00:00:00 --- log: started forth/09.05.29
01:14:14 --- quit: nighty__ (Client Quit)
01:26:05 --- quit: sunwukong ("bye")
03:02:56 --- join: mark4 (n=mark4__@ip70-162-227-36.ph.ph.cox.net) joined #forth
03:04:06 --- quit: I440r (Read error: 110 (Connection timed out))
03:46:06 --- join: GeDaMo (n=gedamo@212.225.121.248) joined #forth
04:03:18 --- quit: proteusguy (Read error: 110 (Connection timed out))
04:25:57 --- join: proteusguy (n=proteusg@61.7.144.97) joined #forth
04:41:12 --- quit: GeDaMo ("Leaving.")
05:02:13 --- join: tathi (n=josh@dsl-216-227-91-166.fairpoint.net) joined #forth
05:26:32 --- join: schme (n=marcus@c83-249-82-162.bredband.comhem.se) joined #forth
05:35:11 --- join: sunwukong (n=vukung@125-14-81-177.rev.home.ne.jp) joined #forth
08:00:24 --- join: SonOfLilit (n=aursaraf@bzq-79-183-229-21.red.bezeqint.net) joined #forth
08:01:02 <SonOfLilit> Hello
08:01:54 <SonOfLilit> I just started to teach myself forth
08:02:05 <SonOfLilit> I'm trying to reverse a string in place
08:02:56 <SonOfLilit> I loop calling a word I called s-switch ( base i1 i2 -- ) which switches the indexes i1 and i2 in the string at address base
08:03:09 <SonOfLilit> this word is ugly
08:03:13 <SonOfLilit> REALLY ugly
08:03:29 <SonOfLilit> and I've tried three different approaches, and it's still ugly
08:04:18 <SonOfLilit> stack manipulation just doesn't feel right for such words
08:04:22 <SonOfLilit> how would you guys implement it?
08:21:11 <SonOfLilit> ah, finally, I found a clean way
08:22:00 <SonOfLilit> value i2 value i1 value b b i2 s@ b i1 s@ b i2 s! b i1 s!
08:22:31 <SonOfLilit> where s@ ( addr index -- c ) , s! ( character addr index -- )
08:23:04 <SonOfLilit> does this overwrite the previous value or leak like crazy?
08:45:50 --- quit: mark4 (Read error: 104 (Connection reset by peer))
08:46:28 --- join: mark4 (n=mark4__@ip70-162-227-36.ph.ph.cox.net) joined #forth
08:50:47 <tathi> Er...comma allots space on the heap, if that's what you mean
08:51:20 <SonOfLilit> no, there's no comma in my code
08:51:28 <SonOfLilit> the comma is part of the english explaining it
08:51:33 <tathi> ah
08:51:45 <SonOfLilit> I only give stack effect diagrams for s@ and s! as they are trivial
08:51:52 --- join: GeDaMo (n=gedamo@212.225.121.248) joined #forth
08:51:53 <tathi> sure
08:52:11 <SonOfLilit> : s@ chars + c@ ; : s! chars + c! ;
08:52:16 <SonOfLilit> I mean value
08:52:30 <SonOfLilit> does VALUE allocate new memory every time?
08:52:41 <SonOfLilit> allot*
08:53:29 <SonOfLilit> and how would you write S-SWITCH ?
08:53:54 <tathi> only when you define a value initially
08:54:02 <tathi> not when you access it or change it, of course
08:54:13 <SonOfLilit> ok, so no crazy memory leak :)
08:54:28 <SonOfLilit> and this seems like the way to go on a forth system without locals
08:54:34 <tathi> yeah, sorry.  I didn't take the time to follow your code carefully.
08:54:41 <SonOfLilit> it's ok :)
08:54:45 <tathi> I don't think the stack manipulation for s-switch should be bad...
08:54:58 <SonOfLilit> can you give it a go?
08:55:05 <tathi> yup, thinking
08:56:15 <tathi> hmm...maybe it is that bad. :)
08:56:28 <tathi> might be better if you just gave it two addresses?
