump, header and c, or...
pmcc --file=c foo/bar/baz.pmc #just create the c file
bacek_ I prefer single shot. 04:19
cotto me too. just checking that we're thinking the same thing, since pmc2c does multiple passes
do you know how to add command-line options to a compiler? 04:20
bacek_ cotto: no...
cotto ok
I was looking and thought I'd try to avoid duplicated effort. 04:21
but since it's not duplicated, I'll keep digging
bacek_ Check "stages" in pct/src/PCT/HLLCompiler 04:22
s/stages/.sub command_line/
cotto nope. stages was the right thing to look at 05:12
bacek_, does pmcc use the post, pir or evalpmc stages? 05:19
05:19 rakudohudson joined 05:34 jan joined 05:54 uniejo joined
dalek rrot: r38706 | petdance++ | trunk (2 files):
removed unused vars, and added more headerizer flags
 06:20
bacek_ cotto: not yet. 06:27
cotto: but it probably will when we implement L1 language 06:28
dalek rrot: r38707 | petdance++ | trunk/config/gen/platform/generic/itimer.c:
return NULL, not 0, as a pointer.
 06:34
cotto getting the thing working is tricky enough now, without worrying about L1 06:35
dalek rrot: r38708 | cotto++ | trunk/runtime/parrot/library/Getopt/Obj.pir:
[getopt] documentation function name update
 06:40
06:47 iblechbot joined 07:20 donaldh joined 08:33 masak joined
cotto going to yapc10? 09:19
purl going to yapc10 is probably kid51 or pmichaud or davidfetter or cotto
09:36 AndyA joined 09:37 zpmorgan joined 09:45 pancake joined
pancake mmh parrot has no manpage? 09:46
cotto care to write one? 09:49
If you're looking for documentation, there's lots in docs/ .
pancake i just wonder why there is no manpage for parrot 09:50
cotto I imagine that a man page isn't a huge priority because Parrot isn't something we expect end-users to care about directly.
i.e. they'll be more interested in the HLLs that run on top of Parrot.
pancake i think that end users will man the program after running parrot -h. so i think that's something people will look for 09:51
yep HLLs are more interesting than manpages, but ok, knowing that there's no manpage i will stop looking for it :P and maybe i will write one 09:52
cotto That'll probably be a common for users to look. It'd be a good idea to put something there.
Feel free to write up a man page that describes Parrot's basic purpose and use, and open a TT. 09:53
pancake ok, i dont have so much free time, but i'll bear in mind 09:54
cotto you can also open a tt requesting a man page 09:55
cotto is looking forward to not having enough time due to having a job 09:56
at some indeterminate point in the future
pancake heh 10:04
i would prefer having some more free time (R) :)
cotto trade? 10:07
purl trade is, like, illegal
pancake uh? 10:09
cotto nm 10:10
time for sleep 10:11
night
10:12 burmas joined
dalek rrot: r38709 | bacek++ | branches/pmc_pct/compilers/pmcc/src/emitter.pm:
Fix creating PMC-specific emitter
 10:37
rrot: r38710 | bacek++ | branches/pmc_pct/compilers/pmcc/src/emitter/pmc/default.pir:
Put default's emitter into namespace
10:53 kid51 joined 11:20 donaldh joined 11:55 Whiteknight joined
dalek kudo: b6c0eaf | jnthn++ | src/ (2 files):
A couple of minor fixes to get wrappers working a bit better. Wins us +20 tests we never passed before in wrap.t.
 12:03
kudo: 0bfbc44 | jnthn++ | src/builtins/guts.pir:
Make sure we hand back return values when invoking with a candidate list.
kudo: 1df5839 | jnthn++ | src/ (3 files):
Factor sub cloning stuff into a macro so we don't have to spread it everywhere; also added it to array stores, so @a = &foo will now work properly.
 12:09
12:09 burmas left 12:12 shadowspar joined 12:25 HG` joined 12:40 rg1 joined 13:00 ruoso joined
cotto pmichaud, ping 13:07
dalek kudo: 7cc0b68 | jnthn++ | src/parser/grammar.pg:
Parse @foo>>.bar in a pretty STD.pm-ish kinda way.
kudo: 85e5b9c | jnthn++ | t/harness:
Twiddle the test harness patch to perl6 on Win32 so env.t works under the harness. I'm still a little bewildered how we ever managed to run any tests...
