pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
TimToady moritz_: no particular speed goal for STD5 other than "fast enough to bootstrap something else" 04:03
most of the slowness comes from using almost none of the built-in Perl 5 regex engine 04:04
so presumably with better builtin support of P6 regexes we can get to something approaching the speed at which gimme5 parses STD.pm 04:05
alternately, we can speed up STD5 by doing more aggressive ratchet optimizations to discover more spots that can map directly into P5 regexen
but the overriding goal of STD5 was to implement correct semantics rather than fast 04:06
gah, the firewall here doesn't allow PROPFIND, so I can't get at svn directly... :( 04:10
rakudo_svn r28991 | pmichaud++ | [rakudo]: 04:46
r28991 | pmichaud++ | * Clean up placeholder variable actions.
rakudo_svn r28992 | pmichaud++ | [rakudo]: change 'process_contextualizer' to 'contextualizer_name' 05:07
rakudo_svn r28995 | pmichaud++ | [rakudo]: methods should be .blocktype('method') instead of using pirflags 05:29
pugs_svnbot r21171 | moritz++ | [spec] unfudged a passing test, autounfudge++ 07:39
pugs_svnbot r21172 | moritz++ | [t/TASKS] add todo for junction tests 11:25
r21173 | moritz++ | [t] statements/for.t: major cleanup 11:40
r21174 | moritz++ | [spec] moved for.t to spec/, updated POD 11:43
r21175 | moritz++ | [spec] fudged for.t. Fails with 'Null PMC access in type()' 11:52
pugs_svnbot r21176 | moritz++ | [spec] enhanced index.t a bit 12:13
rakudo_svn r29006 | moritz++ | [rakudo] implement Str.index, patch curtesy by bacek++ (Vasily Chekalkin) 12:15
r29007 | moritz++ | [rakudo] add S29-str/index.t to spectest_regression. 12:21
r29007 | moritz++ | bacek++ for implementing
rakudo_svn r29008 | pmichaud++ | [rakudo]: coding stds in src/classes/Str.pir . 12:49
pugs_svnbot r21177 | moritz++ | [spec] improved fudge messages in log.t 12:56
r21178 | ruoso++ | [SMOP] p6opaque delegates method calls to the how (still not working)
r21179 | ruoso++ | [SMOP] p6opaque now uses sm0p block, so needs to be renamed
r21180 | ruoso++ | [SMOP] updating Makefile.am 12:59
ruoso interesting... it's the first time I write a long block of sm0p code that simply works at first attempt 13:01
moritz_ ruoso: that's called "practise" ;-) 13:02
ruoso it seems that p6opaque is succesfully dispatching non-repr calls to the metaclass 13:03
pugs_svnbot r21181 | moritz++ | [spec] fixed complex log10() tests 13:05
rakudo_svn r29010 | pmichaud++ | [rakudo]: spectest-progress update: 83 files, 1365 passing tests 13:36
ruoso @tell pmurias, can you please provide me some guidance on how to use your sm0p parser? the kp6 parser is too much slow, and p6opaque already have two blocks... so it's starting to be a real pain.... 13:38
moritz_ wohoo - that makes our graph look quite nice
lambdabot Consider it noted.
pmichaud www.pmichaud.com/perl6/rakudo-tests...-07-03.png 13:40
moritz_ or on rakudo.de/
lambdabot Title: Rakudo - A Perl 6 Compiler 13:41
moritz_ ruoso: it seems that pmurias either forget to commit a file, or his 'brew' script is wrong 13:55
ruoso hmm 13:57
sm0p.pl takes forever to parse p6opaque.sm0p now that it has two blocks
moritz_ (I'm talking about misc/sm0p/ )
ruoso moritz_, I know... 13:59
$ (cd src; time perl ../sm0p.pl .. p6opaque.sm0p p6opaque.c) --> real 1m6.884s, user 1m6.360s 14:05
gah
moritz_ that's a good reason to switch to elf ;) 14:06
pugs_svnbot r21182 | ruoso++ | [smop] some more work in p6opaque, delegation almost working... I need to finish the "setting how" method, so I can effectively test the delegation 14:08
Ontolog Which is more complete Rakudo or Pugs? 14:15
SamB isn't there some kind of test status comparison page which would allow you to form your own answer to that question? 14:16
vixey Pugs 14:17
moritz_ Ontolog: pugs
Ontolog Thanks
moritz_ pugs passes well over ten thousand tests in the official test suite, rakudo 1300 or something
Ontolog Is the Perl 6 specification still being worked on? 14:18
moritz_ yes
Ontolog or is everything now just to be implemented?
