00:00:00 --- log: started forth/01.06.27 00:50:53 heay 00:50:55 heya 01:53:08 --- join: rob_ert (robost86@h3n1fls33o898.telia.com) joined #forth 02:53:14 --- nick: rob_ert -> rob-ert_aw_ay 03:25:07 --- quit: klooie (Ping timeout for klooie[213-84-79-23.adsl.xs4all.nl]) 04:23:55 --- join: cleverdra (jfondren@1Cust129.tnt3.florence.sc.da.uu.net) joined #forth 04:23:55 --- mode: ChanServ set mode: +o cleverdra 04:24:03 --- mode: cleverdra set mode: +oooo aaronl dired I440r nate37 04:24:03 --- mode: cleverdra set mode: +o rob-ert_aw_ay 04:25:07 --- topic: set to ': sum[0..n] dup 1+ 2 */ ;' by cleverdra 04:25:28 g'morning #Forth 04:26:48 --- nick: rob-ert_aw_ay -> rob_ert 04:27:13 good morning... 04:28:35 --- nick: rob_ert -> lighting_rob_ert 04:29:13 --- nick: lighting_rob_ert -> rob_ert 05:08:01 --- nick: rob_ert -> rob_ert_far_far_ 05:08:04 --- nick: rob_ert_far_far_ -> rob_ert_far_far_aw_ay 05:18:37 --- join: nibble (pool@host213-122-72-252.btinternet.com) joined #forth 05:26:02 --- mode: cleverdra set mode: +o nibble 05:26:07 hello nibble. 05:26:43 heya 05:28:41 dired, laws of form is at http://www.ultratechnology.com/laws.html#laws or http://www.rgshoup.com/lof/lof.html 05:28:42 OK, cleverdra. 05:32:39 nibble, I'm reading something right now, and the others are apparently asleep, but if you've any questions just ask them and I'll answer you when I notice. 05:33:09 hello 05:33:24 * aaronl has the high bid on a few SGI's right now 05:33:26 oh. g'morning aaron 05:33:36 aaron - cool. 05:33:43 yup 05:33:48 i just hope i don't win more than one :D 05:35:05 SGI Indy R4400 150Mhz / 32Meg / 1Gb HD 05:48:03 ebay? 06:07:48 --- nick: rob_ert_far_far_aw_ay -> rob_ert 06:18:44 yup 06:18:46 and i won it 06:19:04 now i just need a sweet little 20" TRINITRON SGI SONY NO RESERVE :p 06:20:55 congratulations! 06:25:30 --- nick: rob_ert -> rob_ert_aw_ay 06:36:34 --- quit: nibble (Ping timeout for nibble[host213-122-72-252.btinternet.com]) 06:52:55 cleverdra: any idea where i can find 72pin parity ram? 06:55:32 no, sorry. 07:09:59 --- join: grurp (grurp@pc19960.batc.tec.ut.us) joined #forth 07:24:38 --- join: nibble (involving@host213-122-69-163.btinternet.com) joined #forth 07:29:31 --- mode: cleverdra set mode: +oo grurp nibble 07:29:37 wb. 07:40:03 --- join: klooie (kloo@213-84-79-23.adsl.xs4all.nl) joined #forth 07:40:07 hi. 07:48:19 --- mode: cleverdra set mode: +o klooie 07:48:23 hello klooie =) 07:48:56 oh lookie... 07:48:57 * klooie smiles. 07:55:36 ++ paralellism 07:56:29 : ++ bl parse 2dup karma-keys> 1 + >karma-keys ; 07:57:21 how're you doing today, klooie? 08:01:46 hey clever 08:01:51 i just got a 20" monitor, too! 08:02:27 aaron - cool! I think I've seen one of those :-) 08:02:33 lol 08:02:36 doubt it 08:02:38 it's 13W3 08:02:39 SGI 08:16:26 * aaronl is away: breakfast 08:18:17 --- quit: rob_ert_aw_ay (Read error to rob_ert_aw_ay[h3n1fls33o898.telia.com]: Connection reset by peer) 08:27:16 now THERE is an interesting statement (in c.l.f) 08:27:38 Elizabeth D. Rather: 08:27:58 ... ? 08:27:59 On 32-bit processors, 08:27:59 subroutine threading with in-line code expansion is usually smaller, as much as 08:27:59 25% smaller, as well as being much faster. 08:28:08 [end quote] 08:28:21 huhn. 08:28:39 that hadn't occured to me 08:29:01 in-line of some words, certainly. Enth inlines a number of small words where the inline is actually smaller than the alternative. 08:29:26 enth is subroutine threaded? 08:29:52 Yes. Actually, I may be more correct in saying 'Flux'. I know that the Flux subsystem does that, I dunno about the Enth proper. 