13:12 gryphon joined
pmichaud cotto: pong 13:16
cotto How can I make extra information available to a compiler stage? In pmcc (the PCT-based PMC compiler), I need to have a list of directories available to the past stage, so the compiler knows where to find the .dump of a parent. 13:17
pmichaud nodes are also hashes.
purl okay, pmichaud.
pmichaud can also use contextual variables. 13:18
cotto that'd work 13:19
Also, does Getopt provide a way to specify required args? f=s and f:s appear to do the same thing.
thanks 13:20
pmichaud I don't know. As far as I'm concerned, Getopt is on its way out; it'll be replaced by the module that particle++ is likely building.
cotto ok
it's not blocking anything anyway
It seems like half of Parrot is in the process of being replaced. 13:21
Whiteknight only the bad half
:)
pmichaud I fear we already replaced the good half.
Whiteknight you think?
purl No, I'm a bot.
masak purl: you're an annoying bot. 13:22
purl masak: excuse me?
13:23 whoppix joined
Whiteknight pmichaud: What things do you think were good but got replaceD? 13:24
dalek kudo: 7d6bc0a | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 387 files, 11187 passing, 0 failing
 13:25
rrot: r38711 | NotFound++ | trunk/src/packfile.c:
[cage] avoid a string to cstring conversion
 13:36
13:51 Andy joined
cotto pmichaud, any objections to this patch: 13:58
nopaste "cotto" at 96.26.202.243 pasted "make filename information available to compilers" (12 lines) at nopaste.snit.ch/16524
"cotto" at 96.26.202.243 pasted "make filename information available to compilers, v2" (16 lines) at nopaste.snit.ch/16525 13:59
pmichaud cotto: yes, because I don't know what it's for. I'd like to not put a lot of compiler-specific stuff into HLLCompiler.
dalek rrot: r38712 | coke++ | trunk/t/codingstd/copyright.t:
[t] re TODO this test;

   Keep it todo'd until the underlying task is resolved.
rrot: r38713 | NotFound++ | trunk/src/packfile.c:
[cage] avoid a string to cstring conversion
pmichaud what is the 'filename' option here?
oh, you're trying to get $0 into thelist?
cotto pmichaud, that's the only way I could see for a compiler to find the name of the file it's compiling
*file(s)
pmichaud then it's in the wrong place. 14:01
cotto ok. What's the right place?
pmichaud for one, a compiler can override 'evalfiles' and get it that way. 14:02
cotto yes, but I'd rather not duplicate the code there unnecessarily
pmichaud so, grab the filename and then call the parent's evalfiles 14:03
no need to duplicate code.
cotto of course. Thanks.
dalek rrot: r38714 | coke++ | trunk (2 files):
[t] Make RE's work in both 5.8.8 <-> 5.8.9
 14:13
rrot: r38715 | coke++ | trunk/t/codingstd/c_function_docs.t:
[t] fix test to run in older perls.
 14:19
Coke ping cotto 14:24
purl I can't find cotto in the DNS. 14:25
Coke fine, cotto: ping. =-)
cotto: fixed TT#543.
cotto Coke, pong 14:28
Thanks. Coke++
Coke here's the root cause: 14:29
10:27 <@maukeNSFW> my $re = qr/A$/m; "A\\nB" =~ /$re/ or die;
moritz if you mean the end of string in /m mode, use \\z instead of $ 14:30
cxreg still fails in 5.8.8 14:32
no wait, it's the \\n not matching $ that's the problem. \\z is entirely wrong 14:33
moritz do you want it to match, or not to match?
cxreg it fails to match in 5.8.8
that's a bug
purl No, it's a feature.
cxreg fixed in 5.8.9 and 5.10, apparently 14:34
Coke oh, hey, cxreg. =-) 14:46
14:50 davidfetter joined 14:55 particle1 joined
dalek kudo: ee9d917 | jnthn++ | src/parser/grammar.pg:
Fix parsing of parallel dispatch unicode form.
 15:20
kudo: 65102d2 | jnthn++ | src/ (2 files):
Implement parallel dispatch (>>. and unicode variant). Handles most cases.
15:20 iblechbot joined
15:20 donaldh joined
dalek kudo: f3d2dab | jnthn++ | t/spectest.data:
Add S12-methods/parallel-dispatch.t to spectest.data.
 15:26
kudo: b670f01 | jnthn++ | t/spectest.data:
Add missing #icu to various tests that now need it.