i see
SamB and pugs spawned the testsuite, didn't it?
PerlJam Well ... to be fair, the "official test suite" was written for pugs originally, so there may be bogus tests in there since the design has changed
pmichaud implementation issues often cause minor tweaks to the specification
and I always think of the "official test suite" as t/spec
(which doesn't have 10,000 tests in it yet :-) 14:19
DietCoke PerlJam: that just means that when rakudo overtakes pugs, it will do so violently.
Ontolog haha sorry i started a flame war
moritz_ PerlJam: yes, that why said "well over 10k", not "roughly 17k" ;)
pmichaud moritz_++
DietCoke ah, I like pmichaud's way of thinking. =-)
moritz_ Ontolog: no flame, just discussion
DietCoke we're all friends here. =-)
pmichaud Ontolog: we're still clarifying terms among ourselfs
*ourselves
Ontolog Is there a CPAN for Perl 6? 14:20
pmichaud SamB: yes, pugs spawned the test suite
moritz_ not yet
vixey yes CPAN
PerlJam Ontolog: yes, it's called CPAN :)
Ontolog haha
moritz_ vixey: where?
Ontolog are you serious?
SamB cpan.org isn't it ?
Ontolog shouldn't there be like a seperate CPAN just for perl 6?
vixey moritz_: There's no Perl 6 yet.. but when there is it will work with CPAN
SamB why?
moritz_ SamB: I see no perl 6 modules on cpan (only perl 5 moduls named Perl6::)
SamB doesn't it just have a seperate namespace or something?
vixey Ontolog: that sounds like a bad idea 14:21
SamB oh.
pmichaud we'll undoubtedly need some way to distinguish perl 6 modules from perl 5 ones
moritz_ current cpan.org and CPAN.pm doesn'T have a way to tell which language a module is written in
SamB so that will need some redesign
moritz_ Ontolog: the plan for CPAN6 is to allow packaging of modules in arbitrary languages
SamB and those Perl6:: modules are going to confuse matters
oh, so CPAN6 will replace Debian?
pmichaud goes back to working on implicit vars 14:22
PerlJam SamB: no.
moritz_ SamB: no
SamB hehehehe
moritz_ SamB: Debian is mostly there for binary distribution
SamB Gentoo?
Ontolog I see, so there will indeed be a separate CPAN :p
PerlJam Assuming everything works as we intend, CPAN will likely keep it same name, only the P will change meaning
moritz_ Ontolog: well, we don't know yet if we morph existing CPAN into what we want it to be, or rewrite it and embrace the old stuff
SamB but it will probably be hosted on cpan6.org
moritz_ SamB: both Debian and Gentoo are OS-Specific 14:23
Ontolog ok here's a better question, if I were to write a Perl 6 module that I wanted to share, where would I put it?
SamB sadly so :-(
PerlJam Ontolog: CPAN.
moritz_ Ontolog: currently in the pugs repo
SamB Ontolog: in a public repository
moritz_ Ontolog: that's where other Perl 6 modules live as well
SamB which can be contributed to easily
PerlJam Ontolog: or anywhere else you like. 14:24
SamB maybe pugs repository needs a rename?
moritz_ SamB: why?
SamB if it has everything perl6
moritz_ SamB: everybody(tm) knows it hosts much more than pugs ;)
SamB how about perl6fun
pmichaud looks at perl6fun.org
moritz_ not found
GO GRAB IT! 14:25
SamB grr
why did verisign have to go and give my ISP ideas???
