weekly Rakudo status meetings with phase transitions: Tue 19:00 UTC | IR clogs at irclog.perlgeek.de/phasers/today
Set by moderator on 25 June 2010.
07:20 [particle]1 joined 07:44 [particle] joined 13:38 masak joined 17:57 kalve joined, kalve left 18:19 eternaleye joined 18:32 lue joined
masak \\o 18:59
lue BEGIN { o/ 19:00
masak BEGIN { ... }
dang, the meme was in the air and lue snatched it :)
lue also grabbed the EXTERMINATE phaser for later
jnthn masak terminated his properly though :P 19:01
pmichaud we're done already? ;-)
masak jnthn: lue is still a begineer :P
lue but #phasers isn't over :)
PerlJam phases in
jnthn masak: :P
masak s/nee/nne/
19:01 mberends joined
jnthn masak: Oh, I thought it was a pun on "pioneer" rather than a typo. :P 19:01
pmichaud mberends: o/
masak ENTER { mberends } \\o/
pmichaud I thought it was a pun on 'engineeer' 19:02
jnthn mberends! \\o/
lue
.oO([look at camera] Hello. Welcome to #phasers)
pmichaud *engineer
PerlJam I feel like a begineer some times
masak it wasn't a pun. just accidental Dutch.
mberends == JIT for #phasers today ;)
lue finds himself in a null PMC :( 19:03
.oO(how did I get in here....)
masak lue: let me count the ways...
pmichaud lue: check your callframes
lue the walls are made of... nothing. And yet there are still walls... Halp. 19:04
jnthn So, shall we dig in? :-)
masak let's.
pmichaud yes, let's BEGIN.
jnthn masak: So, how is your rapport with Rakudo this week? 19:05
masak in either sense, it's exceptional :P
I've done some Buf work. not a lot, but still some progress.
Buf IO "works", but there are known encoding issues, and it needs testing in general.
right now, I'm testing Kodi++'s Temporal patch. it's looking good. 19:06
moritz_ appears
mberends masak: are you on the last 20% of the Buf work yet that usually takes 80% of the time?
masak mberends: I'm surprised at how accurate the schedule I threw together has turned out to be. 19:07
mberends \\o/
masak mberends: the only deviation is that some things have come in a different order.
but so far I'm on schedule, and feeling confident about staying on schedule.
PerlJam masak++
jnthn masak++
masak I'll have to be slightly more careful in juggling my time up ahead, as YAPC::EU looms closer. I'm already feeling a slight procrastination pressure on my temples. 19:08
but mostly, life is good.
I probably forgot something that I'll remember in a bit.
.eor
jnthn \\o/
Let's stick with m. moritz_?
moritz_ sorry, distracted.. can we postpone me?
jnthn OK
mberends? :-) 19:09
masak puts moritz_ in a LATER phaser
mberends ok, not much to report.
been working on a port of HTTP::Server::Simple::PSGI
sorear hello.
mberends and it works, so that the perldancer people can build their Sinatra clone on top of it 19:10
I guess I'll have to start a seed implementation
masak mberends++
jnthn Nice 19:11
masak mberends: how does Rakudo ng feel for this task. any especially annoying bugs?
mberends not bugs on the TCP/IP side, but Buf will help to communicate binary data
Rakudo is fine as an embedded webserver. 19:12
masak acknowledged.
mberends I think HTTP/1.1 will need a bit more byte counting to do chunked transfers
but HTTP/1.0 is fine
next up I plan to work that async IO select() implementation, although it seems like to become a poll() implementation instead 19:13
poll() is the new select()
and then we can write telnet clients, irc bots etc more easily
and maybe begin a POE port 19:14
pmichaud
.oO( somehow I never thought I'd read "Rakudo is fine as an embedded webserver." :-)
mberends so poll() could be pretty useful. I'm worried about the short time to do it before R*
.eor
jnthn mberends++ 19:15
pmichaud!
mlue mberends++
PerlJam mberends: I wouldn't worry /too/ much ... there will be other R* releases
Tene mberends: I'd love to work with you on that.
mberends yeah, I should relax :-)
moritz_ right; we want to give people incentives to upgrade to the next release :-)
mberends Tene: cool
Tene I started prototyping a POE clone ages ago.
pmichaud am I up? 19:16
jnthn pmichaud: You are indeed.
pmichaud * Had a great time in Hawai'i
* Now in PDX for OSCON this week
* Fixed the bug with Q-named classes
* Fixed backtracking into subrules and capturing subpatterns
* Made some book edits and review notes
* Refactored startup and phasers a bit
* Eliminated LexPad in favor of a simple Hash
eor
jnthn pmichaud++ # nice fixes
lue pmichaud++
jnthn lue?
lue well, I tried implementing =:=, and found out we need the VAR() macro to make it work right. 19:17
that's about it :D
jnthn OK. :-)
colomon++ can't make it but sent a report. Lemme paste it... 19:18
Done this week:
Checked in trig fixes, numeric spec tweaks. Numeric grant done except
for final report.