08:30:12 then the alternative is "5 bytes" 08:30:25 a one byte call, and a 32bit address 08:31:04 oh, that's right, Enth may not be running in a 32 bit flat environment 08:31:40 I keep thinking Enth runs under windows or Linux ... I'm going to really shoot myself in the foot one day with that. 08:31:45 please explain "flat environment" 08:31:59 non-segmented 08:32:10 Yeah, Enth is in that. 08:32:21 With 32-bit numbers. 08:32:34 yes, w/ 32 bit address space 08:32:54 it is still possible to use segmentation win 32 bit protected mode 08:33:04 but things get REALLY confusing 08:33:12 * cleverdra shrugs. 08:34:03 when I wrote my JSR forth, it was in the DOS environment ... 16bit, segmented 08:34:06 I'll have to investigate Enth to be sure about what's going on, but the comments above this inlining code states that these words are inlined because the inline is smaller. 08:34:27 the comment. 08:34:41 that meant that a call was 3 bytes ... which very few words are smaller than 08:35:08 The calls here are certainly five bytes. 08:35:22 in a 32 bit environ, a call is 5 bytes ... and I'd guess that 1/2 the primites are smaller than that, and would get inlined 08:36:04 peephole instruction reduction would reduce that even more 08:37:48 --- join: rob_ert_aw_ay (robost86@h3n1fls33o898.telia.com) joined #forth 08:38:08 --- nick: rob_ert_aw_ay -> rob_ert 08:38:39 --- mode: cleverdra set mode: +o rob_ert 08:38:51 hello again rob_ert. 08:38:58 Hi 08:39:04 wb, rob_ert 08:39:10 Thanks =) 08:48:36 --- quit: nibble (Ping timeout for nibble[host213-122-69-163.btinternet.com]) 08:56:10 --- join: pi (dan@13.portland-05-10rs.or.dial-access.att.net) joined #forth 08:57:49 hi pi 08:58:04 greets, pi! 08:58:18 --- mode: cleverdra set mode: +o pi 08:58:28 hi 09:00:30 * aaronl is back (gone 00:44:04) 09:00:39 hihi 09:01:25 hihihi 09:02:57 hehehehihihihohoho 09:03:06 go go go joob 09:07:38 i am the walrus 09:08:30 see how they fly like pigs from a gun, see how they run, I'm crying 09:11:58 --- join: futhin (thin@h24-67-113-99.cg.shawcable.net) joined #forth 09:12:12 hello all, hello cleverdra 09:12:28 hi 09:12:32 --- mode: rob_ert set mode: +o futhin 09:12:33 hi rob 09:12:45 i did cleverdra's job ;) 09:12:48 can somebody tell me how to open files in forth? 09:13:02 like load files into blocks or something? 09:13:11 how do i access a file? 09:13:29 No idea =( 09:13:36 hmm.. ok 09:14:08 yes, I can 09:14:14 which forth are you on? 09:14:29 win32forth, but i'm going to need to run the program on linux later 09:14:52 ok, you should be able to do that with ANS commands in w32f 09:15:22 hm.. which? 09:15:30 : X .. S" TEST.FTH" R/W OPEN-FILE ABORT" OPEN-FILE FAILED" ... ; 09:15:44 okay cool 09:15:57 here is the stack pic for open-file ... 09:16:11 ( c-addr u fam -- fileid ior ) 09:16:19 Open the file named in the character string specified by c-addr u, with file access method indicated by fam. 09:16:19 The meaning of values of fam is implementation defined. 09:16:30 If the file is successfully opened, ior is zero, fileid is its identifier, and the file has been positioned to the start 09:16:30 of the file. 09:16:43 Otherwise, ior is the implementation-defined I/O result code and fileid is undefined. 09:17:31 okay, so once you open=file how do you access the stuff in the file? 09:17:40 then there is READ-FILE ( c-addr u1 fileid -- u2 ior ) 09:18:03 and READ-LINE ( c-addr u1 fileid -- u2 flag ior ) 09:18:14 hm 09:18:44 and WRITE-FILE ( c-addr u fileid -- ior ) 09:19:03 and WRITE-LINE ( c-addr u fileid -- ior ) 09:19:18 how are you getting the stack comments? 