% host perl6fun.org 14:26
perl6fun.org A65.200.200.35
perl6fun.org A63.251.179.51
moritz_ SamB: I hope you didn't use a browser? that's what 'whois' is for
SamB: ok, that sucks mightily
moritz_ Ontolog: if you want to put something perl 6 related into the pugs repo, you can easily get a commit bit 14:26
SamB moritz_: indeed, you would be hard pressed to avoid one 14:27
Ontolog moritz_: I was just thinking today about doing something with Perl6 14:34
moritz_: but I'm not sure what haha
moritz_: are you guys at the point where you are writing Perl6 modules to implement parts of the standard or is the whole thing in Haskell? 14:35
moritz_ Ontolog: pugs is written in haskell, but has a prelude (ie standard library) in Perl6... 14:36
PerlJam Ontolog: As far as I'm concerned, rakudo is perl 6 and pugs is an interesting prototype :)
pmichaud rakudo is currently writing the major components of the compiler in Perl 6
moritz_ Ontolog: but pugs isn't being worked on
pmichaud: except runtime.
Ontolog hmm
pmichaud moritz_: I don't consider the runtime as part of the compiler :-)
but yes, we'll be writing the runtime components in Perl 6 soon also
Ontolog but there is nobody in #Rakudo
moritz_ which is slightly ironic, because that's where most compilers start to allow the usage of the native language 14:37
pmichaud the rakudo folks hang out here or on irc.perl.org/#parrot
moritz_ Ontolog: that's because rakudo is being discussed either here or on #parrot on irc.perl.org
pmichaud we don't need our own channel :-)
Ontolog haha
but how am I going to write Perl 6 modules if Rakudo doesn't even implement most of the standard? 14:38
pmichaud ...doesn't even implement most of the standard?
moritz_ use the subset that is already supoprted
Ontolog moritz_: haha ok good idea
pmichaud: err maybe i'm wrong sorry
moritz_ for example most object orientation features are supported 14:39
PerlJam Ontolog: how are you going to write Perl 6 modules if <insert any perl 6 compiler> doesn't implement most of the standard? ;)
pmichaud note that it's difficult to measure how much of a standard is implemented based on the size/composition of the test suite
Ontolog what is Rakudo written in?
pmichaud the compiler is Perl 6
the runtime libraries are PIR
PerlJam pm: it's difficult to measure anything when everything is in flux :)
pmichaud (so far)
Ontolog PIR?
pmichaud PIR is parrot's low-level language 14:40
Ontolog shit
is it easier than Haskell?
moritz_ Ontolog: a variety of languages, which include NQP (not quite perl6), Perl 6 rules, PIR, C
pmichaud *much* easier
moritz_ aye
pmichaud and long-term we'll only be using PIR for the primitives
i.e., those things that can't be easily or efficiently written in Perl 6
Ontolog but moritz_ said C 14:41
which parts are written in C?
pmichaud there is some C, yes -- for very low-level types
pmurias ruoso: ping 14:41
moritz_ Ontolog: the so-called "PMCs" or "Parrot Magical Cookies", which are parrot extensions
Ontolog haha, you said PMCs 14:42
pmichaud the only pieces written in C at the moment are Scalar containers and the string-to-number conversions
Ontolog has been playing MGS4 recently
I see, so the C parts are all done right?
pmichaud well, there will undoubtedly be more as we try to improve efficiencies, but I'm trying to keep as much out of C as I can
PerlJam Ontolog: nothing is "all done" yet :)
moritz_ Ontolog: considering that parrot is mostly written in C, and it's our runtime, I wouldn'T say "all done" ;-) 14:43
Ontolog The special thing about the Parrot is that it can convert between one language and another right?
pmichaud Parrot has a lot of special things :-)
but it's intended to be a runtime environment for multiple dynamic languages
so, it doesn't convert between languages as much as it allows them to run together 14:44
Ontolog it's like a JVM right?
pmichaud yes, but more dynamic
moritz_ yes and no
Ontolog why not just compiler Perl6 code to JVM byte code?