Worried about but cannot fix:
Doubled gather / take bug. (eg trying two greps in a row, or two ... series)
Defining new operators in modules (eg Vector module issues)
Using <ABC::note> in grammars (eg ABC module issues)
Want to fix:
Series type issues (eg generating Rats when we should be generating Ints)
Range iterator (suspect switching it to series internally will break
spectest because of gather / take bug)
Modifying hyper to handle lazy arguments
Parse decimal numbers at compile-time rather than run-time
.eor
[Coke] ~~
jnthn colomon++ for grant being pretty much done \\o/
lue \\o/ colomon++ 19:19
jnthn Who else...ooh, I guess I can go. :-)
THIS WEEK
* Worked on pls portability. pls itself is now not so bad, but the makefiles it generates are about as portable as a cannon. Kinda stalled on that. Can pick it up again this week if nobody else will.
* Lots of work on error message improvements, especially in multi-method dispatch
* Fixed one more problem with subset types, though others remain
* Ripped out the Int ~~ Num hack and fixed up the tests. Not too much fallout. 19:20
* Ripped out REJECTS, which is gone from S03. Zero fallout.
* Got rid of the BEGIN special casing and updated the rest of the phasers implementation to be able to provide for BEGIN
* Did a first very basic cut of the feed operators ==> and <==.
* Switched custom operators to work in terms of BEGIN phasers too, fixing a pre-comp problem. Fear I may have introduced a new problem - got a good idea of how to fix it.
* Did a first cut of custom circumfixes
* Fixed bug with attribute initializers which led to them having the rong outer scope
* Also started to unbreak .wrap and .unwrap a bit
* Oh, and got myself a shiny new blog too. 6guts.wordpress.org/
* Still thinking through various future meta-model-y things for post-R*, though focused more on R* things this week, and likely will be the same up until R* is released.
* Hacked on Test::Mock. Quite happy with it.
NEXT WEEK
* Fix more bugs in the run-up to R*. Suggestions welcome.
* pls can haz portability?
* Look at the distribution a bit too as well as the compiler
* Procrastinating for YAPC::EU
WORRIES
* Very concerned still about the role outer chain bug.
.eor
PerlJam jnthn++ 19:21
[Coke] I've done nothing! (but am still planning on releasing Atlanta on Thursday.)
jnthn :-)
moritz_: Readier?
moritz_ yes
jnthn Go!
moritz_ DONE
* usual patch applying, testing
* experimented with s:g///. Got a prototype working 19:22
* pondered my possible Hague grant application
WORRIES
* the book is not a coherent whole
* we haven't done much wrt distribution setup for Rakudo *
GENERAL_PANIC
.eor 19:23
masak is the lack of bullets under "GENERAL_PANIC" good or bad?
PerlJam moritz_: was "GENERAL_PANIC" a directive? :)
pmichaud I think he was my commanding officer at one point. :)
moritz_ masak: it hasn't been shot yet. That explains the lack of bullets 19:24
jnthn :-)
Who did I miss who has something to say?
ash_ has anyone heard from cygx on try.rakudo.org?
lue [ pmichaud: I had him as a Colonel :) ] 19:25
ash_ because i am about to start working on the back end, since i haven't seen any commits from him in a while
moritz_ ash_: no, haven't seen him
ash_ well, i should have time this week to try to get more of try.rakudo.org working, so hopefully that will be useful to someone 19:27
jnthn ash_: Sounds good. :-)
pmichaud (will be useful to me )
moritz_ btw feather3 is now a physical host, not a vm
Juerd++
Tene Theoretically, I've been recovering a bit lately, so I might do something this week, but I'm not very hopeful, as I've said that before.
PerlJam Do we have a press date for the book from chromatic? Or do we need to tell him when? 19:28
sorear hasn't done much at all with Rakudo lately
pmichaud chromatic said (iirc) that he needs about 2 weeks notice from when we think we're ready until he can have ready for print
but we definitely need some more review on the book (see my notes added last night) 19:29
masak would like to have a projected press date set
PerlJam indeed.
pmichaud yes, a press date would be good.
I expect there to be another Rakudo Star release in August, perhaps that's a good date.
