--- Log opened Sun Jan 29 00:00:45 2006
00:53 -!- Arnia [n=jgeldart@1-40-101-159.adsl.legend.co.uk] has joined #nemerle
00:53 -!- Arnia [n=jgeldart@1-40-101-159.adsl.legend.co.uk] has left #nemerle ["Kopete 0.10.4 : http://kopete.kde.org"]
00:54 -!- Arnia [n=jgeldart@1-40-101-159.adsl.legend.co.uk] has joined #nemerle
01:18 -!- kanru [n=kanru@59-121-147-130.dynamic.hinet.net] has quit ["Leaving"]
01:33 -!- nazgul` [n=d47f5f7d@lilith.ii.uni.wroc.pl] has quit ["CGI:IRC (EOF)"]
03:18 -!- Arnia [n=jgeldart@1-40-101-159.adsl.legend.co.uk] has left #nemerle ["Kopete 0.10.4 : http://kopete.kde.org"]
04:38 -!- ChanServ [ChanServ@services.] has quit [ACK! SIGSEGV!]
04:49 -!- ChanServ [ChanServ@services.] has joined #nemerle
04:49 -!- ServerMode/#nemerle [+o ChanServ] by irc.freenode.net
08:03 < malekith> user317: are you running make install from the ROOT account? or are you making a package?
08:04 < malekith> d: it only works for tuples
08:04 < malekith> d: I didn't think it was needed for anything else
08:05 < malekith> d: though your example might be a good candidate too, but it would be def foo(Pair where (x, y)) { ... }
08:25 -!- nazgul` [n=nazgul@NATW6-Rozanka.ip.WRO.Korbank.PL] has joined #nemerle
08:25 -!- mode/#nemerle [+o nazgul`] by ChanServ
08:30 < user317> malekith: i am using gentoo: make DESTDIR=${D} install  should install everything into the diretory that i specify
08:51 -!- nazgul` [n=nazgul@NATW6-Rozanka.ip.WRO.Korbank.PL] has quit [Remote closed the connection]
09:05 < malekith> yes, lemme check
09:06 < malekith> user317: and where is gacutil trying to write? 
09:07 < malekith> user317: this works for rpm: http://pastebin.com/528447
09:18 < user317> thanks, let me try that
09:33 < d> malekith but why limit the possibilities here ? wouldn't such a simple transformation work for any patter ?:
09:34 < d> def (pattern) { ... } #==> 
09:34 < d> def (args) { match (args) { | pattern => ... }
09:34 < d> or with a "where" keyword or whatever, but so, that any patter could be used
09:35 < user317> thanks that worked
09:55 < d> and one more thing, could anyone tell me what is wrong with the GetEnumerator method here http://terror.org.pl/~d/rlist.n ? it's at the very bottom along with the error msg.. (remember this is polymorphic recursion, so some solutions used may seem weird at the first glance)
09:56 < d> i never used yield before, so i'm kinda clumsy with this..
09:56 < malekith> d: I think one should only use 'always-match' patterns in function parameters
09:57 < malekith> because:
09:57 < malekith> def foo(x :: xs) ...
09:57 < malekith> def foo ([]) ...
09:57 < malekith> simply won't work
09:57 < d> hm
09:57 < d> so why not
09:57 < malekith> d: you cannot use yield in local functions... sorry.. :/ 
09:57 < d> just give a warning
09:58 < d> if someone feels too much like programming haskell ;]
09:58 < malekith> ;)
09:58 < d> hm. so i won't be able to create an enumerator for rlists ?;/
09:58 < malekith> yield doesn't remember the stack
09:58 < malekith> this is why it is so restricted
09:58 < d> mhm
09:59 < malekith> but I guess it can be written
09:59 < malekith> using Yield
10:00 < d> is the capital Y there on purpose ?;]
10:00 < d> if so, then i don't know what you're talking about ;]
10:01 < malekith> no
10:01 < malekith> lemme think a bit
10:01  * malekith is wrapping his mind around poly-recursion ;-)
10:01 < d> i need to use pass a function as one of the arguments, there's no way to do it otherwise and in such a case, the GetEnumerator method wouldn't match the signature..
