00:00:00 --- log: started forth/05.01.09 00:28:54 --- join: p-Imperator (CapStone@cs24160141-160.satx.rr.com) joined #forth 00:28:54 --- quit: Testament (Read error: 104 (Connection reset by peer)) 00:56:29 --- quit: Sonarman ("leaving") 01:21:48 --- join: zol1 (~l@ppp-62-245-209-126.mnet-online.de) joined #forth 01:25:45 i've put a version of the wiki markup to html translator on a web server, actually running it as cgi there: http://fwiki.logilan.com 01:44:02 --- join: Frek (14402-iden@h229n2fls31o815.telia.com) joined #forth 02:14:33 --- quit: Teratogen ("SKYKING, SKYKING, DO NOT ANSWER.") 02:21:38 --- join: bbls (~bbls@80.97.121.133) joined #forth 02:21:39 hello 02:41:21 hi 02:41:55 hi zol1 03:17:19 --- join: aum (~aum@60.234.138.239) joined #forth 03:17:58 i *sooo* need 4 21" LCD monitors side by side 03:18:13 what for? 03:18:42 --- join: qFox (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth 03:18:47 code editor window, 4 simulator windows, this xchat window, various pdf manuals... 03:19:04 :) 03:19:07 what are you coding? 03:19:25 a forth vm for PIC18Fxxx microcontroller 03:20:05 sweet chip, the 18F452 - got the whole VM and a few primitives in just over 512 bytes 03:20:17 :) 03:21:06 oh, i forgot - assembly listing windows as well 03:27:12 18 machine instruction cycles per vm fetch/execute opcode cycle - must be a way to bum it down 03:27:49 hmm, i've found something 03:28:03 the name of the os bought by microsoft was QDOS 03:28:12 Quick and Dirty Operating System 03:28:14 haha :) 03:29:16 to this day, windows remains a poorly debugged, gui-tarted hack of ms-dos, which is a hack of pc-dos, which is a hack of cp/m... 03:29:45 not pc-dos, qdos 03:30:09 guess it fits though - pentium is a hack of 486, which is a hack of 386, hack of 186/86, hack of 8080, hack of 4040, hack of 4004... 03:30:39 well, i'm not sure if all backward compatibility means a hack 03:34:51 aum, afaik does the pic17xxx (same as 16xxx or 15xx) not use bytes for instructions 03:34:58 but 12, 14 and 16 bit words 03:35:11 i.e. one address increment = 2 bytes of 8 bits 03:35:20 pic16f uses 14-bit words 03:35:31 pic18f uses 16-bit words 03:36:17 maybe it would make sense to talk about code size, using bits as metric ? 03:46:55 --- quit: aum () 04:38:54 crc, you're running ubuntu linux ? 04:39:00 yes 04:39:24 had a look (throuhg gnoppix) at it 04:39:39 but i preferred gnoppix when it was still debian based 04:40:12 gnoppix isn't debian based anymore? 04:40:29 it is ubuntu based now 04:40:41 ubuntu is debian based. 04:40:59 ubuntu sounds so black africa :) 04:41:02 ah 04:41:16 The name is an african word ;) 04:42:29 ah 04:42:35 * crc will only use Debian-based Linux distros, or Slackware 04:43:10 the image links of wiki markup processor don't show now because the images are on forthfreak.net (which is just being built on) 04:43:51 web server turned off because of sql local compilation to distro-based change 04:46:06 wil be back to operation in < 15 minutes i was told 04:47:05 is online again now 04:47:23 how good is suse? 04:47:40 i returned after a short detour to debian quickly 04:47:52 why? 04:48:12 i wans't pleased with several things, such as the configuration manager 04:48:28 either you turn it off, or you lose manual changes to config files 04:48:39 ah 04:48:57 also, the remote installation stuff (rpm based) is/was much worse than debians 04:49:06 there's now better tools for rpm 04:49:13 but not at that time 04:49:26 i've used red hat before, now i have knoppix 04:49:36 SuSE feels more bloated than Debian 04:49:41 3rd point i don't like is that suse patches the kernel, and it is not vanilla compatible anymore 04:50:09 installation of suse without X seems to be impossible 04:50:18 That can be done 04:50:21 (well, you can delete of course) 04:50:35 so it's a linux for windows users? 