00:00:00 --- log: started forth/06.08.28
00:04:34 <JasonWoof> Quartus: Oh, before I go to bed in a sec, I just had a though.
00:04:56 <JasonWoof> Quartus: If you want to jot down some notes about what you had to do to get vor to build on windows I'd include that in the source distro
00:04:57 <ayrnieu> You had a thought?
00:05:00 <ayrnieu> JASONWOOF HAD A THOUGHT
00:05:18 <JasonWoof> 3am here, so they don't come real fast :)
00:05:37 <Quartus> JasonWoof, I'll let you know.
00:05:51 <JasonWoof> mostly just thoughts of sex, gaming and dental care atm
00:06:13 * JasonWoof starts brushing his teeth
00:06:31 <Quartus> what was that make command again?
00:06:33 --- join: segher_ (n=segher@dslb-084-056-165-158.pools.arcor-ip.net) joined #forth
00:06:43 <JasonWoof> mingw-make -f Makefile.win
00:07:08 --- quit: segher (Read error: 60 (Operation timed out))
00:09:33 <Quartus> It immediately fails to find sdl.h
00:09:50 <Quartus> SDL.h that is.
00:10:09 <JasonWoof> oh
00:10:22 <JasonWoof> maybe you do actually have to install sdl
00:10:31 <Quartus> And a chorus of subsequent errors, as is typical of C. 
00:10:33 <Quartus> Ok, where do I get it?
00:10:40 <JasonWoof> libsdl.org
00:11:06 <JasonWoof> oops
00:11:21 <Quartus> No .h in the win32 package.
00:11:58 <Quartus> Looking in the source snapshot.
00:12:07 <JasonWoof> one of my source files includes "SDL/SDL.h" and the rest "SDL.h"
00:12:20 <Quartus> There's no sdl.h in the whole tree of the two tars you had me download.
00:12:59 <JasonWoof> you might have to remove the SDL/ from the include paths at the top of sound.c
00:13:09 <JasonWoof> should be SDL.h
00:13:17 <JasonWoof> weird
00:13:33 <Quartus> Ok. But I have to unzip all these sdl*.h files to somewhere useful, hang on.
00:13:44 <JasonWoof> looks like it needs SDL_mixer.h SDL_image.h and SDL_keysym.h
00:14:13 <JasonWoof> the first two might be in seperate packages
00:14:21 <JasonWoof> keysym I assume comes with the main libsdl
00:15:02 <Quartus> rocks.c:1:23: SDL_image.h: No such file or directory
00:15:09 <Quartus> Any thoughts as to where that can be found?
00:16:34 <JasonWoof> seperate package from libsdl.org
00:16:39 <Quartus> think I found it.
00:16:45 <JasonWoof> you'll need to get the dev packages for sdl-image and sdl-mixer
00:17:58 <Quartus> ok.
00:18:01 <Quartus> Getting mixer now.
00:19:20 <Quartus> f:\mingw\bin\..\lib\gcc-lib\mingw32\3.2.3\..\..\..\..\mingw32\bin\ld.exe: cannot
00:19:20 <Quartus>  find -lSDLmain
00:19:20 <Quartus> mingw32-make: *** [vor] Error 1
00:19:43 <Quartus> SDLmain, I guess I need.
00:20:35 <Quartus> checking the mingw sdl stuff
00:21:17 <JasonWoof> on my box SDLmain.a came from the "libsdl" package
00:21:29 <Quartus> What directory should I put the libs in to be found?  I'll look at the makefile.
00:21:47 <JasonWoof> put them right in the source directory
00:21:50 <JasonWoof> with main.c and all that
00:21:56 <Quartus> me too, but no joy.
00:22:12 <JasonWoof> well, the dynamic libs should go there anyway
00:22:28 <JasonWoof> I want to package the dynamic libs with the exe so people don't have to install SDL to play the game
00:22:42 <JasonWoof> but as I understand it .a files get built into the exe
00:22:59 <Quartus> Do I need more than libSDLmain.a ? Coz it ain't linking.
00:23:47 <Quartus> should the makefile include the 'lib' prefix?
00:24:25 <JasonWoof> you can add paths for it to look for libraries with -L
00:24:41 <Quartus> yeah, but you have to add . to that path?