PerlJam pmichaud: how soon after R* do you think the "lexical regex" problem could be resolved? Maybe we just don't change the book
sorear chromatic is on IRC and non-idle this very minute. I just pointed him at #phasers
TimToady maybe just a temp footnote saying "for now you have to write <foo=&foo> 19:30
pmichaud PerlJam: we still need some specification work
PerlJam: I wouldn't plan on it being ready in August. Lots of other things to do first.
PerlJam okay
pmichaud and I expect we'll have a ton of "little bug reports" immediately following R*
jnthn Aye, I expect so too. 19:31
TimToady and we're really talking full-fledged implicit grammar derivations here, "slangs"
pmichaud right.
jnthn Aye, those are some work.
pmichaud I might be able to cheat a bit in the meantime, along the lines of "regexes declared outside of grammars look locally" 19:32
19:32 chromatic joined
pmichaud but I don't want people building a lot of code that we know will break. 19:32
chromatic Book press date: whenever we can
pmichaud book press: I tentatively propose Aug 26
PerlJam When's YAPC::EU?
jnthn Start of Aug
pmichaud (which is the likely month + 1 release date for Rakudo Star)
YAPC::EU is Aug 4-6
Rakudo hackathon on Aug 3
19:33 colomon joined
moritz_ btw we could get a room for the hackathon for ~150 Euro per day 19:33
from the hotel
masak \\o/
chromatic The book will likely take two weeks from submission to print.
moritz_ chromatic: does that include proper typesetting?
pmichaud so, we'd need to have the book in a "done" state by Aug 12
moritz_ or do we have to tweak the latex ourselves?
jnthn pmichaud: That's quite ambitious given the next week or so will be quite tied up with readying R*, and YAPC::EU will happen after that, which many of us will also be at. 19:34
lue like LaTeX
pmichaud jnthn: agreed, it's ambitious. If too ambitious, let's postpone. 19:35
chromatic The only typesetting I've ever had to tweak is linebreaking.
19:35 mikehh joined
chromatic I haven't had to work with a lot of Unicode in any of our books before though. 19:35
pmichaud jnthn: I'm tossing out proposals for people to shoot down. :)
moritz_ chromatic: I'm also talking about positioning floats, selecting fonts etc.
PerlJam pmichaud: we're a too agreeable bunch :)
[particle] PerlJam: agreed. 19:36
chromatic moritz_, that's all doable, but the earlier the better.
pmichaud PerlJam: then consider this an application of "the veto rule"
jnthn pmichaud: It's only a data point, but at a personal level, I'm highly unlikely to find much time between now and then. I know that there's a lot to do between now and the R* release date, I know that I won't get much in the way of writing done at YAPC::EU 'cus there's people to meet, talks to see/give, beer to drink, etc. After YAPC::EU I've got a few days offline doing family bits. 19:37
PerlJam chromatic: and what about general proofing? Will it be submitted as PDF? Who generates the PDF? (where my PDF might look different from your for some reason)
moritz_ chromatic: well, doing that early doesn't make sense, because when we tweak the text we destroy the layout again
pmichaud jnthn: ...propose a different date?
PerlJam What's the next big perl conference after Aug 26? 19:38
pmichaud jnthn: would it make more sense to say that we have the final copy ready on 26, and print is two weeks after that?
chromatic I mean font selection and some of the other stuff.
PerlJam maybe we could should for a date that falls just before a conference :)
jnthn pmichaud: That would certainly allow me a lot more chance to look at it.
chromatic I can generate the PDF, but if someone else wants to do it and suggest changes and submit it for proofreading, that's good too.
jnthn pmichaud: Particularly given I can focus on book rather than an iminent first distribution release. :-) 19:39
pmichaud anyway, we can (and should) continue with monthly book PDF releases 19:40
jnthn Yes, +1 on that.
PerlJam pmichaud: agreed
moritz_ we haven't sticked to those for quite some time
pmichaud I vote we try to have to have something in a nearly complete state by August 26
moritz_ but if somebody wants to throw out a release, feel free
chromatic The last wc I did showed about 25k words, or 100 pages.
180-200 pages seems right. 19:41
jnthn pmichaud: I think I've a chance of being able to contribute some decent amount before that date.
*a very good chance...
If it works for @other, works for me.
pmichaud any objections?