pmichaud because JVM isn't dynamic enough :-)
moritz_ Ontolog: because JVM doesn't suppport dynamic types very well
PerlJam This sort of "education session" would make a good perl.com article (or something with wide distribution)
pmichaud PerlJam: ...write it! 14:45
Ontolog haha
PerlJam pm: wish I had the time to.
moritz_ Ontolog: and no continuations and stuff
Ontolog Just post this IRC log
moritz_ I could make a summary of it
pmichaud yeah, the IRC log itself is pretty good
Ontolog asks good questions in the right sequence :-)
Ontolog haha thanks 14:46
pmichaud goes back to implicit vars. 14:47
ruoso pmurias, I'm going for lunch now... I should be back in a hour, maybe a little more... 14:48
pmurias ok 14:49
[particle] pmichaud: can't you take an .implicit_lunch? 14:52
pugs_svnbot r21183 | pmurias++ | [sm0p] fixed brew
DietCoke is sure some intrepid soul will make a perl6 on JVM. 14:54
ruoso DietCoke, the problem with that is that it will perform worst than Java... because it will be a runtime on top of another... 15:00
ruoso now really going for lunch 15:01
Juerd One of feather's hard drives just failed 15:12
moritz_ D'oh
Juerd So I'm going to replace it quickly. If you have open files, save them :)
(feather doesn't have hot swappable hardware)
SamB Juerd: shutdown does display a message ;-P 15:13
moritz_ SamB: it's nice to inform us in advance nonetheless ;)
SamB true, true
after all you might be wasting time on IRC rather than actively editing 15:14
but then, that's what autosave files are for
Juerd SamB: I know, but from my experience, several people pay more attention to IRC than to wallmessages.
And here, people who aren't logged in can also see what's going on. 15:15
After all, you will see several people timeout after a while :)
Or connection reset by peer, etc.
Ontolog shit 15:18
I just installed parrot and I get: parrot: error while loading shared libraries: libparrot.so.0.6.3: cannot open shared object file: No such file or directory
what did I do wrong?
moritz_ Ontolog: can you run it locally? (ie ./parrot) 15:19
[particle] Ontolog: you installed parrot 15:24
Ontolog yes did make install 15:25
after make test and and make and perl Makefile.PL
I will try to run it locally hold on
same result
[particle] and you read what it says, about install not working well
and you must run make reallyinstall if you really want to install it 15:26
Ontolog i did
run reallyinstall
[particle] well, ok then.
Ontolog i read that it said if you want to develop parrot on the same machine it doesnt't work well
yes so what is the biff?
why can't it find the librarY?
[particle] hrmm, maybe that message is confusing, then
Ontolog /usr/local/lib does indeed have libparrot.so.0.6.3
[particle] we parrot developers run parrot from the build directory 15:27
none of us install it
Ontolog haha
[particle] if you want to have an installed parrot, and develop parrot, you're asking for trouble
Ontolog well im not so much at the "wanting to develop parrot" stage yet
i'm more at seeing wtf it is all about
is there a make uninstall? 15:28
[particle] no
hence the problems
Ontolog shit! you guys screwed me! I'm suiing
sueing!
haha
yeah running parrot in the build directory works
what a load of crap
how do I find and delete everything parrot under /usr/local? 15:29
I will run it from the build like you guys
[particle] i'm a windows guy, mostly, so i won't be much help with that
Ontolog sorry my what a load of crap statement not about parrot about my rushing into installing it haha
a windows guy working on Perl 6? 15:30
hehehe
did you every try linux (or bsd or some such)?
[particle] i have at least seven different linux virtual machines 15:30
Ontolog i see
i like to run linux on the host and windows as a guest 15:31
[particle] i've got BeOS running, too... well, Haiku anyway
Ontolog so your not exactly "a windows guy" :p
DietCoke ... you have to ignore a warning about not installing to install it. =-)
Ontolog DietCoke: that warning was about not installing it and developing it at the same time!!!!!!!!!!111!!!1!!
DietCoke catches up.