04:50:36 but by just using the installation management i didn't manage 04:50:40 yes 04:50:57 You have to boot in "expert" mode, choose the text-only YaST tool, and proceed to manually select the base packages 04:51:14 whatever i did, it has seen X as required 04:51:21 heh 04:51:32 that was suse 5.2 or 5.3 04:51:49 I used 6.4 and 7.1 ;) 04:52:00 there was no yast at that time 04:52:04 no yast2 i mean 04:52:13 yast2 sucks. 04:52:18 only yast, the text mode version 04:52:21 I liked the original yast better 04:52:36 i like editing config files better :) 04:52:43 hehe 04:53:07 yast is too much an "all or nothing" solution 04:53:10 Well if you had to use a configuration tool, yast wasn't as bad as yast2 is 04:53:12 I agree 04:53:51 * crc remembers having to go through yast to set up everything or you'd lose your changes after installing/removing packages 04:54:57 nowadays suse users have among linux users a bit the status of AOL users among the rest of us 04:55:25 "too dump to install a real distribution" 04:55:29 dumb 04:55:52 It's not as bad as Lindows or Desktop/LX though. 04:56:04 right, lindows seems to be worse 04:56:14 never tried it though 04:56:42 it's debian based, but geared towards windows users 04:56:50 KDE as the default desktop :( 04:57:20 no matter what desktop 04:57:27 there's still ctrl-alt-f1 04:57:45 is it with lindows ? 04:58:11 or do they protect user sanity ? 05:06:28 I think that's still allowed 05:06:49 But the default account is 'root' with no password IIRC 05:07:28 less brain-strain 05:07:44 heh 05:07:59 On ubuntu, they do away with a 'root' entirely 05:08:25 You have to use 'su' or 'sudo' to temporarily gain root access 05:09:05 what about a suid wrapper, setting uid 0 and calling a shell ? 05:09:30 You could do that 05:09:46 I do 'sudo /bin/sh' often enough 05:11:11 main() { 05:11:11 setuid(0); 05:11:11 seteuid(0); 05:11:11 execl("/bin/sh","","-i",0); 05:11:11 } 05:11:21 something like that could do 05:12:07 how would i write that in forth ? 05:12:59 what are the syscall numbers for each function? 05:13:12 the user ids 05:13:20 ? 05:13:45 (Using RetroForth/GenericDL this should be easy) 05:14:17 execl would be system oir the like 05:14:30 but i have no idea how to set user id from forth 05:15:24 i only can guess what i really does to the started process 05:15:41 (i estimate it patches the credency record) 05:17:08 credence record it seems to be called 05:19:12 that doesn't seem to leave the uid and euid once execl is called :( 05:28:57 I can use sudo though... 05:28:58 " /lib/libc.so" LoadLibraryAs libc 05:28:59 libc " system" MapFunctionAs system 05:28:59 : (zt) over + 0 swap c! ; 05:28:59 : exec 10 parse (zt) 1 system cinvoke ; 05:29:00 exec sudo sh 05:30:13 right 05:30:36 * crc will be simplifying the FFI in retro8 05:30:51 or suid a shell which doesn't protect against, like bash does 05:31:16 which is the only reason for the wrapper 05:31:36 How can I suid a binary? 05:31:41 chmod 05:31:47 chmod *what* 05:31:55 2xxx or 4xxx 05:33:52 chown root.yourgroup file ; chmod 4650 file could do 05:34:13 yourgroup so you got file execution privs 05:35:32 that works ;) 05:35:50 Now I have a Forth shell running as root 05:36:07 if you break it you onw both pieces 05:36:12 own 05:36:19 --- part: bbls left #forth 05:36:38 this is cool 05:37:09 * crc should do some work on expanding his shell-in-forth code 05:37:41 simple tools like "cat" "ls" "more" and all built in would reduce the reliance on external tools. 05:37:54 right 05:38:05 then comes sed 05:38:08 *g* 05:38:54 how do you handle regexprs ? 05:40:00 do you know busybox ? 05:41:54 --- nick: zol1 -> zol_afk_30min 05:52:00 I don't 05:52:28 Actually, I could, but I have no reason to at this point 05:52:58 I'm aware of busybox 05:53:07 I just dislike C for the most part 06:21:26 --- join: bbls (~bbls@80.97.121.133) joined #forth 06:21:33 what's the best forth for win32? 06:22:54 RetroForth ;) 06:23:04 i need win32 functionality 06:23:08 (mainly opengl) 06:23:10 Or Win32Forth for a bloated Forth with bindings to pretty much all API's 06:23:22 RetroForth/Windows supports OpenGL 06:24:13 http://retroforth.org/board/index.php?board=5;action=display;threadid=150 06:24:28 Alexey has been doing a lot of work with OpenGL and RetroForth 06:25:12 does opengl support exists in the main package? 