09:19:24 or are you just typing it out? 09:19:42 and, finally, CLOSE-FILE ( fileid -- ior ) 09:20:01 I'm typing the words, and pasting the stack effect from the dpANS document 09:20:26 i don't have to open-file first before using read-file right? 09:20:44 you DO have to open the file before reading from it 09:20:48 ah 09:20:58 ok 09:21:05 otherwise you have no "fileid" to use 09:21:11 ok 09:21:22 hm 09:21:34 on unix, you can cheat by using "1" for fileid, which should send to stdout 09:21:38 * cleverdra returns. 09:21:43 so read-file does what? 09:21:45 but that's REALLY bad practice 09:21:59 READ-FILE ( c-addr u1 fileid -- u2 ior ) 09:22:03 like after i use read-file how do i access the text within the file? 09:22:03 ok ... 09:22:12 like, does read-file put the stuff into a block or buffer or something? 09:22:18 futhin - www.forth.org has a link to the dpANSForth standard which describes all these words. If you have gforth (for example), it's documentation also describes file words 09:22:20 give read-file the address of a buffer and its size 09:22:29 futhin - yes, one of the arguments to read-file is the buffer 09:22:34 ah 09:22:42 that is c-add and u1 09:22:56 indicate which file you want it to read from ... that's fileid 09:23:11 how do i know what address is safe to give read-file? 09:23:18 futhin - if you're on unix, look at `man 2 read`, as read-file is basically a wrapper for that syscall. 09:23:28 it return ior on top of stack ... 0 unless a catastrophic error occured (like wrong fileid) 09:23:35 futhin - typically, you create a buffer first. 09:23:42 how? 09:23:51 under ior is the number of bytes that read-file actually read 09:23:51 futhin - CREATE and ALLOT work. 09:23:56 ok 09:24:25 u2 might be less than u1 if you attempted to read more chars than available 09:24:49 create filebuf 257 chars allot 09:25:04 if I'm writing an application, I usually use PAD for my buffers 09:25:10 filebuf 1+ 256 some-fileid read-file throw filebuf c! 09:25:18 i need to basically read a line in from the file, match it, and repeat with next line.. thru the whole file.. 09:25:38 futhin - what are you trying to do? 09:25:55 processing an eggdrop textfile 09:26:06 irc log 09:26:33 actually I should've said char+ instead of 1+ :-/ I'm trying to do ANS code anyway -- but then I shouldn't used CHARS and such. You get the idea though from that code, right? 09:26:45 futhin - processing is a very general sort of operation 09:26:47 my friend wants me to code it so that he can see if forth is fast or not, cause he's coded a perl processor :) 09:27:04 he wants to compare speed heh 09:27:38 ok, in that case ... 09:28:01 futhin - o/~ this is not a good comparison o/~ 09:28:03 open the file ... get its size ... allocate a buffer for the entire file ... read the whole file in 09:28:03 i'm just matching each line to something, like a public message would search for < > in and return that to caller (we haven't decided how to treat the matching yet) 09:28:12 futhin - what MrReach is saying. 09:28:55 what happens if the file is bigger than ram? 09:29:00 close the input file, open the output file ... start parsing 09:29:03 than ram available 09:29:04 futhin - then it doesn't fit, see? 09:29:15 and then what happens? it fails? 09:29:21 futhin - painfully. 