-Ofun: xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by apple-gunkies on 11 November 2005.
luqui for anybody who knows: where is the "\^J" etc syntax documented 00:37
Now that ^ is a sigil, this causes problems with escaping 00:38
leo luqui: sorry for the dumb question: what should "\^J" mean? 00:50
escape-wise
luqui: did you mean "\cJ" perhaps? 00:57
luqui oh, a control character 01:07
currently implemented in pugs
leo the \^ would be really overkill then, I'd drop that (nothing against TIMTOWTDI, but ...) 01:09
luqui ahh, \c would be a much better choice for that 01:10
I'll look at the docs
leo well perldoc perlre ;-)
luqui but is that also in strings? 01:11
leo as a side note - this pod isn't the proper place to explain perl5 string escapes, but it does
luqui it appears so 01:12
leo that sayeth the text
and it works iirc
there are about 5 working constructs to escape non-ascii, I never heard of "\^J" 01:13
luqui yeah, I hadn't either 01:14
just changed it to \c
leo parrot supports: 01:15
\xhh 1..2 hex digits
\ooo 1..3 oct digits
\cX control char X
\x{h..h} 1..8 hex digits
\uhhhh 4 hex digits
\Uhhhhhhhh 8 hex digits
\a, \b, \t, \n, \v, \f, \r, \e
enuff I think ;-)
luqui :-) 01:16
leo was pondering /N{foo} too, when reading perlre 01:17
luqui is ^A traditional ASCII 0 or 1? 01:18
leo because of cross-compilation - the "parser" or/and the runtime has to support it
luqui doesn't see why... 01:19
leo ^A = 1 ; ^@ = 0
luqui aa soo 01:20
leo I don't know how fit haskell/pugs is WRT unicode, but parrot supports it (or supports more unicode), *if* the ICU lib is installed 01:22
luqui ?eval "foo".substr:1 01:29
evalbot_7950 Error: Unimplemented 3-ary op: substr
svnbot6 r7951 | luqui++ | Delete Language::AttributeGrammar. I'll resurrect it at some point. 02:19
r7952 | luqui++ | Added mandatoryWhiteSpace combinator (needs cleanup).
r7952 | luqui++ | Changed \^X control char to \cX, as per spec. Also fixed math so that
r7952 | luqui++ | \cA refers to char 1 rather than 0 (where \c@ refers to 0).
r7952 | luqui++ | Implemented method-call listop form: $obj.method: arg1, arg2.
r7953 | luqui++ | Added some tests for these changes. The listop method needs
r7953 | luqui++ | many many more, though.
luqui ?eval "check it out".substr: 4, 6 02:34
02:34 evalbot_7950 is now known as evalbot_7953
evalbot_7953 \"k it o" 02:34
luqui woot and whatnot
Alias_ autrijus: ping? 02:45
dduncan I have yet to track it down, but there seems to be a problem with pugs smoke that started to manifest after I switched to svk ... the smoke report doesn't mention a pugs revision number in it, while otherwise seeming to generate correctly, and after uploading it to the server, the server ignores it / doesn't display it, perhaps since undef is lower than the last highest versions ... 05:21
I uploaded two repository / darwin over the last few days, one just now, that don't show up
is anyone else successfully uploading smokes and using svk? 05:22
gaal dduncan: please fix in util/yaml_harness.pl around line 196 05:25
also possibly add an early error/warning in smoke-client against reports with missing r numbers 05:26
dduncan will have a look ...
gaal dduncan++
$work &
dduncan may have to wait for tomorrow, as I must sleep
gaal do you know what the svk command is to get the right upstream r number? 05:27
dduncan no, but the 'make' process is able to get it when building pugs itself I think, so I can try looking there ... I know it involves invoking svk 05:29
as in, seeing how make does it
gaal of course :) well, i really have to go now, so good luck
dduncan right ... I'll try to do this tomorrow ... but doesn't look too complicated
if all goes well, should checkin it in about 12 hours from now 05:30
gaal cool.
autrijus gaal: I'm still stuck at $work, will checkback tonight (in 6hrs), but excellent article :) 06:02
svnbot6 r7954 | autrijus++ | * Trivial copyediting over gaal++'s excellent tpr article.
r7954 | autrijus++ | * Update animals.p6 to use more elegant (and less misleading)
r7954 | autrijus++ | given/when clauses.
Alias_ autrijus: ping
autrijus Alias_: yes?