08:57:50 <tathi> : exchange ( addr1 addr2 -- )  over @ over @ rot ! swap ! ;
08:57:59 <tathi> er...I mean c@ and c!
08:59:58 <tathi> and...that's still wrong
09:01:07 <GeDaMo> What problem are you trying to solve?
09:01:18 <SonOfLilit> Gave myself an excersize
09:01:20 <tathi> he's trying to reverse a string in place
09:01:27 <SonOfLilit> reversing a string in place
09:01:39 <SonOfLilit> we're working on a clean word to reverse two characters
09:02:20 <tathi> it's true that stack manipulation often isn't great for this sort of thing
09:02:37 <tathi> too many things to deal with at once
09:03:02 <GeDaMo> Why doesn't your exchange word work for that?
09:03:15 <SonOfLilit> it replaces each location with itself
09:03:16 <tathi> it does, just needs another swap (before rot).
09:06:28 <SonOfLilit> : exchange-addrs ( addr1 addr2 -- ) over c@ over c@ swap rot c! swap c! ;
09:06:28 <SonOfLilit> : exchange ( addr i1 i2 -- ) >r over r> chars + >r chars + r> exchange-addrs;
09:06:28 <SonOfLilit> : exchange-addrs ( addr1 addr2 -- ) over c@ over c@ swap rot c! swap c! ;
09:06:28 <SonOfLilit> : exchange ( addr i1 i2 -- ) >r over r> chars + >r chars + r> exchange-addrs;
09:06:28 <SonOfLilit> : exchange-addrs ( addr1 addr2 -- ) over c@ over c@ swap rot c! swap c! ;
09:06:35 <SonOfLilit> uh, sorry
09:07:02 <SonOfLilit> connection problems combined with stupid irc client made me think paste of two lines doesn't work
09:08:03 <SonOfLilit> yeah, this workd
09:08:05 <SonOfLilit> works*
09:08:37 <SonOfLilit> gotta say the VALUE version is much cleaner though
09:09:02 <SonOfLilit> or is my exchange at fault?
09:09:47 <tathi> no, I think there's just too much stuff on the stack for it to be comfortable
09:10:47 <tathi> you could write it using a bunch of different stack-manipulation words, but I don't know that any of them are any better.
09:11:12 <tathi> 2dup c@ swap c@  rot c! swap c!
09:11:45 <SonOfLilit> no, I'm talking about exchange, not exchange-addrs
09:12:21 <tathi> oh, I see.
09:13:03 <tathi> I think I'd try to write the loop to give you addrs rather than indices
09:13:16 <GeDaMo> I agree
09:14:22 <SonOfLilit> well, here it is then for your hacking pleasure:
09:14:56 <SonOfLilit> : reverse ( addr n -- ) dup 2 / 0 do 2dup i - 1- i exchange loop 2drop ;
09:15:10 <SonOfLilit> I can't think of a better way to write it than factoring out the indices-to-addresses code
09:15:37 <SonOfLilit> wait, maybe I can
09:15:40 <SonOfLilit> just a moment
09:16:04 <tathi> you could do the calculation in the loop, where you have `I` available instead of having to keep everything on the stack
09:16:17 <SonOfLilit> yeah
09:17:58 <SonOfLilit> : c+ chars + ;
09:17:58 <SonOfLilit> : exchange ( addr1 addr2 -- ) over c@ over c@ swap rot c! swap c! ;
09:17:58 <SonOfLilit> : reverse ( addr n -- ) dup 2 / 0 do 2dup i - 1- c+ over i c+ exchange loop 2drop ;
09:18:14 <SonOfLilit> much better
09:18:21 <GeDaMo> There's a 2/ word
09:18:44 <SonOfLilit> yay, so now my code is 80 chars wide and not 81 :)
09:18:45 <GeDaMo> It does a right shift rather than a division
09:19:18 <SonOfLilit> but... it doesn't pass the test
09:19:22 <SonOfLilit> let me have another look
09:21:18 <tathi> ok, or you could use indices and do : exchange ( addr i1 i2 -- )  rot tuck + >r + r> exchange-addrs ;
09:21:46 <tathi> er...well, unless you want chars in there
09:22:02 <SonOfLilit> what does tuck do?