09:29:30 then you will get an error from ALLOCATE ... but that is incredibly unlikely on Win32 or Linux 09:29:34 futhin - well, not really. It obviously won't work. 09:29:44 hm. Yes, what MrReach said. 09:30:36 I have 128MB of swap on both machines, so I can prob read 10MB files into memory w/o too many probs 09:30:37 : make-fbuf: ( fileid -- ) file-size throw allocate throw create , does> @ ; 09:31:06 does file-size return an ior? 09:31:13 I don't have documentation on this machine. 09:31:16 I believe, lemme check 09:31:25 i'm not quite sure how the perl program works, but i think it generally reads in all of the file, and if that doesn't fit, it probably does some automatic swapping as necessary?? 09:31:42 ( fileid -- ud ior ) 09:31:44 futhin - Perl probably does *not* itself do swapping. 09:31:48 thanks MrReach. 09:32:04 : make-fbuf: ( fileid -- ) file-size throw d>s allocate throw create , does> @ ; 09:32:05 well my friend said the os would do the swapping itself.. 09:32:16 WARNING: the "ud" on the stack is a double ... two stack entries ... the uppper contains the most significant 32 bits 09:32:23 or mayble 1+ allocate... 09:32:28 futh, yep, it prob would 09:32:48 so how large is the largest file? 09:32:57 prolly 8.5 megs 09:33:07 which should be fine i guess 09:33:21 dired, 8.5 * 1024 * 1024 09:33:22 ok, open w32forth ... I'll do the same here 09:33:22 8912896 09:33:32 That'll fit in a 32 bit number. 09:34:13 8912896 allocate .s [2] 6553612 0 ok.. 09:34:41 my computer just allocated 8,912,896 bytes without barfing 09:35:04 yeah, same here 09:35:09 your outer interpreter prints DEPTH dots too? =) I love that. 09:35:13 i've got 32 megs of ram :) 09:35:25 now type "drop free drop" 09:35:28 That used to be Enth, until Enth had the last line of the screen set up to display the stack and depth constantly. 09:35:42 I've got 128MB, and the same in swap 09:35:54 hm, cool 09:36:27 20000000 allocate .s [3] 0 24707084 0 ok... 09:36:44 that just allocated 20megs on my machine, no prob 09:37:22 nevermind the extra 0 on the stack 09:37:48 now type "drop free drop" 09:37:50 --- quit: pi (Read error to pi[13.portland-05-10rs.or.dial-access.att.net]: EOF from client) 09:38:38 jeesh this channel is almost a REAL channel with the number of ppl in here hehe 09:38:41 how does your machine handle it? 09:38:47 --- nick: rob_ert -> rob_ert_aw_ay 09:38:52 heya, I440r 09:38:56 --- nick: rob_ert_aw_ay -> rob_ert 09:38:58 20000000 allocate .s [2] 15532044 0 ok 09:39:02 and there's actually some source in here 09:39:10 I think channels are always real, unless they aren't. 09:39:37 if you advertise this channel on clf then we might get too many people :P 09:39:39 futhin - congratulations! 09:39:50 ok, futhin, does that convince you that you're prob not going to bomb out while slurping up a big text file? 09:40:13 the 20meg failed to 15 megs didn't it? :) 09:40:31 : make-fbuf: ( fileid -- ) file-size throw drop 4 + allocate throw create , does> @ ; 09:40:34 hi :) just got up - gota get a coffee 09:40:43 most OS's are most efficient when reading large ammounts of data atomically ... that is in one system call 09:41:05 some-fileid make-fbuf: THIS-BUF \ cell-count plus all the bytes in the file 09:41:20 cleverdra: I would create a word that accepts a filename ... and returns a memory address of the buffer holding that file 09:42:16 MrReach - yeah, that'll work too. ( fileid -- buf len ) ... actually a better idea. You can pass that straight to one of the string-searching words. 09:42:39 ok, um, different question now 09:42:49 Isn't there actually a ANS word that'll do what futhin wants? To search for a substring in a string? 