PerlJam none here
pmichaud moritz_: I'll cut a release for July. 19:42
moritz_ pmichaud: +1
masak pmichaud++
jnthn Between now and R*, I figure there's a few things to focus on. 1) Must-fix bugs. 2) Preparing the distribution release and making sure it installs on various sytems. 3) Preparing release announcement and other related things. 19:45
At this point, can we identify which things fall into "1" and who can tentatively look at them? 19:46
(And same for 2 and 3.)
colomon bugs must be fixed by this thursday?
jnthn That's largely down to how strongly pmichaud++ feels about the Rakudo in R* being the Thursday release. 19:47
mberends jnthn: I know you know 1) Zavolaj stopped working some time around 3rd July, affecting MiniDBI
moritz_ that's something that should really be fixed
jnthn mberends: I'll get to that one either tonight or tomorrow at latest. 19:48
PerlJam jnthn++
pmichaud I'll do 3 soonish.
(as in, before thursday)(
my plan has been to basically work on 1 and 2 until next thursday
jnthn Yes, having osme time to review 3 would be good.
moritz_ disappears for a while 19:49
PerlJam is the release going to be prepared in the rakudo/star repo? How are the various things like MiniDBI going to be integrated? submodules?
jnthn PerlJam: There's a script in there that, afaict, fetches all of the pieces.
PerlJam ah. Clearly I haven't looked in a long while.
oh, no. I just forgot :) 19:50
jnthn pmichaud: I'll take on the native call thing. What are our other things that fall into "1"?
PerlJam jnthn: does blizkost need any love or is it working well? 19:51
jnthn pmichaud: The role outer fix and the chained gather iterator ones seem amongst the most serious.
colomon +1
jnthn PerlJam: Last time I tried it, it was. I can give it another look.
pmichaud I thought chained gather iterator was fixed already. Do we have a golfed example I can work on?
jnthn pmichaud: See #perl6 19:52
Just fed one to the evalbot
It's the double-grep
pmichaud okay, that should be a very easy fix.
jnthn Or in general, gather/take back to back
I hoped you'd say that. :-) 19:53
pmichaud I didn't realize it was still a bug.
colomon or try doing two series with a Zop
jnthn pmichaud: On the role outer bug - which of us should take that? Should I, and you can focus on distribution issues a bit more, or do you want it, or? 19:54
colomon I'll try to write some tests for it when I get home
jnthn pmichaud: I figure it's potentially going to take whoever takes it on quite some effort.
pmichaud: Given we've both spent some time on it so far without reaching a fix. 19:55
pmichaud jnthn: we still have the problem of the MAINLINE not seeing the correct outer.
that needs to be resolved first, and then see where we are.
jnthn pmichaud: Right, that's a pre-cursor to it.
pmichaud jnthn: so, I'd say whichever of us gets to it first :) 19:56
jnthn OK.
pmichaud once we have that working, then I think the role outer bug is really better suited for you to fix.
but that precursor has to be taken care of, and either of us are equally well situated to look at it
I suspect I can find the bug with a few hours' work
I just didn't have the hours until this week :)
jnthn Understand. :-) 19:57
19:57 eternaleye joined
pmichaud let's ask a slightly different question. Does anyone feel there are any outstanding bugs that should block an R* release if not fixed by the 29th? 19:59
(I say "no.")
masak doesn't
PerlJam those two are the only big ones that I can think of.
jnthn I'd be uncomfortable releasing with the two I just mentioned. They're quite painful to work around.
Beyond that, I have none.
masak right, I also meant beyond those two. 20:00
pmichaud I think it's safe to assume that gather/take will be fixed (before Atlanta, even)
jnthn asn't sure if the question was taking those as given or not
PerlJam wiki.github.com/rakudo/rakudo/whats...nto-rakudo mentions and installer that's not-proto. Does pls qualify?
pmichaud I have no idea about the role outer bug -- that's really outside my control.
PerlJam s/and/an/ 20:01
masak we could ship with the role outer bug. it would become a FAQ on #perl6, but generally people would understand.
pmichaud (at least, once I get the main outer fixed)
masak and there's quite a simple workaround.
PerlJam masak: simple buy yucky
s/buy/but/
masak PerlJam: all workarounds are yucky.
jnthn I'm fairly confident I can implement a "bandaid" style solution if it comes to it.
pmichaud if we assume there are no blocker bugs, then from there it's a matter of finding the bugs that cause us (or R* users) the most pain and fixing those as well.
masak PerlJam: they don't get much *less* yucky than that. 20:02
jnthn Not quite the right fix, but righter than having everybody write workarounds.
Anyway, we'll see what comes.
pmichaud anyway, I'm hoping the mainline outer will help there.
jnthn pmichaud: I hope so too.
pmichaud also, the spec for YOU_ARE_HERE just changed, so need to look at that a bit.
jnthn Yeah, I noticed the change, didn't have change to catch onto the implications yet. 20:03
pmichaud I do expect that R* will have monthly releases for the first few months.
to keep a tight cycle on bugfixes. that only makes sense.