10:01 < d> s/use//
10:03 < d> btw i used 'b in there, 'cause in RList [Pair ['a]].Enumerate 'a becomes Pair ['a] and the enumerator is for 'a, so i just gave up type checking (dunno if that's a correct approach)
10:03 < malekith> http://pastebin.com/528479
10:03 < malekith> but I guess I got the ordering wrong
10:04 < d> hmm
10:04 < d> but i want to yield 'a elems
10:04 < d> not Pairs
10:04 < malekith> and you do
10:04 < malekith> yield elem.fst;
10:04 < d> remember, that the deeper the recursion
10:04 < d> the deeper the pair nesting
10:04 < malekith> yes
10:05 < malekith> but it compiles and it seems to me it will work
10:05 < d> really ? great :)
10:06 < malekith> there should be FromList static method
10:06 < malekith> of maybe a constructor taking lis
10:06 < d> there will ;]
10:06 < d> but that needs a bit more thinking
10:06 < d> because consing every elem is not optimal
10:07 < d> hm, it works indeed. wiiiicked ;>
10:08 < malekith> :)
10:09 < malekith> I guess I even got the ordering almost right
10:09 < malekith> it seems just reversed
10:09 < d> no really, it seems to be totally ok
10:09 < malekith> or even not reversed
10:11 < d> and as i said before - couldn't [this] be matched by default if the function takes no arguments ?
10:11 < d> so "match" and the two annoying curly braces could be ommitted ;]
10:12 < malekith> I just posted an email about this to devel-en
10:12 < d> mhm
10:12 < malekith> I think this can (and should) be done
10:13 < malekith> unfortunately
10:13 < malekith> Rev (acc : list ['a])
10:13 < malekith> won't work (inside some variant)
10:16 < malekith> that is you cannot say: 'Rev (match this, acc : list ['a])
10:16 < malekith> or maybe you should be able... 
10:18 < d> yeah, i like the syntax with an argument prefixed with "match"
10:19 < d> and with "match this" - well this is not as obvious as in, say python, where "this" is explicitly written among the arguments, but it seems reasonable enough to me 
10:39 < CIA-10> d * r6090 /nemerle/trunk/lib/rlist.n: Add enumerators for Pair and RList (thank's to malekith).
11:07 < d> irclog grepping at the nemerle site could be usefull ;-d
11:08 < d> i mean even a field that asks google "query site:http://nemerle.org/irc/" would be enough
11:23 < malekith> i guess you cannot filter google results based on directory
11:23 < malekith> only based on hostname
11:24 < malekith> but I may be mistaken
11:36 < d> it worked for me just a moment ago
11:36 < d> ;]
12:08 < malekith> oh... strange. maybe they added it
12:08  * malekith works hard on getting his theorem prover to where it was 30 revisions ago... 
12:08 < malekith> (performance-wise)
12:09 < malekith> i'm close http://nemerle.org/~malekith/temp/plot.png
12:53 < d> heh
12:55 < d> l3.Length = 10000000 (time: 00:00:01.7827830)
12:55 < d> rl3.Length = 10000000 (time: 00:00:00.3186440)
12:55 < d> that's better.
12:55 < d> i fixed Length a bit.
12:58 < CIA-10> d * r6091 /nemerle/trunk/lib/rlist.n: Fix Length, add the simplest version of FromList.
13:38 < d> malekith is it *really* worth (performance-wise) to repeat some code from a certain match-branch instead of making a  recursive call that will lead to that branch ?
13:45 < d> malekith i'm thinking of situations like in Nth where when we match One (x, ps), then we make a recursive call with Zero (ps), instead we could make these computations instantly, without this recursive call. code will be slightly longer and less concise, but possibly faster, and it's that speed that i started writing this for in the first place..
13:55 < malekith> d: The problem is possibly not a call, but an allocation and subsequent match
13:56 < malekith> but you'll never know until you profile
13:57 < d> mhm. i'm making tests now, we'll see..