06:25:17 or i have to download some extras? 06:25:21 no 06:25:43 Alexey's package (on that link I posted) has RetroForth with the OpenGL files 06:27:39 --- nick: zol_afk_30min -> zoly 06:58:48 --- join: Testament (CapStone@cs24160141-160.satx.rr.com) joined #forth 07:08:04 gtg, bye 07:08:04 --- part: bbls left #forth 07:16:09 --- quit: p-Imperator (Read error: 110 (Connection timed out)) 08:45:54 fwiki markup supports named image links now : [http://destination.domain http://server.domain/image.png] displays image.png, connecting to destination when clicked. 08:46:54 test and demo here: http://fwiki.logilan.com/ 09:26:38 back 09:27:08 wb 09:29:19 zoly: it's looking good 09:29:57 yes, not too bad. though i'm not 100% satisfied 09:30:03 but it is first attempt anyway 09:30:07 Yup 09:30:21 it's more complete than RetroWiki though 09:31:02 this implementation is just to spot problems, to know how to really factor it 09:31:09 call it a prototype 09:31:12 ok 09:31:27 also allows me to experiment with syntax 09:32:12 ." hello" 09:32:14 hellook 09:32:36 Should I have the prompt code spit out a space or a CR before displaying "ok"? 09:32:57 you decide 09:33:01 though i prefer cr 09:33:15 i don't like ok tucked away behind last output 09:33:37 that makes sense 09:33:47 but i hate empty lines between oks as well 09:33:52 hmm 09:50:40 --- join: p-Imperator (CapStone@cs24160141-160.satx.rr.com) joined #forth 09:50:40 --- quit: Testament (Read error: 104 (Connection reset by peer)) 09:50:49 --- join: Topaz (~top@sown-85.ecs.soton.ac.uk) joined #forth 10:21:43 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 10:25:10 --- join: tathi (~josh@pcp01375108pcs.milfrd01.pa.comcast.net) joined #forth 11:06:49 are here users with other browsers than firefox who want to do a quick test of input forms for me ? 11:07:08 I could try Dillo 11:07:22 ok please do 11:07:35 http://fwiki.logilan.com/cgi 11:07:43 add the end of file is an Edit button 11:08:03 --- quit: Topaz (Remote closed the connection) 11:08:06 posting contents is not finished 11:08:38 The form shows; no contents in it 11:08:43 would interest me to know whether the edit form works 11:08:47 right, empty 11:09:22 I could enter text, clicking "save" gives a 404 about "save.cgi" not existing 11:09:29 cirrect too 11:09:31 correct 11:09:33 Preview doesn't work either; just clears the input box 11:09:41 seems to be fine 11:09:58 One thing: the relative path to edit.cgi didn't work; I had to add "edit.cgi" to the URL to get to it 11:10:18 hmm 11:10:29 That could be a problem with Dillo though 11:11:08 hmm...the edit.cgi link works in Lynx 11:11:11 I don't think it should though. 11:11:35 I'm getting Edit 11:11:56 right. supposed to mean "in the dir of the current server.domain 11:11:59 " 11:12:06 you shouldn't have the http: 11:12:12 that's not a valid url I don't think. 11:12:15 That's not valid html IIRC 11:12:17 ah 11:14:10 yes, links makes something else from that url again 11:14:28 /cgi/http:edit.cgi ... 11:15:01 :) 11:15:19 heh 11:16:04 yah, if you've got http you also need "://host/path" at least. 11:18:40 hmm, yeah, looks like the RFC says you never put the scheme (http/ftp/whatever) in a relative URL. 11:19:00 http://www.ietf.org/rfc/rfc1808.txt 11:28:21 crc, works again now. 11:28:32 no white screen no more 11:29:35 still has a bunch of broken links 11:29:45 just moved some files to other dirs 11:29:52 you tried the source ? 