09:43:02 futhin - OK. 09:44:30 suppose i read in a line from the file and it looks like " hi there!" i need to match the <> before i can say it's a public message. i should also probably recognize the as separate (delimited by space) 09:44:46 --- nick: rob_ert -> robert 09:44:50 --- nick: robert -> rob_ert 09:44:57 : slurp ( adr cnt -- buf-adr ) R/O OPEN-FILE THROW DUP FILE-SIZE THROW DUP ALLOCATE THROW ( fileid size buf ) ... 09:45:04 I don't understand the last six words of your question, futhin. 09:45:30 what futhin wants .... 09:45:31 uh " hi there!" the has a space after it 09:46:00 wh ois jennifer ?? :) 09:46:02 yeah. I see that. Are you distinguishing that from cases where does not have a space after it? 09:46:05 --- nick: rob_ert -> rob 09:46:09 is a match on "\n<[::PRINT::]>\w.*" 09:46:11 Do you just want to search for at the beginning of a line? 09:46:12 mrreach explain throw to me again ? 09:46:20 no.. will always have a space after it 09:46:46 it realy doesnt seem like a real forth word because forth words tend to be very descriptive of thats going on and i cant picuter forth throwing anything heh 09:47:06 --- nick: rob -> rob_ert 09:47:29 i'm saying, i need to match the <> in the first "word" which is "" and the first word is obvious because it has a space between it and the rest of the text "hi there!" 09:47:29 I440r: if TOS is 0, do nothing and continue execution ... otherwise unnest the return and parameter stack back to the nearest CATCH, leaving the TOS as it's output 09:48:00 I440r - throw is part of the catch/throw mechanism implemented by dpANSForth for error handling in Forth programs. It's stack comment is ( n -- ). If n is 0, nothing happens. If it is either positive or negative something happens. 09:48:14 I440r: it means "throw an error to whoever is catching" 09:48:32 Oh, I didn't know about it leaving TOS when caught. I don't use throw much. 09:48:51 I440r - some other languages use throw/catch like this 09:48:52 abort" is easier :P 09:48:57 cleverdra: I think it does, lemme check 09:49:08 but that leaves just one error handler 09:49:16 I440r - yeah, that's what many people said =) Others said that throw/catch was needed. 09:49:17 um, how do i match the < > in "" ?? i can't use WORD? 09:49:21 I440r: abort" is usually implemented using throw, on systems that have throw 09:49:34 Does anyone know about NETS? It was an alternative error-handling mechanism submitted to ANS, that was thought to be cooler. 09:49:39 yea i recall that from the last time we talked about it 09:49:42 futhin: I think you need MATCH 09:49:58 see match 09:49:58 Error: match is undefined 09:50:04 futhin - please stop saying ``< > in ""'', it's confusing. 09:50:23 futhin - you want to do something with all lines in the file that begin with , yes? 09:50:31 i need to match the "<" and the ">" in "" does that help? :) 09:50:39 no 09:50:43 uh 09:51:11 futhin: I think we have all see IRC logs ... @:^> 09:51:13 no it's not, because you don't define 'match' and I don't understand how you could match "<" or ">" in "" specifically or what purpose this has. 09:51:19 these are all the different lines that i'm matching: 09:51:19 see match 09:51:19 Error: match is undefined 09:51:21 er 09:51:29 \join ([^ ]+) \([^ ]+\) joined/ && 09:51:29 \part ([^ ]+) \([^ ]+\) left #/ && 09:51:29 \quit ([^ ]+) \([^ ]+\) left irc: (.*)/ && 09:51:29 \kick ([^ ]+) kicked from #[^ ]+ by ([^ ]+): (.+)/ && 09:51:29 \mode #[^ ]+ mode change '([^']+)' by ([^!]+)/ && 09:51:30 \nick Nick change: ([^ ]+) -> (.