15:55 mikehh joined
NotFound Coke: ping 15:57
15:58 Theory joined 16:02 ruoso joined
dalek rrot: r38716 | NotFound++ | trunk/compilers/imcc/pbc.c:
[imcc] attempt to fix TT #629
 16:02
rrot: r38717 | coke++ | trunk/t/compilers/imcc/syn/regressions.t:
[t] add regressions for RT #41788
 16:22
Coke NotFound: pong
NotFound Coke: please take a look at TT #629 16:26
Coke NotFound: it fixes the original issue, yup. 16:40
NotFound Coke: Did you agree on leaving the comment and the ticket opened?
Coke should probably add that as a regression test.
NotFound: no opinion.
dalek rrot: r38718 | NotFound++ | trunk (14 files):
[cage] drop include/ from path in .include directives, TT #511
 16:48
kudo: 89539a9 | jnthn++ | src/ (3 files):
Implement parsing of $foo.=bar to work as STD.pm does, rather than falling back to infix:<.=> as we did before. This means that .=foo alone is now a mutating call on $_ and also that >>.= now works.
 16:52
16:57 Andy joined
particle- where are my directors? 17:04
jonathan In the jacuzi, with the champagne! 17:05
particle- i must have missed the memo...
Coke NotFound: I added a test for #629 17:07
dalek rrot: r38719 | coke++ | trunk/t/compilers/imcc/syn/regressions.t:
[t] add a (passing) test for TT #629
 17:09
NotFound Coke: good :)
particle- www.parrot.org/ now has a 'donate' menu item at the top 17:12
ugly, but working.
NotFound Now we need one 'Receive donations', for developpers only ;) 17:17
pmichaud #ps in 60 17:31
dalek kudo: 62773e0 | jnthn++ | src/builtins/guts.pir:
Make sure we don't get unwanted flattening going on when forming the result list from a parallel dispatch.
 17:38
17:58 NotFound joined
NotFound hi 17:59
purl bonjour, NotFound.
NotFound Building perl6 froze my laptop
18:03 AndyA joined
Tene considers harassing chromatic for socialization. 18:08
18:09 darbelo joined 18:10 allison joined 18:13 barney joined
Util NotFound: If you are not building with a MS compiler on Win32, then try sial.org/pbot/36403 to eliminate that freeze. 18:17
NotFound Util: ubuntu i386, but short of ram 18:19
18:20 cghene joined 18:21 mikehh joined
Util The patch should fix it. It shortens the final pbc_to_exe step from around 2 hours to 5 seconds on low-RAM machines. 18:22
NotFound Util: I know, I just comment my problem to show the importance of aplying that patch ;) 18:23
Util The patch is uncommitted only because I am re-working it for a grevious limitation under MS compilers.
Oh! thanks.
18:24 ilia joined, chromatic joined
chromatic will be 10-15 late to #ps 18:24
dalek rrot: r38720 | pmichaud++ | trunk/runtime/parrot/library/PGE/Perl6Grammar.pir:
[pge]: Update Perl6Grammar.pir to handle utf8 input a bit better.
 18:25
18:30 jhorwitz joined 18:31 mikehh joined 18:37 viklund joined
jonathan is disconcerted by the unalphabeticity of #ps 18:38
Coke awwwww. Better chromatic decides, eh friend? 18:39
cotto partcl needs a hug 18:44
Coke ... or a mercy killing.
particle- `svn rm partcl` 18:45
where's mdiep, anyway?
purl mdiep is Matt Diephouse <mailto:matt@diephouse.com>
particle- learn your interrogatives, purl 18:46
purl particle-: sorry...
allison well, the problem isn't partcl, it's parrot, at the moment
cotto "It's not you, it's me."
18:46 Theory joined
particle- allison: did you get back to bradley kuhn? 18:46
Coke mdiep has a real job and a fiance and is busy. =-) 18:47
allison particle: aye, I sent him the new link, and let him know the donation would be welcome
particle- a fiance? is she tall?
wonderful.
ooh, i need to register a talk for yapc::na. 18:48
keynote, in fact.
following larry.
whee!
Whiteknight great, I missed #ps again 18:56
what I need to do is quit my $real_job so I don't get distracted on tuesdays
Coke ps is still going on. 18:57
Whiteknight yeah, but I missed the reporting portion of it
Coke there. 18:58
Whiteknight bah!