11:29:56 Yes 11:30:00 that's fixed this very moment 11:30:40 http:/pages/wikiwords 11:30:48 I just reloaded it :) 11:30:53 Still broken on those links 11:31:23 --- join: Topaz (~top@user-71-194-151-83.e7even.com) joined #forth 11:31:27 probably same reason as the Edit link problem. is also a relative link with leading http: 11:31:36 yup 11:31:52 the edit link works now 11:32:15 wikiwords link too 11:33:02 the link to fwiki.forth is broken still 11:33:11 Other than that, it looks good 11:33:49 fixed too 11:33:58 good 11:34:14 nice. working base. 11:34:41 user- and form interaction would be the next 11:34:43 yup 11:38:57 if you want to make an html link to a file foo in the top level directory for the domain, you do 11:39:59 right 11:41:41 you should only use the "http://hostname/" bit if you're linking to a different domain, different protocol, or for some reason you have to generate a redirect. 11:42:12 indeed i don't need them for links generated by markup>html. 11:42:26 but the markup uses http: for link detection 11:42:39 --- quit: I440r (Read error: 104 (Connection reset by peer)) 11:44:02 but as those are 2 different things, it can just ref dirs and files by /... 11:44:09 (done) 11:44:35 except with the Edit button 11:44:42 which is a named link 11:44:47 (currently) 11:45:05 it needs the http: portion 11:45:26 or fwiki markup wouldn't recognize it as web link 11:47:08 how many forth specific wikis are there ? 11:47:22 herkamire, retro, forthfreak 11:47:29 forth-ev 11:47:36 Those are the ones I know of 11:47:37 (broken wiki, that is) 11:47:56 c2 11:48:03 not specific but with forth sections 11:48:06 c2 isn't forth-specific though. 11:48:20 There's a forth section on factor.sf.net/wiki 11:48:25 I just check for : to detect non-wiki links 11:48:45 hehe 11:48:51 attacked by link spammers 11:48:52 because you may well see https: ftp: mailto: 11:48:57 and many obscure ones 11:49:11 forthfreak:ForthImplementations 11:49:13 sftp: telnet: mmx: etc 11:49:21 * crc does custom ones ;) 11:49:23 besides, ":" is way easier to look for 11:49:57 news: 11:50:48 the http:... is only used for named links 11:51:07 i suppose for other URIsit would be allright to use unnamed links 11:51:19 so the user can see what kind of link it is 11:51:33 ehm 11:51:37 not true 11:51:52 RetroWiki's flaw is that the wiki pages are Forth source; security issues abound. 11:52:18 indeed 11:53:26 what do you mean named links? 11:53:39 wasn't it onetom who had a forth interpreter as irc bot online for a while 11:53:50 many of us did 11:53:51 which got hacked every other day ? 11:54:09 onetom hooked up gforth running as root :) 11:54:17 named links: not url shows in text, but substitution text 11:54:21 I don't know anything about one getting hacked 11:54:27 such as [http://url other text] 11:54:30 oh, I don't do that 11:54:36 I like knowing where links go to 11:54:37 RetroBot will be coming back soon I think 11:55:08 i think first he forgot that there was a "system" word :) 11:55:43 my impression with onetom was that he didn't really care that much. 11:56:38 I just do 11:57:15 I do c2[ pagename] 11:57:27 or [ pagename] or forthfreak[ pagename] 11:57:29 i'm kwiki-oriented 11:57:34 same wiki as forthfreak 12:07:48 I'm rather flexible on syntax 12:08:15 two reasons i tried to settle for kwiki syntax 12:08:45 a: if every wiki a user visits uses a different markup, that can be quite annoying 12:09:03 b: should forthfreak pages ever be migrated to fwiki, the effort would be much lower 12:09:12 Both are good reasons 12:09:35 My wiki has only been hacked once, likely due to the bizarre syntax 12:10:01 hacked = link spammed / defaced ? 12:10:10 I just figured that wikis tend to have different syntax already 12:10:17 and I wanted to try out my syntax ideas 12:10:18 defaced the main page 12:11:35 oh. I thought you meant hacked into the computer 12:11:45 No 12:12:08 Until recently the syscall word was kind of ignored by everyone but a few of us 12:12:32 (Even if I remove it, it's easy to add one again since pages are forth source) 12:12:59 heh 12:13:16 put the markup-relevant words in a voc, and seal it ? 12:13:30 I don't have support for vocabularies. 