00:24:44 <JasonWoof> ie: -L. should make it look in the current dir
00:24:50 <JasonWoof> dinno
00:24:56 <JasonWoof> I don't know that much about the paths and such
00:25:03 <JasonWoof> linux tends to have that stuff already set up
00:25:19 <Quartus> ok, you actually do, which is astonishingly stupid.  Ok, now it wants the image.a.
00:26:34 <JasonWoof> ooh :) tonight I turn my ringer off :)
00:26:39 <JasonWoof> got woken by the phone twice this morning
00:26:57 <Quartus> there doesn't appear to be an approprate .a for mingw for sdl_image, at least not one I've found so far.
00:27:08 <JasonWoof> weird
00:28:37 <JasonWoof> true enough
00:28:45 <Quartus> so I'm kind of stuck without that.
00:29:24 <Quartus> undefined reference to `IMG_Load'
00:30:11 <Quartus> Likewise, they're hiding a mixer.a for mingw pretty damned well, if it exists at all.
00:30:41 <Quartus> Are those functions in the dll or something?
00:31:24 <JasonWoof> maybe
00:31:32 <Quartus> How should I proceed?
00:31:41 <JasonWoof> did you try removing -lmixer -limage from the makefile?
00:32:05 <Quartus> Yes, then I get a raft of  undefined references, mostly to mixer stuff, but also to IMG_Load which I imagine is in sdl_image.
00:33:07 <JasonWoof> ok
00:34:05 <Quartus> I made that change to sound.c, as well, btw.
00:37:17 <JasonWoof> this is weird
00:37:32 <Quartus> It compiles otherwise.
00:38:31 <Quartus> There are .lib files in the VC6 stuff, but I'm guessing ld won't recognize'em.
00:39:37 <Quartus> I could go through the source and remove all calls to the offending libraries, and see what I get. :)
00:40:21 <JasonWoof> hehe
00:40:27 <JasonWoof> no good if it can't load the graphics
00:40:52 <Quartus> Well, maybe there's .a files in some older version of the SDL stuff?  Did this ever build under mingw?
00:41:21 <JasonWoof> after googling a very little bit I've heard suggested to build SDL-image (and maybe sdl too) from source
00:41:32 <JasonWoof> or the last post here: http://www.gamedev.net/community/forums/topic.asp?topic_id=92297
00:41:39 <Quartus> and that's going to magically build the .a files?  jebus, that's an uphill climb.
00:41:39 <JasonWoof> says how to make .a files from the dlls
00:42:04 <JasonWoof> we have built it on windows before, I'm pretty sure it was with mingw
00:42:18 <JasonWoof> I mostly just watched as tathi figured it out though
00:42:40 <Quartus> That link the message gives for pexport isn't valid any more.
00:43:23 <Quartus> Found another copy.
00:45:05 <Quartus> Ok, now I need dlltoo.
00:45:14 <Quartus> dlltool.  This is a PAIN IN THE ASS Jason. :)
00:45:27 <JasonWoof> this explains how to install mingw and sdl: http://www.libsdl.org/extras/win32/mingw32/README.txt
00:45:35 <JasonWoof> sorry
00:47:05 <Quartus> Ok, it builds.  Popup error: "The procedure entry point SDL_strlcat could not be located in the dynaminc link library SDL.dll."
00:47:47 <JasonWoof> huh
00:48:15 <JasonWoof> not a function I've heard of 
00:48:24 <JasonWoof> it's not the vor sources
00:48:41 <JasonWoof> does it run anyway?
00:48:56 <Quartus> Stops dead.
00:49:24 <JasonWoof> did you install msys?
00:50:04 <Quartus> Got a new SDL.dll.  Game is running.
00:50:11 <Quartus> Oddly it started up once, and closed immediately. Second time, it runs.
00:50:31 <Quartus> And the keyboard controls are ...?
00:50:37 <JasonWoof> arrow keys
00:50:46 <Quartus> and that's it? Just get out of the way?
00:50:49 <JasonWoof> or nthc if you use dvorak
00:50:51 <JasonWoof> yep :)
00:51:31 <Quartus> Makes a nice boom noise.  Seems to work.
00:51:47 <JasonWoof> pass -m and see if the music works
00:51:51 <JasonWoof> ./vor -m
00:52:18 <Quartus> Yes, extremely homosexual techno music is now playing incessantly.
00:52:31 <Quartus> Wait, I'll get my leather pants and vest.
00:52:32 <JasonWoof> lol
00:52:43 <JasonWoof> now you see why it's disabled by default :-P
00:53:03 <Quartus> heh
00:53:27 <Quartus> Anyway, it all works, after one amazingly annoying process.