Infinoid Whiteknight: if you quit your $real_job, you won't even know it's tuesday 19:00
Whiteknight right, but I'll be on IRC every day 19:01
#ps will just form around me
19:05 particle1 joined
cxreg Coke: i'm less sure now that the /m bug is previously known 19:10
nobody ive asked had heard of it (tye, jjore)
Coke cxreg: whee!
moritz I know that the scoping of (?m ) changed between 5.8 and 5.10 19:12
but I'm not sure if that's the same issue
cxreg tye suspected it had something to do with that, so it sounds right 19:13
he thought that maybe ?m: didn't work prior to 5.8.9
as a subexpression, anyway 19:14
Coke moritz: that is the issue we were having, yes. 19:15
qr//m resulted in a (?m), which failed, but using an explicit /m at m/ or s// time worked.
19:20 donaldh joined
pmichaud how do I get tickets to show up on the roadmap report in trac? 19:35
(#14) 19:36
particle- associate them with a milestone iirc 19:37
or is it release?
i'd have to look at trac to verify....
pmichaud I did associate with a milestone, it didn't work.
Ticket #661
Coke pmichaud: it's set to milestone 1.2 19:40
19:41 ilia joined
pmichaud ...it should be set to...? 19:41
Coke no, that's right. 19:42
nopaste "NotFound" at 213.96.228.50 pasted "This is how reading from a piped command must look like?" (5 lines) at nopaste.snit.ch/16529 19:43
Coke trac.parrot.org/parrot/query?statu...estone=1.2 shows it.
pmichaud why doesn't it appear in report #14, then? 19:44
Coke because that report is for 'milestone' tickets only. 19:45
and yours is a TODO.
I would argue that yours should stay a todo.
pmichaud ...because? (ooc)
19:47 ilia_ joined
Coke milestone are the big ticket items that were originally on the roadmap document. 19:48
like "this entire subsystem must be completed"
particle- how do i get the hex/octal value of a character > 127 19:49
in p5?
pmichaud I'm not sure that I agree with "originally on the roadmap document" clause 19:50
I think that it should just be "this must be completed"
I'm fine with arguing that it doesn't have to be completed for 1.2
but if we get to 1.4 and don't have pipes, we're likely in for some hurt.
Coke perl -e 'printf "%x\\n", 257'; 19:51
if it just has to be completed, then the milestone is sufficient.
PerlJam particle-: my $oct = sprintf("%o", ord($char)); # same for hex but use %x instead of %o
Coke ticket type and milestone are orthoganal. 19:52
NotFound pmichaud: I'll make them work at leaste in linux in a short time.
Coke pmichaud: I appreciate your pain. I'm still trying to recover from 1.0
pmichaud Coke: if milestone is sufficient, when/where do those tickets get reviewed prior to release? 19:53
(again, ooc)
Coke up until this week, in the weekly PS meetings. 19:54
unless we just haven't gotten to that yet.
pmichaud ...except the ticket that I just filed doesn't show up in the list that is in the weekly PS meeting.
*is reviewed in
Coke Ok. You need to take it up with allison, then, as she (SFAIK) created the milestone ticket type. 19:55
the report I look at is here:
trac.parrot.org/parrot/roadmap
sorry.
allison pmichaud: that's a different report, that's the milestone report, instead of the roadmap report 19:56
pmichaud which one is the different report?
The /14 report is the one I'm asking about. 19:57
allison aye, /14 is only roadmap items 20:00
pmichaud how do things get on the roadmap? Or, what's the criteria for placing something on the roadmap, such that it isn't forgotten from one #ps to the next? 20:01
Coke my 2c: we shouldn't be relying on the /14 report for "stuff we need to get done for a release." That's why we have the trac milestones in the first place. the roadmap report is for big ticket items. (which makes it more helpful for planning than doing.) 20:03
allison Coke: aye, but should we be reviewing every ticket tagged for a release in every #ps meeting?
20:04 ilia joined
allison we added the roadmap review to make sure those items didn't get missed in the larger list of todos 20:04
Coke allison: (reviewing) no.
pmichaud if a ticket is tagged for a release, that's because we expect it to be done by that release, yes? I'm not saying we need to review them at every #ps, but at some point we do need to be saying "are we getting the tasks done that we expect to be done"?
Coke given how frequently we just bump-to-next-release, I can tell you, "no, we're not." 20:05
chromatic Yeah, otherwise we won't get them done and we won't improve our scheduling and estimating abilities.
allison tickets are tagged for a release for various reasons 20:06
sometimes it's no more than someone saying "I'd really like this to be a high priority"
pmichaud there should be some way for us to accept that a ticket is high priority, though. 20:07
Coke (note that there are currently 30 tickets marked with the 1.2 milestone.)
pmichaud (not simply that someone says "I'd like it to be high priority", but rather that we as the parrot development team agree that "this is high priority")
Coke pmichaud: there's a priority tag.