20:03 mberends left
colomon +1 20:03
pmichaud (I've even toyed with the idea of biweekly releases)
20:04 mberends joined
masak pmichaud: that does not make sense if stars are based off monthlies :) 20:04
[Coke] I would not plan on biweekly. lot of churn.
jnthn pmichaud: +1. I expect to be digging into big deep changes, but they won't find their way into Rakudo in the first couple of months after R*.
pmichaud: And even when they do, they shouldn't cause regressions. 20:05
moritz_ chromatic: one more thing... which paper size the book going to be printed to?
s/to/on/
jnthn pmichaud: And being responsive to bugs is a Good Thing.
pmichaud masak: (monthlies) right, which is why I've kept to "monthly" for now
but if we move off of monthlies, then biweekly works for a short period of time.
masak pmichaud: I'm also skeptical about bi-weekly.
pmichaud okay, we'll keep monthly. 20:06
maybe we'll make sure to have a way for people to get R* with the latest compiler
jnthn masak: What's left before we can s/proof_of_concept/pls/ ? 20:08
masak jnthn: same as after the Lund hackathon: testing that more/all projects can be installed; nice packaging. 20:09
the 'packaging' wording includes, but is not limited to, a Perl 5 wrapper script, and README/documentation.
jnthn eek, urgent errand, brb 20:10
(5 mins)
masak I'll be happy to guide people through testing it more, and I hope to find the tuits to write wrapper/documentation myself.
PerlJam Should STD be part of R* ? 20:12
moritz_ no 20:13
PerlJam why not?
moritz_ because it has quite heavy demands on Perl 5 modules, and is often not what people expect 20:14
masak a person coming to Rakudo for the first time will probably neither think of reaching for STD, nor think to look for instructions to do so. 20:15
TimToady
.oO(nobody expects the STD inquisition...)
lue We have one thing, that's parsing and error messages. Okay, TWO things.... 20:16
PerlJam Who is R* for?
I imagine that there are developers out there who would download and use R* and would benefit from some extra help in understanding what went wrong when things blow up 20:17
[particle] that's what #perl6 and p6c is for 20:18
TimToady
.oO(STD is just chopped liver :)
PerlJam [particle]: is there some reason only one path to enlightenment is preferable here? 20:19
[particle] STDs are dangerous
lue TimToady: at least STD doesn't rely on the Norweigan Blue, lovely plummage.
colomon is p6c really appropriate? 20:20
[particle] we shouldn't promote it as something that *everyone* should take a look at
PerlJam ah, so back to my other question ... who is R* for?
[particle] R* is for everyone, the lowest-common-denominator being that they have some excitement about perl 6 and want to try it
jnthn back 20:21
[particle] at least, that's what the first release is for
or first few releases.
masak PerlJam: I have a feeling shipping STD will Rakudo will be quite a bit of work with relatively little benefit to people.
PerlJam masak: I kind of have that feeling too, but I thought I'd throw the idea out there anyway.
TimToady it's just a CPAN module :)
PerlJam TimToady: and that's why I even thought about it. 20:22
jnthn Right, people can install it from CPAN now too.
PerlJam okay, so the "documentation" that comes with R* will say, "Have fun! If you need help, ask on #perl6" or something
masak yay 20:23
[particle] yes, or it will point to rakudo.org, which will point to #perl6, p6u, p6c, etc
PerlJam right.
jnthn OK, any more for #phasers today? 20:27
pmichaud are we about done with #phasers?
I'm going to head over to OSCON and get my registration stuff
and then make another attempt at nom-fetching
masak END { #phasers }
lue DESTROYALL { ... }
jnthn o/ 20:28
chromatic moritz_, 6x9 I think, the same size as the PIR book. 20:30
moritz_ inches? 20:32
chromatic yes 20:33
masak counts on fingers
chromatic Hm, maybe that was 5x8.
\\usepackage[centering,width=3.75in,height=6.75in,includeall,hcentering,vcentering,marginparsep=0in,marginparwidth=0in]{geometry}
moritz_ uh 20:34
that's even smaller than A5
is that a good choice for a book with many code listings?
chromatic Maybe, maybe not. If it looks bad in PDF at that size we can go larger. That's no problem.
moritz_ great 20:35
21:13 masak left 23:16 colomon joined 23:32 mberends joined 23:54 [Coke] left

Powered by Cro and the Raku® Programming Language.

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

Thank you!