Ontolog it's cool I have the list of all the files that got installed 15:33
and I will delete all them shits
masak I've been thinking 16:04
often enough I want to access adjacent elements in for loops 16:05
it bugs me that I have to resort to an index variable just to acheive that
masak is there a way I could get around this using Perl 6? 16:06
Auzon for @array -> $a, $b {}
masak Auzon: well, then I get them pairwise
that's not exactly what I want
cjfields is that working with Rakudo? 16:07
Auzon Oh, processing multiple arrays in parallel?
cjfields yes 16:07
masak Auzon: that could be the solution, if I could shift the same array out of place one step
Auzon I know where it is. Give me a minute
masak I'm thinking something along the lines of for @a -> $e { ... $e.prev ... }
Auzon oh. hm.
masak but that's syntactically confusing, because .prev is more of a macro than a method on $e 16:08
Auzon Let me check S04
moritz_ masak: my @x = ...; my @y = @x; @x.shift; for @x Z @y -> $a, $b { ... }
masak moritz_: yes. that works 16:09
moritz_ or better, for @x Z @x[1..*] -> $a, $b {..} 16:10
masak now we're talking 16:10
moritz_ ;)
Auzon There we go, that's like what I was thinking.
only note is that the final element of @x is repeated in the second array
moritz_ is it? 16:11
perhaps @x Z @x[1..@x.end] then
Auzon Or maybe that was with ranges instead
masak likes Perl 6 16:11
moritz_ ranges don't sound Haskellish enough for that sort of problem ;) 16:12
Auzon ok. zip stops as soon as the first one is exhausted 16:14
moritz_ the first item in a smartlink, does it have to be =head1? or an arbitrary =head\d ? 16:14
Auzon roundrobin can be used to avoid this issue
moritz_ after the L<S\d\d/ that is ;)
Auzon moritz_: I think it can match anything in the table of contents. 16:16
moritz_ Auzon: ok, thanks
moritz_ Juerd++ # brining feather back up in about no time 16:22
moritz_ even commitbit works again 16:22
Juerd moritz_: I'm confused that you call this no-time 16:26
moritz_ Juerd: I did something in between, so the time had passed switfly ;-) 16:27
Juerd moritz_: So did I - we had a small power outage in one of our breaker/fuse groups 16:27
The one that has feather, so the shutdown+powerup was nicely done around that. No suffering on feather, but plenty on 2 other machines. 16:28
moritz_ wait. Where did our precious pugs_svnbot go? 16:40
Auzon Good question
I was just checking :)
moritz_ maybe I should start one on timtowtdi.org
Auzon Who runs it normally? 16:41
moritz_ SLAVES says it's diakopter
but istr that he's on vacation
Auzon If Rakudo had sockets, I'd offer to make a pure Perl 6 one :P 16:42
But yeah, we can start one on timtowtdi.org
moritz_ Auzon: do that. After you soc project ;-)
s/you/your/ 16:43
pugs_svn r21183 | pmurias++ | [sm0p] fixed brew 16:43
r21184 | Auzon++ | [gsoc_spectest] reorganization and smartlinks for filetest.t
r21185 | moritz++ | [t] documentation updates:
r21185 | moritz++ | * first shot at t/HOWTO
r21185 | moritz++ | * updated t/README
r21185 | moritz++ | * mentioned Junction oddities in t/deprecated-syntax.pod
r21185 | moritz++ | * deleted an empty directory
r21186 | Auzon++ | [gsoc_spectest] fixed broken smartlink, Auzon--
Auzon thanks moritz_++ :) 16:45
moritz_ any idea why dev.pugscode.org/changeset/21186 is empty? svn diff shows it correctly 16:46
lambdabot Title: Changeset 21186 - Pugs - Trac
Auzon It's a whitespace change
apparently. 16:47
All I did was bang on the link until it works
moritz_ oh yes, and I had the "ignore whitspace changes" box ticked
Auzon To be fair, I didn't see the change when editing either :P
moritz_ it could be freakin' unicode non-breaking space 16:48
moritz_ and once you did a copy & paste it worked 16:48
Auzon I had to copy/paste from ack's output, though. My browser didn't copy it the same
moritz_ (Auzon's browser)-- 16:49
Auzon IWBNI whitespace didn't matter when matching
pugs_svn r21187 | Auzon++ | fixed some POD errors 16:50
moritz_ I think that double quotes causes them to be matched more liberally
Auzon Interesting 16:51
moritz_ don't know if it actually works ;)
Auzon: thanks for the fixes 16:52
Auzon: what do you think of t/HOWTO?