(yes I had not booked ticket nor applied for visa -- later today)
Alias_ If I were to nudge brisbane.pm about having a meeting, would you be interested in taking a drive up there one evening to talk? 06:03
autrijus of course yes
Alias_ Probably in the trolltech offices there
autrijus sure. if they all went to osdc I may need to think of something new though
Alias_ I'm note sure how many will be
autrijus if not, then I'll just repeat my osdc talk, which would be least effort
Alias_ Coverage isn't great
autrijus try inquire a bit
Alias_ brisbane.pm is one of those sometimes alive, sometimes dead groups :) 06:04
autrijus . o O (Weremongers)
obra hi
autrijus greetings obra-san. 06:05
obra autrijus: how's .tw?
autrijus busy. too many forgotten previously arranged meetings and deadlines
some 1283 unread mails
but that's all part of course :)
obra heh
autrijus I'm doing fine though and finally beat jetlag, which is a definite improvement 06:06
can't do much creative things under jetlag
(gotta run) 06:07
bbiab :) &
obra later
understood
gaal autrijus: thanks! I'm at $work myself and the article has been pushed to another issue, so there's no rush. 07:46
also, when ({}) Just Works now that ~~ CODE has been implemented, so I don't need direct help with that (though I'm still *very much* interested in more general pointers about tracing Parsec parsers!) 07:47
clkao it seems dduncan and i have exclusive timezone 09:50
gaal "underlapping schedules" is my term for that kind of thing 10:44
nothingmuch who runs the log bot? 13:09
Juerd The logbotmaster! 13:13
15:20 stevan__ is now known as stevan
robkinyon Anyone here understand math at a high level? 15:24
blazar just threw me a curveball and I want to discuss w/someone before I reply
particularly in terms of sets, iterators, type inferencing, and the like 15:25
17:12 Lopo_ is now known as Lopo
dduncan I'm back 17:47
so will now look at that yaml harness thing and hopefully have a fix in an hour
seen gaal 18:30
jabbot dduncan: gaal was seen 7 hours 46 minutes 2 seconds ago
gaal hi dduncan 18:37
dduncan hello
I seem to have tracked the problem and determined a solutin
gaal great!
dduncan it appears to be more widespread than I though too 18:38
./pugs -v doesn't work either
so what I will do is ...
gaal so PugsConfig.pm is broken too?
there must be 'svn info's sprinkled around
dduncan update version_h.pl to work with svk correctly, it doesn't, so pugs own internal version is 0 when under svk 18:39
gaal my Q#1 is whether svk can get the upstream r number w/o talking over the wire
dduncan and update the harness to use the same pugs_version.h which that makes and that pugs' build uses
gaal my Q#2 is if it's correct to put in an r number at all if there are loca changes
dduncan gaal, I've determined that for #1 yes, but you have to invoke it twice 18:40
at least, for a standard setup of a svk user having a remote mirror and a local copy ...
gaal yes i don't think we need to worry about pugs forks at this point ;-)
dduncan the first 'svk info' will give us the local rev #, not the remote one, but also the name of the local mirror 18:41
using that, we call svk info again with the name of the local mirror, and that gives us the rev of the remote one
gaal brb, something's up with my cat
dduncan in any event, I noticed that the test harness is redoing (differently) the work in version_h, rather than reusing its output 18:42
gaal i'm not actually sure which came first :) 18:44
but by all means, unify! 18:45
dduncan the current version_h is naive, assuming a plain 'svk info' on the cwd has the right info, but it doesn't in many cases, so result = 0
the test harness ignores svk entirely, with a todo item
thanks for pointing me in the right direction, though, as I never used the 'info' command before 18:46
which the test harness you pointed me to did
gaal re: which svk to use, that could be controlled by config.yml 18:47
dduncan I notice even the handling of svn is different in the two files ... version_h looks at the .svn files, the harness does a shell call to svn info
gaal yeah
peeking at .svn is both better and worse. 18:48
dduncan I should point out that the output of svn/k info contains enough info, or seems to, that we can autodetect the user's setup enough to get version, I think 18:49
gaal in any case: the harness probably shouldn't do this at all, but get the version from pugs (or version_h)
okay, go for it :)
dduncan true ... if pugs is fixed to work, the harness can use that
so I'll fix this in 2 steps, first just get pugs to contain the version always, and separately the smoke 18:50
gaal sounds good to me
dduncan++
bbiab &
theorbtw1 BTW: When you upgrade any of that, check the case where svk is installed but has never been used. 19:26
19:26 theorbtw1 is now known as theorbtwo
theorbtwo I rewrote version_h (IIRC) to fix that case. 19:28
r0nny yo 19:29
dduncan hello
well, commit it ... then I won't have to fix it
otherwise I'm prepared to, right now 19:30
theorbtwo dduncan: I did -- several months ago.