+)/ && 09:51:32 \duplicate Last message repeated (\d+)/ && 09:51:34 \ 09:51:36 uh 09:51:38 ugh 09:52:08 futhin - do you want to find an arbitrary name? and and , but all in <> pairs and at the beginning of lines? 09:52:14 yeah 09:52:18 ok =) 09:52:35 COMPARE ( c-addr1 u1 c-addr2 u2 -- n ) 09:52:35 an arbitrary name, i need to recognize the < and the > in those names before they are "matched" :) 09:52:53 i don't know the names ahead of time 09:53:02 all i know is that they have < and > 09:53:03 futhin - OK, we've told you how to slurp a file into a buffer, now one way to do what you want is to step through the buffer by line and compare the beginning chars of it with COMPARE ( Thanks MrReach ) and your string. 09:53:31 futhin: ok, then ... compare one possibility at SOL, then the next, then the next 09:53:37 futhin - you *must* know the name when you're searching, otherwise "searching" is pointless. Please define "ahead of time". Do you know them at runtime? That's early enough to use COMPARE. 09:54:00 when a possibility is matched, process the line (however) ... search for beginning of next line 09:54:20 cleverdra: no, he doesn't need the name 09:54:22 argh 09:54:32 heh, i'm not being very clear on this 09:54:35 cleverdra: he only needs to know it's a "talk" line 09:54:39 : next-line ( buf - buf' ) begin dup c@ nl = ?exit char+ again ; 09:54:54 rather than a channel join or part, or mode change, or whatever 09:55:00 i'm taking in a line, and trying to determine if it's a public message, or if it's an action, or whatever. 09:55:08 futhin - OH! You want to search for lines that have , where foo is any time? You don't want to match specific names, but all names? 09:55:20 * MrReach laughs. 09:55:37 yeah, i want to match the "<" and ">" i think 09:55:38 OK, that's a bit easier. Do "talk" lines start always with '<'? Are there any other lines that start with that? 09:55:59 I suspect that futhin is processing eggdrop logs into HTML files for channel logs 09:56:19 futhin: is that right? 09:56:21 * cleverdra has never seen an eggdrop log. 09:56:30 here's an example 09:56:30 [00:00] nopers 09:56:31 [00:00] he's kawaiiiiiii 09:56:31 [00:00] Action: lulu comforts Nezzie 09:56:36 it looks very similar to mIRC logs 09:56:46 futhin - so, the timestap is part of the lines? 09:56:50 yes 09:56:57 but i can use an offset of 7 09:56:59 I suspect that futhin is processing eggdrop logs into HTML files for channel logs 09:57:01 futhin: is that right? 09:57:05 nope 09:57:27 i'm just matching lines and returning that to the "caller" 09:57:36 : +offset ( addr -- addr', account for [00:00] ) 8 + ; 09:57:47 oh, ok ... I rewrote the log portion of the eggdrop bot to write in HTML, and to start a new file at midnight 09:58:11 : is-talk-line? ( addr -- f ) c@ [char] < = ; 09:58:22 uh 09:58:26 MrReach - cool =) 09:59:13 c@ returns char at offset 9? 09:59:28 no, the addr given to is-talk-line? will be offset 09:59:33 wait a sec =) 10:00:00 futhin it doesnt return a char - it returns one stack item - it fetches a char -and zero extends it to however wide one stack item is 10:00:06 16 bits or 32 or waht ever 10:00:33 offset 9 is the "<" IF the line has it.. (i'm simplifying the address & offset for the moment) 10:00:58 what is [char] ?? 10:01:29 is [char] like ascii ? 10:03:08 hey I440r 10:03:09 : processs-talk-lines ( bufaddr n xt -- ) >r s>end+begin begin dup offset+ is-talk-line? if dup r@ execute then next-line 2dup = until 2drop rdrop ; 10:03:37 ugh 10:03:39 heh 10:03:40 isnt rdrop r>drop ? 