Auzon I like it :)
moritz_ it should be enhanced a bit
Auzon Yeah, I'll see if I can find anything to add 16:53
moritz_ biggest problem is that I don't know how much knowledge from the reader I can assume
I planned to add something about putting related test in blocks to ease fudging
but I didn't have the motivation to explain fudging ;-)
anyway, ISTR that pmichaud wanted to blog about something similar, maybe it's a start for him 16:54
Auzon Did you read his post about how the test suite improved significantly recently? We got some nice press :) 16:55
moritz_ aye, I did ;) 16:56
Auzon I think what he talked about was a followup to that post 16:57
moritz_ if you look at that progress graph, the yellow curve is our doing
(mostly)
the green one is combined effort of testers and implementors
masak moritz_++ Auzon++ 17:01
moritz_ @karma Auzon 17:02
lambdabot Auzon has a karma of 13
moritz_ lambdabot: you lost 17:03
perlbot: karma Auzon
perlbot Karma for Auzon: 106
moritz_ much better
irclog.perlgeek.de/perl6/2008-05-22#i_306074 17:04
lambdabot Title: IRC log for #perl6, 2008-05-22
Auzon Yeah, I remember that ;)
moritz_ Auzon: do you happen to have German ancestors? your name sounds a bit like it ;) 17:09
Auzon Yep, I do
Auzon I don't know if I still have close relatives in Germany, but I do have some in England. 17:10
(where close is roughly 4 steps)
moritz_ ;)
4 steps isn't all that much, considering that cousins are already 3 steps away 17:12
Auzon ok, well, maybe <10 steps then :P 17:13
masak Juerd recently taught me the necessity of decode/encode of all string data that crosses the I/O barrier of a program. what's the lay of the land in Perl 6 re the same thing? do we still decode/encode?
moritz_ masak: Str is a decoded string 17:14
masak which in practice means...?
moritz_ masak: I think that the IO stuff isn't fully specced, but the language design very strongly encourage the usage of Str over Buf
masak: so you'll probably add an encoding information on open() 17:15
masak as in Perl 5, then
moritz_ in this respect yes
but what's different is that you have different data types for text strings (Str) and byte strings (Buf) 17:16
which means that if you write a sub foo(Str $x), you force everybody who uses that to decode their strings
masak I'm not sure I see how that translates to real code
and what if I don't give my variable a type, what then? 17:17
moritz_ neither am I
masak: oh, you will. When you want to match a string against a regex, you'll want it to be a Str
Juerd moritz_: Well, using Str over Buf because it's recommonded is a stupid idea. Using it because it matches what you're doing is much better. 17:18
masak would a literal "like this" be a Str?
moritz_ masak: string literals are Str 17:19
Juerd And where you have I/O, you have bytes, thus Bufs.
And some conversion (that is: decoding/encoding) is necessary.
The nice thing here is that Buf know their encoding
masak what does my Bufs = "encode me"; do re encoding?
oops, my Buf $buf = "encode me"; 17:20
Juerd masak: Use the default encoding, which'll probably be either something that fails on everything, or ISO-8859-1, or UTF-8
masak what decides? how would I know?
Juerd masak: Larry does
masak :) 17:21
moritz_ Juerd: so it's not locale dependent?
Juerd I'm hoping for "something that fails on everything".
moritz_: Locale dependent would be good only for STDIO.
moritz_: And there I wish that locales are only used when explicitly requested.
It doesn't make sense all of the time.