dduncan well, the current version has a problem
the problems I describe are the current version
so unless someone undid your changes ...
dduncan back in 15 mins 19:31
19:42 lampus_ is now known as lampus
xinming anyone here know about haskell? 20:36
nothingmuch xinming: sort of 20:39
but the people on #haskell know much better
and they're nice
xinming nothingmuch: hmm, But i am too new, might be bothersome to them. :-/
nothingmuch nonesense, they're nice
but i am willing to help first here
if you feel more comfortable 20:40
xinming hmm, how to make a lambda contain another lambda? 20:41
hmm, well, I don't mean the normal lambda, I mean the lambda contain the pattern match 20:44
integral you use a case-of 20:45
\x -> case x of ...
SamB newbies not considered harmful 20:46
xinming (\a -> case x of [] -> 0; case x of (x:y) -> x + "outer lambda" )
oops
SamB though I guess if we don't understand your questions it won't be very useful for you to come to #haskell... 20:47
xinming SamB: I saw your in haskell, But I wish to ask more complex question there. :-)
I was attracted by the example Lemmih show me 20:48
SamB usually when your lambdas get that complicated you try to stick the function in a "let" or a "where"
integral I don't like those since I can't then test the function individually from ghci; but then it can't close either *sigh* 20:49
xinming SamB: haskell interpreter can't define the function on the fly... that's why I feel lambda more confortable. 20:50
SamB well, if it is self-contained, you could just stick a "let" in front and paste it, then test it
in GHCi
not hugs though
er, well, I guess not unless you use ; and put it all on one line... 20:52
xinming (\a -> case a of [] -> 0; case a of (x:y) -> x + y )[2,3] 20:53
SamB usually tries to put functions so complicated he wants to test them interactively at the top level
integral xinming: I gave you can example in #haskell
xinming integral: Ok
integral you mean: (\a -> case a of { [] -> 0; [x,y] -> x + y; _ -> error "oops") [2,3]
pawel Juerd: Are you there? 20:56
pmurias a bit OT:Are there any plans to extend a haskell compiler to generate pasm/pir/pil/? 21:38
SamB why would you want any of those? 21:41
GHC generates C-- internally
integral pmurias: you should feel free to grab a copy of something like jhc and hack away :)
SamB when it isn't compiling to bytecode, that is
pmurias i suppose it would improve the integration of haskell libs with parrot.. 21:43
dduncan okay, first part of fix is committed 21:45
pmurias and make writing hybrid haskell/tcl-or-pir-or-perl6 easier 21:46
dduncan now pugs revision is properly gathered when client is svk
so pugs -V actually works
integral golly, that bit of code is nothing but hassle :)
jdv79 seen stevan 21:47
jabbot jdv79: stevan was seen 5 days 4 hours 11 minutes 36 seconds ago
svnbot6 r7955 | Darren_Duncan++ | r832@Darren-Duncans-Computer: darrenduncan | 2005-11-14 13:44:49 -0800 21:48
r7955 | Darren_Duncan++ | updated util/version_h.pl so that 'make' updates /src/Pugs/pugs_version.h properly when svk is configured with the standard //svk/trunk and //svk/local duality ... it now looks up the version from the trunk referred to by the local, and not just the local
dduncan the second fix is done, and I'm doing a smoke to test that that works ... will commit in an hour or so 22:04
yaml_harness now gets its revision from pugs itself, rather than svn
j0sephi picasa.hu ;-) :) 23:10
theorbtwo Nifty. 23:13
(And not always SFW.)
svnbot6 r7956 | Darren_Duncan++ | r834@Darren-Duncans-Computer: darrenduncan | 2005-11-14 15:29:29 -0800 23:30
r7956 | Darren_Duncan++ | update util/yaml_harness.pl so get_revision() gets revision from pugs itself rather than from svn ... so it works with svk too
quicksilver stevan: ping 23:52
stevan: when you have a moment come to irc.perl.org#dbix-class, mst + I have some thoughts about class::C3 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!