Whiteknight there is a priority field in tickets
pmichaud Whiteknight: sure, but if that is never reviewed/acted upon, it's not helpful. 20:08
the only reason I marked #661 as milestone 1.2 was because allison said "would be nice for 1.2". But I'd consider it somewhat critical for 1.4.
Whiteknight pmichaud: you can lead the horse to water, but you cant make them close the ticket 20:09
...or something like that
pmichaud if others disagree on the critical-ness, that's fine. But my larger question is "when we as a group decide that something is really important, how do we indicate that so that we're certain to yoke a person to it?"
Coke mark it as critical. 20:10
Whiteknight and assign it to an owner
pmichaud Coke: but when do such tickets get noticed?
allison pmichaud: the main value of those markings is providing a quick-glance categorization
I always look at the list of milestone tickets before the release
Coke pmichaud: when the volunteers look at them. or when the person requesting the ticket says 'hey, this is important, is anyone going to work on it.'
Whiteknight to make sure the ticket gets noticed it needs an owner, and should probably end up on the roadmap if it's a big deal
allison and, if they're marked as critical, the appear at the top of the list and get more attention 20:11
pmichaud I'm asking "how do things end up on the roadmap"?
Coke I don't think whether they're on the roadmap or not matters to getting them resolved.
allison Whiteknight: the roadmap is for things that we specifically planned in advance that needed to happen
dalek rrot: r38721 | NotFound++ | trunk/src/io/unix.c:
[core] fix pipe open in unix, tested only for reading
pmichaud allison: and the roadmap is immutable?
allison Coke: aye
Whiteknight allison: Okay, but there should probably be a mechanism to adjust it over time. 20:12
Coke obviously not immutable, since we just pushed several things out to the next release.
allison pmichaud: nope, we change it every week, as we review
Coke s/we/allison/ =-)
pmichaud I mean as far as "adding new items to the roadmap"
particle- critical is not the same as urgent.
allison can both add and remove tasks
particle- ^^whiteknight
allison but it is definitely a group review,
Coke pmichaud: is your primary concern not "what's on the roadmap", but "whether tickets that are important to you/rakudo get worked on?"
Whiteknight so at #ps we should be able to propose new tasks to add to the roadmap?
allison while marking a ticket for a milestone is individual, and more random 20:13
NotFound Here is a working pipe. At least the minimal pir code I tested works.
pmichaud Coke: not necessarily rakudo specific. I'm curious about when tickets make it onto the roadmap in general.
allison Whiteknight: yes, but keep in mind that they're not just random tickets, they are critical tasks in the long-term plan
pmichaud For example, I brought up the issue of concurrency support in today's #ps, specifically because it *doesn't* appear in the roadmap.
Coke I think roadmap has some special meaning here that we're not agreeing on. =-)
pmichaud is that because we think it's implemented sufficiently well? 20:14
Whiteknight allison: right,I'm not saying it will happen often, just working out the mechanism
Coke pmichaud: TT#618, for 3.0 release.
allison pmichaud: it was in the original roadmap at PDS, but was removed in a later discussion
particle- if only we had a project manager....
Coke ah.
particle-: project manager is useless without resources to /manage/
purl Hmm. No matches for that, Coke.
pmichaud Coke: TT#618 is "advanced concurrency support"
Coke pmichaud: i defer to allison.
allison Coke: ah, aye, there were two concurrency tasks, one immediate, one long-term
Whiteknight salivates at the idea of TT #618 20:15
allison the reason I know, is I saved it off to my parrotsketch notes when I removed it from the roadmap
Whiteknight that is going to be quite the fun ticket to work on
pmichaud Whiteknight: just because something is slated for 3.0 doesn't mean you have to wait until then :-)
Whiteknight pmichaud: I've got a lot of things on my near-term todo list to tackle first
allison ||||concurrency/threading||allison||||||
||||threads useful in hlls||||||||
those were on the roadmap for sometime around 1.3 IIRC 20:16
Coke IM mind, "roadmap" is 'big picture' items that are useful for someone outside of parrot know when to expect something. It's more marketing. If you're concerned about something getting done for release 1.2, that's doesn't have to be on the "roadmap".