10:03:46 yeah 10:03:58 [char] is like ascii I think 10:04:00 okk - thunked maybe they had changed it hehe 10:04:28 "[char] n" is a bit like "[ char n ] literal" 10:04:35 : foo ascii a emit ; 10:04:36 [char] = Skip leading space delimiters. Parse name delimited by a space. Append the 10:04:36 run-time semantics given below to the current definition 10:04:47 the a character is compiled as a literal 10:05:06 : s>end+begin ( addr n -- addr-end addr ) over + swap ; 10:05:10 i440r - yeah. 10:05:42 ruthin dont read the ans spec for descriptions of what words do - its written in gobbldegook :P 10:05:43 futhin - ick. Get some good documentation =) the dpANSForth stuff I only touch when I can't find anything else. 10:06:06 yeah, i'm annoyed that i can't type "man " in win32forth :P 10:06:16 --- nick: rob_ert -> rob_ert_ver_y_far_aw_ay 10:06:16 where do i get good documentation? 10:06:21 futhin - some Forths have something like that. 10:06:43 futhin search ftp.forth.org for fprimer.zip 10:06:47 futhin - uh, I like the gforth docs. Categorized by word function. 10:06:53 thas all dos related tho - 10:07:07 futhin - do you see how you'd do this, now? 10:07:30 not quite.. how do i match ? :) 10:07:44 c@ [char] < = 10:07:48 cleverdra: i want every single word in isforth described in detail on the isforth web site - 10:07:53 and examples of their use given 10:08:17 i440r: and example code so that lamos can just cut'n'paste :P 10:09:11 :) 10:09:17 like in turbopascal help, you can like get help on a particular command and it'll give you code using that particular command, if it happened to match what you wanted to do with it (and it usually did) then all i had to do was cut'n'paste it :P 10:09:19 futhin - you're not matching , you're matching the '<' that begins with and that no other line will begin with. If you want a *specific* , you'll have to write another word and maybe generalize PROCESS-TALK-LINES some more 10:10:00 I440r - that's a good plan. Documentation is *** GOOD ***. 10:10:13 and necessary! 10:10:35 not necessary, just good. Very good. Good with cherry on top. Does anyone know a better word than 'good' here? =) 10:10:36 I440r: it would help if you worked on isforth :-/ 10:10:49 cleverdra: yes :) 10:11:02 aaronl: i want to but i need tcn to gimme his stuff again doh! 10:11:13 Heheheheh 10:11:16 I440r - is isforth supposed to work now? I got it compiled a few days ago, but I can't even tell if it is reading stdin. 10:11:35 cleverdra: the version you have doesnt work 10:11:41 the version tcn has is starting to 10:12:34 ah, OK. 10:12:42 see why i need it ? hehe 10:13:40 futhin: one you know that the char at the proper offset is "<", then yo KNOW the line is a talk line ... it's up to the processing part to extract the name 10:15:36 the name can be extracted by looping through lines looking for a ">", when it is found, return the address of the start of the name (the character after "<") and how many chars are in the name 10:15:52 looping throught the characters, rather 10:16:00 mrreach why not just use parse ? 10:16:04 or word 10:16:09 ascii > word 10:16:09 ? 10:16:13 futhin - yeah, you see the xt stuff I'm doing in process-talk-lines? That's so you can call arbitrary Forth words upon an address where "<" is at. Then you can parse for a ">" and print the name or, or print the rest of the line out, or do whateer. 10:16:18 that would work 10:17:08 I440r: I've spent a lot of thought in how to document in source files ... do you have a system that you use yet? 10:17:33 i have some general guidelines that i try to follow like 10:17:34 dont do 10:17:35 code 10:17:36 comment 10:17:37