09:22:11 <tathi> ( a b -- b a b )
09:22:17 <tathi> swap over
09:22:21 <SonOfLilit> mm
09:25:44 <SonOfLilit> this fixes my bug:
09:25:45 <SonOfLilit> : c+ chars + ;
09:25:45 <SonOfLilit> : exchange ( addr1 addr2 -- ) over c@ over c@ swap rot c! swap c! ;
09:25:45 <SonOfLilit> : reverse ( addr n -- ) dup 2/ 0 do 2dup over i c+ rot rot i 1+ - c+ exchange loop 2drop ;
09:26:07 <SonOfLilit> is there a short for rot rot?
09:26:37 --- quit: Deformative (Remote closed the connection)
09:26:43 <GeDaMo> -rot
09:28:17 --- quit: sunwukong ("bye")
09:30:36 --- join: Deformative (n=joe@c-71-238-45-45.hsd1.mi.comcast.net) joined #forth
09:33:04 --- join: bifft (n=sixforty@adsl-070-148-228-084.sip.asm.bellsouth.net) joined #forth
09:37:32 --- part: bifft left #forth
11:04:40 --- quit: Al2O3 ("I'm heading out, will be back online soon")
12:33:20 --- quit: schme (Read error: 113 (No route to host))
13:15:18 --- join: kar8nga (n=kar8nga@e-62.vc-graz.ac.at) joined #forth
14:17:56 --- join: aguaithefreak (n=aguai@114-42-211-66.dynamic.hinet.net) joined #forth
14:18:34 --- quit: aguaithefreak (Client Quit)
14:35:51 --- quit: GeDaMo ("Leaving.")
15:02:16 --- quit: sleepydog (anthony.freenode.net irc.freenode.net)
15:02:17 --- quit: madgarden (anthony.freenode.net irc.freenode.net)
15:03:28 --- join: sleepydog (n=david@64-252-82-123.adsl.snet.net) joined #forth
15:03:28 --- join: madgarden (n=madgarde@CPE001d7e527f89-CM00159a65a870.cpe.net.cable.rogers.com) joined #forth
15:08:01 --- join: maht (n=maht__@85-189-31-174.proweb.managedbroadband.co.uk) joined #forth
15:12:38 --- quit: kar8nga (Remote closed the connection)
15:19:57 --- quit: maht_ (Read error: 110 (Connection timed out))
16:51:59 --- quit: ENKI-][ ("WeeChat 0.2.6")
16:57:26 --- join: ENKI-][ (n=weechat@c-71-234-190-248.hsd1.ct.comcast.net) joined #forth
17:09:15 --- quit: sleepydog ("leaving")
19:56:07 --- join: Dialupas (n=Rodo1@v-209-98-172-217.mn.visi.com) joined #forth
19:56:49 <Dialupas> Anyone around?
20:00:24 --- quit: Dialupas (Read error: 104 (Connection reset by peer))
20:10:52 --- join: Al2O3 (n=Al2O3@24.9.43.112) joined #forth
20:38:34 --- quit: Al2O3 ("I'm heading out, will be back online soon")
20:47:51 --- join: Deformati (n=joe@c-71-238-45-45.hsd1.mi.comcast.net) joined #forth
21:19:50 --- join: Def (n=joe@c-71-238-45-45.hsd1.mi.comcast.net) joined #forth
21:19:51 --- quit: Deformati (Read error: 54 (Connection reset by peer))
21:27:32 --- join: crc (n=charlesc@c-68-80-139-0.hsd1.pa.comcast.net) joined #forth
21:29:18 --- quit: crc (Client Quit)
22:47:53 --- quit: mark4 (Read error: 54 (Connection reset by peer))
22:49:30 --- join: mark4 (n=mark4__@ip70-162-227-36.ph.ph.cox.net) joined #forth
23:05:34 --- quit: nighty^ ("Disappears in a puff of smoke")
23:59:59 --- log: ended forth/09.05.29