pmichaud Coke: I'm fine with such a case not being on the roadmap. I'm not fine with such things not ever being reviewed to make sure that they're not forgotten.
allison Coke: that's an excellent way of explaining it
particle- coke: agreed.
pmichaud if you're saying that such reviews are taking place, then I'm simply not seeing them from where I sit. 20:17
particle- roadmap:product::milestone:release
allison pmichaud: yes, I agree
Coke No, I'm not.
allison pmichaud: I did the 1.1 review last month
pmichaud allison: that was done when? 20:18
allison would now, a week before the release, be a sensible time for some group review?
pmichaud (relative to the 1.1 release)
Yes, I think that'd be reasonable.
allison pmichaud: reviewed each ticket, closed the ones that were completed, rejected ones that were entirely inappropriate, gave feedback to any that needed it, moved some to 1.2, and moved others to no roadmap marking 20:19
how about right now?
purl puts on her wizard robe and hat
pmichaud unfortunately, I have to leave in a few minutes.
particle- sorry purl, this party is humans-only.
NotFound Can someone tell if the pipe fix for unix is useful? If not, stop crying about lack of working pipes }:) 20:20
allison pmichaud: okay
pmichaud NotFound: in order to be useful for rakudo, we probably need both unix and windows support.
allison I will look at every ticket marked for 1.2 this week
pmichaud NotFound: if getting it working under unix is a useful first step for that, I'd take it.
allison and, we can talk on the mailing list for any questions that need discussion 20:21
pmichaud that works for me.
NotFound pmichaud: I don't see how can we make working in general without a first test case checked.
pmichaud NotFound: I don't even know what the PIR interface for pipes looks like -- I haven't seen it documented.
Perhaps I'm not looking in the right place.
NotFound pmichaud: yeah, that's te reason why I ask for some confirmation. 20:22
pmichaud NotFound: I'd be glad to write a test case if I knew what the API is supposed to look like. :-)
allison the PIR API?
pmichaud In general, I think a ticket review two weeks prior to each release might be useful.
allison or C API?
purl C API is probably the main focus in the docs
pmichaud allison: whatever I need to write from PIR so that I can run a subprocess and capture its output.
allison pmichaud: we could add it to the release manager guide
pmichaud +1 (add to release manager guide) 20:23
nopaste "NotFound" at 213.96.228.50 pasted "This is the test case I worked in" (9 lines) at nopaste.snit.ch/16530
pmichaud I suggest two weeks because that gives us a little time to make progress on tickets before the actual release, instead of simply moving the ticket to the next release.
allison pmichaud: it won't work now, but you'd open a read-write pipe
pmichaud: open with mode flags 'rwp'
pmichaud NotFound: that test case looks fine to me. 20:24
NotFound: I can work from that.
NotFound pmichaud: it works now in linux, and probably in other *ix
Whiteknight I'm heading home now. Later
pmichaud NotFound: why not call a perl script to test it in a os-independent manner? 20:25
NotFound: we know the location of 'perl' from parrot_config
allison or, invoking parrot?
pmichaud or, invoke parrot. :-)
NotFound pmichaud: good idea. I'll try to make a test based on that, and TODO it for non-linux
pmichaud NotFound: fair enough, although it should work even for non-linux
allison a small PIR script that just prints a line to stdout?
pmichaud even just capturing the output of "parrot -v" 20:26
or "parrot --help"
or whatever.
NotFound pmichaud: hey, I've done it in less than 30 minutes, don't ask for miracles ;)
allison pmichaud: ah, excellent
pmichaud NotFound: sorry, we expect miracles here.
NotFound I don't make them for windows since a long time ;) 20:27
pmichaud Yes, getting windows to work requires extra-miracle-points.
NotFound Did we know the path to parrot from parrot? 20:28
pmichaud it's in the config
['parrot']
NotFound Ah, nice
pmichaud oh wait, that's not it. 20:29
['build_dir'] and concatenate that with '/parrot' or '\\parrot.exe'
dalek kudo: 19490ae | pmichaud++ | (3 files):
Get src/parser/grammar.pg to be utf8 again. This involves bumping
 20:30
pmichaud (the .exe suffix can be obtained from ['exe']
NotFound I think that just /parrot might work in all supported platforms. 20:31
dalek rrot: r38722 | allison++ | trunk/docs/project/release_manager_guide.pod:
[release] Add a note about reviewing the current milestone tickets
Coke parrot_config build_dir slash test_prog
particle- NotFound: please append ['exe'] 20:32
Coke test_prog should include the .exe, no?