For example, if stdin is a file, it makes no sense to make it locale dependent -- the file might be encoded with whatever the locale was when the file was written, but that information is *lost*, so it's better to specify a single encoding than to rely on heuristics. 17:22
"Assume that it hasn't changed" is stupid. If you assume that things never change, you wouldn't need to store things in a file :) 17:23
moritz_ you do store pi in a file, although it'll never change ;) 17:24
smtms you never know. someone might discover that pi is dependent on speed just like mass is 17:25
masak pi is dependent on local curvature :) 17:26
moritz_ smtms: no. That's the difference between math and physics. ;-) 17:27
masak: not pi, only the ratio between circle circumfence and diameter
smtms btw, you are not storing pi in a file, but just a finite approximation of it 17:28
masak moritz_: good point
masak but then you need a new name for that ratio. lots of people call it 'pi' 17:28
Juerd moritz_: I wonder though how localedependent pi is ;) 17:30
Does it get reset back to 22/7 when you set it to ancient greek? :)
moritz_ not pi again.. they'll just parameterize the ratio depending on the curvature 17:33
ruoso hmpf.... pmurias already gone away... 17:40
moritz_ ruoso: but he fixed something in the mean time 17:50
ruoso hmm... I'll take a look... thanks
moritz_ 'sh brew' now works (if you have perl 5.10.0 installed and as the default 'perl' in $PATH) 17:51
ruoso cool 17:53
Auzon moritz_: I'm working on moving operators/arith.t into the spec, but it's a pain to fudge since it uses its own testing routines. Do you have any suggestions? 18:54
moritz_ Auzon: first move it 18:55
Auzon: when you have done the move, I'll take a look
Auzon OK
pmurias ruoso: i'm back again ;) 18:57
ruoso cool
would you like to test replacing the sm0p parser by yours? 18:58
pmurias tried it already
my parsers has nasty bugs with the code emitting 18:59
Auzon moritz_: done
or should be in a second
pmurias nasty bugs means stringing array refs
moritz_ At revision 21187. 19:00
Auzon Yeah, svn-- is hanging
ok, done 19:01
pugs_svn r21188 | Auzon++ | [gsoc_spectest] staging commit for fudging arith.t
pmurias ruoso: KP6sm0p.p6 is the proper thing to use? 19:02
ruoso yes
pmurias i'll copy that over and try to make it work 19:03
ruoso take a look at sm0p.pl in the parent directory
it calls the compiled one 19:04
KP6sm0p.pl
but you could change sm0p.pl to call the appropriate script
moritz_ Auzon: I'm now trying to fudge it. Part of the trick is to use #?DOES for the tryeq and tryok subs 19:07
Auzon So what does DOES do? 19:10
moritz_ it says "this sub or block is equivalent to $x tests" 19:11
#?DOES $x at least
so it basically register it as a known test function
Auzon cool
fudge++
moritz_ I'm now at 175/190 tests (run or skipped) 19:13
pugs_svn r21189 | moritz++ | [spec] started to fudge artih.t - got rid of all compilation and run time 19:14
r21189 | moritz++ | errors, doesn't PASS yet
moritz_ Auzon: we can remove tryok, because todo is now handled differently 19:15
Auzon: same with tryeq, I think - it's just a fancy "is" that's not needed anymore 19:17
Auzon tryeq uses ==, and is uses eq for comparisons. Does that matter? 19:19
moritz_ good point 19:20
I'll keep tryeq for a moment
s/a/the/
Auzon But we should be able to replace tryeq with one line or so.