12:13:48 can be added 12:13:53 Not easily 12:13:58 hmm 12:14:13 usually, high level implementation suffices 12:14:18 You still have to think that I could do: 12:14:23 (if you have vectored create/find) 12:14:31 : foo [ $hexcode , $hexcode , .... ] ; 12:14:36 and get a syscode word 12:14:52 why ? [ is not in the sealed voc 12:14:53 I don't have any vectored words 12:14:58 neither is : 12:15:24 create : -5 allot $code , $code , .... 12:15:31 no create in sealed voc 12:15:35 no allot 12:15:37 You'd have to disable a lot of useful words 12:15:49 no, only NOT useful for wiki markup 12:16:06 zoly: part of my goal was to allow page-specific markups to be created 12:16:23 perfect, put those into the sealed voc 12:16:36 You can't if you don't have access to the compiler 12:17:01 that's what they mean by "you can't have your cake AND eat it" 12:17:05 :) 12:17:51 One of my users even came up with a scheme to enable access to the compiler with just: @ ! + - 12:18:06 no @ or ! in sealed voc ... 12:18:29 how do I write the sealed vocabulary then? 12:18:46 is a normal voc until you seal it 12:19:16 fromon that moment, only words in sealed voc are in voc search list 12:19:39 That would require a lot of changes to RetroForth to implement. 12:19:57 if you create a word to make another voc context, you can break out. 12:20:18 i don't know retro well enough, but vocs can usually be implemented hi-level 12:20:24 no change to underlying system 12:20:26 how? 12:20:35 except where's find involved 12:20:52 a voc is just a linked list 12:20:56 I've never seen a set of vocab words that don't rely on the underlying system 12:21:00 s/is/can be/ 12:21:18 find is hard coded to search the forth and macro wordlists; and nothing else 12:21:18 only thing you want a voc to do is contain headers 12:21:27 and find to be able to search through them 12:21:51 (of course create to create words) 12:22:18 what aspect of these would have to be part of the kernel ? 12:22:27 find 12:22:32 (execution vector for find assumed) 12:22:50 A way to hide the real vocabularies below this 12:22:59 cause you can replace your kernel find with a voc-enabled one in that case 12:23:32 interpret is an assembly word that calls find 12:23:38 So I'd have to vector that 12:23:41 no 12:23:44 ? 12:23:54 if you vector find, interpret still can find through find 12:23:55 So I'd have to write a custom parser/interpreter then 12:24:03 why ? 12:24:05 just find 12:24:09 or don't 12:24:14 Not if I don't vector find 12:24:18 I don't like vectors. 12:24:19 interpreter doesn't need to know how find finds 12:24:50 vectors obscure the code 12:25:13 vector is a way of saying "i'm not bothered with the details here" 12:25:48 My "find" is non-standard anyway 12:25:57 a limited set of vectors is quite useful 12:26:01 like i/o 12:26:05 or boot 12:26:18 i count find to that as well 12:26:24 boot is only essential if you are turnkeying or metacompiling 12:27:09 * crc will not vector words in RetroForth. 12:27:12 if you do, you can take advantage of a vectored boot 12:27:25 zoly: I don't turnkey. 12:27:53 When I do, there'll be a way to do it without making boot a vectored word. 12:28:18 wif your feeling that strong about not vectoring, then you possible can't add hi-level vocavularies 12:28:22 :) 12:28:37 The closest thing I have to vocabularies is localized wordsets: 12:28:39 loc: 12:28:42 : foo ... ; 12:28:45 : bar ... ; 12:28:53 : baz foo .. bar ... ; 12:28:56 ' baz 12:29:01 ; loc alias baz 12:29:28 foo and bar are removed from the dictionary 12:30:08 foo and bar removed means, their headers removed ? 12:30:14 yes 12:30:27 variable (loc) 12:30:36 : loc: last @ (loc) ! ; 12:30:39 a right 12:30:46 : ;loc (loc) @ last ! ; 12:30:52 something between : and :noname 12:30:52 that's it 12:31:00 right 12:31:13 They allow me to factor words out without polluting the dictionary 12:31:31 Makes some things much cleaner overall 12:31:57 debugging may become more difficult 12:32:09 I debug a word at a time. 12:32:20 I only add the loc: ;loc pairs after the words are debugged. 12:32:47 some conditions are sometimes not tested for, and only show later 12:32:57 such as? 12:33:05