particle- test_prog => 'parrot'
NotFound If I remember well, CrateProcess takes care of that
Coke ah. parrot_config build_dir slash test_prog exe 20:33
particle- coke++
on my system, that translates to C:/dev/rakudo/parrot\\parrot.exe 20:34
Coke wonky slashy.
particle- and yes, windows does permit that
NotFound So you see that trying to use the nicer slash is a loss of time ;)
particle- C:\\dev\\rakudo\\parrot>C:/dev/rakudo/parrot\\parrot.exe
parrot -[acEGhprtvVwy.] [-d [FLAGS]] [-D [FLAGS]][-O [level]] [-o FILE] <file>
Coke for windows, yes. 20:41
trust me, you won't want to have to back and fix that for VMS. =-)
(though really I suppose we should pass all of that to a library routine for canonicalization) 20:42
Coke updates partcl's and paraplegic's front pages to note the lack of workage. 20:45
20:46 raiph joined 20:47 raiph left
dalek kudo: fcf5d97 | pmichaud++ | src/parser/grammar.pg:
Clean up some places where latin1->utf8 conversion did the wrong thing.
 20:48
Coke pmichaud: can I give you a commit bit on the APL project? 20:49
(not that I expect you to fix anything, just that you wrote more than half of what's in there.)
pmichaud Coke: sure.
Coke googlecode id?
pmichaud I think "pmichaud@pobox.com" works. 20:50
Coke made you an owner in case I get hit by the bus. 20:51
... at least, I will have done that by the time the page refreshes. :|
pmichaud Coke++ # thanks
Coke what, i just made it look like the dead project was half your fault. =-) 20:53
pmichaud fine with me!
I'm good at creating dead projects :-)
particle- like languages/perl6
dalek rrot: r38723 | coke++ | trunk/config/gen/makefiles/root.in:
Fix some build dependencies.
 20:54
rrot: r38724 | NotFound++ | trunk/src/io/unix.c:
[cage] delete a debug statement commited by mistake
nopaste "NotFound" at 213.96.228.50 pasted "Platform independent pipe test" (29 lines) at nopaste.snit.ch/16531 21:00
NotFound Something like that?
Coke NotFound: note that 'parrot' is available as config['test_prog'] 21:01
but yes, something like that.
particle- hence coke++'s earlier comment about "parrot_config build_dir slash test_prog exe"
NotFound Oh, I don't read that carefully enough 21:02
Coke I mean, who is going to build parrot and call it "not_parrot", but hey, it's there.
NotFound And what must we expect in the test output? /This is Parrot.*/ ? 21:04
particle- NotFound: yes 21:05
there is a streams test for exactly that 21:06
dalek rrot: r38725 | NotFound++ | trunk/t/op/io.t:
[test] add a test for pipe reading, TT #661
 21:21
cnum-dynpmcs: r30 | darbelo++ | trunk/ (4 files):
- Add decnumber parameters to Makefile CFLAGS.
 21:26
darbelo Now, I can get back to *useful* work. 21:27
21:27 Whiteknight joined
darbelo hates working on multiple platforms. 21:27
Whiteknight back in my day we had to work on zero platforms, and we were lucky to do it! 21:29
darbelo I'm too used to embedded work, I guess. 21:31
Whiteknight oh really? you do embedded stuff? do tell
dalek website: allison++ | Election of Bilbo Baggins for 2009-2010 Board Term 21:32
website: www.parrot.org/content/election-bil...board-term
darbelo I study electronics engineering. Last embedded work i did was on an 8-bit microcontroller. 21:33
Whiteknight I did electrical myself
NotFound Talking about limited platforms, parrot for Nintendo DS will be a nide thing :)
nice
Whiteknight darbelo: Which controller were you using? I've worked on about a dozen of them myself 21:34
darbelo A MCS-51 derivative. It automated a spark plasma sintering experiment. 21:35
Whiteknight oh, very cool. MCS-51 I've never used 21:36
pmichaud Util: ping
Util pongs
darbelo You might know it as the intel 8051.
pmichaud Util: how far away might a pbc_to_exe fix be? Apparently rakudo doesn't build on feather (which I didn't realize)
at least, it doesn't build for some individuals
darbelo es.wikipedia.org/wiki/Intel_8051 21:39
Whiteknight oh, then yes. I had never heard that name for it before
darbelo Sorry: en.wikipedia.org/wiki/Intel_8051
dalek kudo: 3412a2f | pmichaud++ | src/parser/grammar.pg:
Enable parsing and limited processing of � quotes � .