moritz_ aye 19:22
Auzon Are you still working on it, or can I take over? 19:23
moritz_ wait a second, I'll make a commit
Auzon ok
moritz_ then you can take over
Auzon Great :)
pugs_svn r21190 | moritz++ | [spec] arith.t more fudging, not quite there yet
moritz_ Auzon: it's yours ;-) 19:24
and it's really ugly to fudge, because it doesn't have too many test description, so it's hard to find the failing tests in the source file 19:26
Auzon yep
moritz_ but you aked for it ;) 19:27
Auzon I can use the same way I use to check for what line is causing a parsefail 19:28
moritz_ well, parsefail gives you a line number immediately ;) 19:29
that's why I love them
Auzon Not the one I had yesterday :-/
er, Tuesday
cj TimToady: so, by "There's more than one way to do it" do you mean "We just hacked something together, and it may work in mysterious ways"? 19:33
PerlJam cj: that's an odd interpretation. 19:34
moritz_ perl has been remarkably successful with its Timtowtdi approach, and not all of it seems like "hacked something together" ;-) 19:35
cj PerlJam: heh. blame the guy sitting next to me :) 19:38
PerlJam: hey, go update this... maybe reify what I've said... 19:39
Auzon moritz_: I found a way to get a bit of a hint on what line to look at by having tryeq report the $lhs and $rhs :) 19:41
pmurias TimToady: a|b is broken in gimme5 19:49
moritz_ pmurias: how so? it's parsed as inifx:<|>(a, b) it seems 19:51
Auzon heh. Rakudo does poorly on these corner cases. 121 skip, 69 pass.
pmurias token foo {a|b}
moritz_ pmurias: ah, I thought you meant 'a|b' as input 19:52
pmurias return 1 instead of proper match objects
moritz_ Auzon: mostly it doesn't do Inf and NaN yet 19:52
Auzon The UV/IV tests hurt too
moritz_ aye. How lucky sombody wrote them already ;-) 19:53
cj PerlJam: en.wikipedia.org/wiki/Rob_Levin#.23...-Operative
pugs_svn r21191 | Auzon++ | [gsoc_spectest] fudged arith.t for Rakudo
Auzon moritz_: You can add this file to the spectest_regression list now :)
moritz_ Auzon: running spectest_regression now. If it passes here, I'll commit it 19:54
Auzon great :) 19:56
only ~11000 tests left outside of spec/ ;)
moritz_ my crappy ISP disconnects me every ~30s
moritz_ 4 TODO passed 20:00
Auzon Cool :) 20:01
pugs_svn r21192 | moritz++ | [spec] use is_approx() instead of tryeq_sloppy for two tests, and remove their 20:02
r21192 | moritz++ | TODO marking
moritz_ runs the tests again, and gets some food in the mean time 20:03
pugs_svn r21193 | pmurias++ | [sm0p] initial version of the new grammar 20:08
r21194 | pmurias++ | [smop] added a SM0P_FASTER env var to used the new parser 20:11
ruoso pmurias++ 20:12
pmurias i'm now getting autotools to check if the new parser actually works
ruoso :) 20:13
pmurias ruoso: have you tried it yourself 20:15
ruoso not yet... 20:16
still getting the dependencies
pmurias fixing bugs...
other people want to use the internet& 20:22
rakudo_svn r29035 | moritz++ | [rakudo] added S03-operators/arith.t to spectest_regression. Auzon++ 20:22
pugs_svn r21195 | moritz++ | [spec] un-TODO two more tests in arith.t 20:26
pugs_svn r21196 | moritz++ | [t] removed duplicate test 20:55
r21197 | moritz++ | [spec] while.t: cleaned up two useless tests
pugs_svn r21198 | moritz++ | [spec] moved statements/repeat.t to spec/, cleaned up a test 20:58
r21199 | moritz++ | [util/perl6.vim] 'repeat' is a keyword like 'while' 20:59
r21200 | moritz++ | [spec] fudged repeat.t 21:04
rakudo_svn r29037 | moritz++ | [rakudo] add S04-statements/repeat.t to spectest_regression 21:11
pugs_svn r21201 | ruoso++ | [smop] still more work on p6opaque, few bugs fixed, some slime bugs still. 21:27
moritz_ likes ruoso++'s names very much 21:28
smop, s1p, slime ...
ruoso heh... i just hope slime communicates how lame the interpreter is ;) 21:29
TimToady @tell pmurias you can use [a|b] as a workaround until I have svn access again 22:55
lambdabot Consider it noted.
ruoso geez... p6opaque have yet another sm0p block... it should be taking ~ 2 or 3 minutes to parse... 23:02

Powered by Cro and the Raku® Programming Language.

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

Thank you!