14:01 < d> on a 2500000-element rlist, looking up each elem with Nth takes 20.25secs for normal Nth and 28.95 for the Nth without the additional call and matching ;-d
14:01 < d> but differences in results on small lists are more significant
14:02 < d> 0.34 vs. 0.001
14:02 < d> for a 20-elem rlist
14:02 < d> so i guess it's worth it
14:02 < d> whoops
14:02 < d> thats 0.034 vs. 0.001
14:03 < d> but still.. ;]
14:03 < d> i hope these relations hold on other CPUs ;]
14:06 < malekith> that is expanded version is SLOWER on long lists? that's very strange
14:06 < d> ah
14:07 < d> another mistake
14:07 < d> that's 18.95
14:07 < d> not 28.95
14:07 < d> sorry
14:07 < malekith> so it's always faster, good
14:07 < d> yup.
14:07 < d> just the difference is bigger for smaller lists
14:07  * malekith will be trying to run System Zapisy on his home box
14:07 < d> and we usually use small lists like that
14:07  * malekith expects PROBLEMS
14:07 < d> heh
14:07 < d> what for ?:)
14:08 < malekith> testing
14:08 < d> i see
14:08 < malekith> to see what's really needed for it to run
14:08 < d> possibly a complete rewrite ;]
14:08 < malekith> I'm the new admin of this shit
14:08 < malekith> hehe
14:08 < malekith> maybe, but I don't have time for such things
14:08 < malekith> maybe some student :)
14:09 < d> yeah, some student already did that
14:09 < d> but never finished
14:09 < d> cause he got his title before it was done
14:10 < malekith> heh. anyway it's possible we'll use something called ipsos or upsos also for ''system zapisy''
14:10 < malekith> in one year or so
14:11 < d> never heard of it ;-d\
14:12 < malekith> Need to get 25.8MB of archives (25.8MB to download).
14:12 < malekith> After unpacking 72.3MB will be used.
14:12 < malekith> pff
14:12 < d> heh
14:13 < d> perhaps adding caching to the current software would solve the perfoirmance problems
14:13 < d> there is some general stuff for apache, mmcache or something. never used it, just heard of it.
14:13 < malekith> are there still performance problems? 
14:13 < d> during the day the system is opened ? HUGE
14:14 < d> it took a few minutes after each click
14:14 < d> for the site to load
14:14  * malekith didn't use the system for a while
16:04 < d> the errors are really scary when you forget $ in LCs. perhaps some friendly hint would be in order.. after all many people are going to be used to the notation without the $..
16:06 < d> oh. hgat was a different error
16:15 < CIA-10> d * r6092 /nemerle/trunk/lib/rlist.n: Slightly optimize Nth, FoldLeft and FoldRight.
16:33 < d> and btw, "detached" type specifications, such as those in haskell, are great. they are good documentation (when type inference is being used everywhere) and they could be a nice solution now, that type inference on global methods is not implemented. people could write code as if it was implemented and just add a line with a detatched signature for each method.
16:34 < d> just a thought..
16:34 < d> :)
18:01 -!- serras [n=serras@62-14-190-160.inversas.jazztel.es] has joined #nemerle
18:01 -!- serras [n=serras@62-14-190-160.inversas.jazztel.es] has left #nemerle ["Abandonando"]
18:34 < d> wow, i love these LCs, they're really handy during testing
18:53 < malekith> ;-)
18:54  * malekith is looking into extension methods
19:15 < CIA-10> d * r6093 /nemerle/trunk/lib/rlist.n: Fix Equals, add FromList (xs, i) with O (|xs|) time complexity.
19:16 < d> not even a reverse was required :]
19:18 < d> hm. 645 lines already ;-d mostly comments though ;]
19:23 -!- nazgul` [n=d47f5f7d@lilith.ii.uni.wroc.pl] has joined #nemerle
19:23 -!- mode/#nemerle [+o nazgul`] by ChanServ
23:54 -!- nazgul` [n=d47f5f7d@lilith.ii.uni.wroc.pl] has quit ["CGI:IRC"]
--- Log closed Mon Jan 30 00:00:46 2006