Util pmichaud: well, I *could* commit the patch from last Tuesday, which would break the build on Win32 with MS compilers. I expect to have a breaks-nothing version in Perl (instead of PIR) in another 30 minutes, but I don't know when I can convert that version to PIR. 21:40
Infinoid darbelo: cool, 8051s are fun. I've worked with them and AVRs 21:42
AVR has a gcc port, so I like those more
pmichaud Util: I'm okay with moritz++'s suggestion to have separate code paths for gcc versus msvc 21:44
Util pmichaud: are you OK with pbc_to_exe moving from PIR to Perl5, since it is supposed to be a stop-gap anyway? 21:46
pmichaud Util: that's not really my call, but yes, I'd be okay with that too. 21:47
darbelo The atmel 8051 have less stuff on them tha others, but are good when you're on a tight budget. 21:48
Or, as was my case, when some of the peripherals MUST be phisically separate from the main board. 21:49
NotFound Someone wanting to try to implement Parrot_io_open_pipe_win32? I have some C++ code that can be borrowed from, but I don't have a windows development system and don't have plans to get one. 21:50
Infinoid I love all microcontrollers, actually. I mean... it's a processor, it runs code, it costs $1. That's just awesome.
I wouldn't want to do video rendering on one, but there are a lot of tasks its just perfect for
the available compilers do vary a lot in quality tho 21:51
darbelo Heh, the end-users wanted a screen on it. Managing the 128x64 display was something of an adventure. 21:53
Infinoid I've hooked them up to serial display controllers, but haven't tried doing anything graphics-intensive 21:54
chromatic I think moving pbc_to_exe to Perl 5 is a step backwards.
We're trying to install less Perl 5 code with Parrot.
davidfetter MOAR PERL6! 21:56
PerlJam What does pbc_to_exe do exactly? Could a parrot program do it? 21:57
moritz PerlJam: I think it just generates a .c file
which contains a dumped .pbc file, and a call to a function 21:58
PerlJam yeah, I that's what I thought too (haven't bothered to actually look though :)
moritz the problem with the current version is that it stores the dump in an array of unsigned bytes 21:59
which is a *huge* array
and takes much memory in the compiler
the better solution is to use a string instead, but msvc doesn't like strings larger than 16k
(at least constant strings)
dalek rrot: r38726 | chromatic++ | trunk/docs/book/ch03_pir.pod:
[book] Revised the second quarter of chapter 3.
 22:00
22:01 bacek joined 22:10 slavorg joined
dalek rrot: r38727 | NotFound++ | trunk/src/io/unix.c:
[cage] some cleaning of Parrot_io_open_pipe_unix
 22:17
22:22 Whiteknight joined
mj41 "Test results diff" is back on TapTinder, e.g. tt.ro.vutbr.cz/report/pr-Parrot/do?...hat+I+mean 22:25
I also setup client with Visual C++ 2008 Express Edition. See "Buld status page" tt.ro.vutbr.cz/buildstatus/pr-Parrot/rp-trunk 22:26
allison moving across town 22:27
22:36 kid51 joined
Whiteknight moritz: On Windows, the better idea would probably be to include the .pbc as a resource, instead of using an array or a string 23:00
Of course, that would be a lot of platform-specific code to write to make that happen
pmichaud I'd be fine with leaving Windows as slow-to-compile while speeding up the unix side. :-) 23:07
Whiteknight The vast majority of Windows users are not likely to be compiling these things themselves 23:09
it needs to be just fast enough for our packagers
chromatic And people building/rebuilding Rakudo. 23:11
Whiteknight If we attached the .pbc as a binary resource, the compilation on windows would be very speedy indeed 23:13
I'll have to double-check my copy of Petzold to make sure I remember how to do that though 23:14
23:20 donaldh joined
dalek rrot: r38728 | whiteknight++ | trunk/DEPRECATED.pod:
[DEPRECATED] add note about get_addr and set_addr opcodes, which will change semantics post-1.4. See TT #218 for details
 23:35
23:40 gaurav joined
dalek rrot: r38729 | whiteknight++ | trunk/src/gc/generational_ms.c:
[docs] Add some function-level documentation to src/gc/generational_ms.c. I dont know what everything does, but I took some educated guesses
 23:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!