00:53:31 <JasonWoof> but still, if you could create a windoze shortcut that passes -m that would be cool
00:53:39 <Quartus> You can.
00:53:52 <JasonWoof> I can?
00:53:59 <Quartus> I mean the universal you. Anyone can.
00:54:06 <Quartus> I can't, I mean it'd be path-specific to my system.
00:54:16 <JasonWoof> yes, but most people aren't that sophisticated
00:54:27 <JasonWoof> oh
00:54:36 <JasonWoof> no relative shortcuts
00:54:37 <JasonWoof> booo
00:54:41 <JasonWoof> oh well
00:55:02 <Quartus> You need an installer.
00:55:14 <JasonWoof> what would I have to do to get something in that dir that people could double click to run ./vor -s?
00:55:19 <JasonWoof> make a batch file?
00:55:32 <Quartus> Oh, sure, that'd do it.
00:55:41 <Quartus> @./vor -s
00:56:00 <Quartus> What's -s?
00:56:03 * JasonWoof blinks
00:56:09 <JasonWoof> -s is no sound
00:56:12 <Quartus> Oh ok.
00:56:14 <JasonWoof> oh, sorry
00:56:23 <JasonWoof> well, that too actually
00:56:33 <Quartus> a batch file -- playvor.bat   -- @./vor -whatever
00:56:38 <JasonWoof> would be cool to have three: no sound, just explosions, with music
00:56:54 <JasonWoof> I'm just in the habbit of typing: ./vor -s
00:56:59 <Quartus> Is there sound other than the explosion? I'm not hearing it.
00:57:04 <JasonWoof> no
00:57:11 <Quartus> Anyway they can just click on the vor file, it'll play with no switches.
00:57:18 <JasonWoof> just right
00:57:47 <JasonWoof> so I can just do:  echo '@./vor -s' > vor-nosound.bat
00:58:23 <Quartus> .\    sorry
00:58:46 <JasonWoof> after we got a review in Linux Tab somebody contacted me asking if we wanted him to write new music for vor
00:59:13 <JasonWoof> I replied, blathering on a bit about small file sizes and making sure that he can give us compatible music
00:59:16 <Quartus> Heh.
00:59:17 <JasonWoof> then never heard from him again
00:59:25 <Quartus> Who was it?
00:59:26 <JasonWoof> oops, that's Linux Tag
00:59:48 <JasonWoof> I forget
01:00:07 <JasonWoof> was pretty exciting to find we had a review in a magazine
02:22:21 <Quartus> Slava, I've improved that quicksort a lot. Remind me to give you an update.
02:28:05 --- quit: slava ()
04:46:18 --- quit: Cheery ("Download Gaim: http://gaim.sourceforge.net/")
04:55:44 --- join: PoppaVic (n=pete@0-1pool65-71.nas22.chicago4.il.us.da.qwest.net) joined #forth
06:38:45 --- join: Ray_work (n=Raystm2@199.227.227.26) joined #forth
07:06:03 --- join: madwork (n=foo@derby.metrics.com) joined #forth
07:16:04 --- join: Cheery (n=Cheery@a81-197-19-23.elisa-laajakaista.fi) joined #forth
07:20:21 --- join: nighty (n=nighty@66-163-28-100.ip.tor.radiant.net) joined #forth
07:20:32 <k4jcw> Quartus,  you around?
07:23:58 <nighty> lo
07:25:47 <k4jcw> hey
07:51:08 --- join: nighty_ (n=nighty@66-163-28-100.ip.tor.radiant.net) joined #forth
07:51:26 --- quit: nighty (Read error: 113 (No route to host))
07:54:51 --- quit: k4jcw (Remote closed the connection)
07:56:59 --- join: k4jcw (n=jcw@adsl-065-006-151-062.sip.asm.bellsouth.net) joined #forth
08:03:50 --- quit: virl (Read error: 110 (Connection timed out))
08:13:00 <k4jcw> Opinion poll:  I've added support for socket, bind, listen, accept, etc.  Oddly, the words are called "socket", "bind", "listen", "accept", etc.  Current I do not support dictionaries.  Should I rename these words to "net-socket", "net-bind", etc?
08:14:15 --- join: Ray-work (n=Raystm2@199.227.227.26) joined #forth
08:28:56 --- quit: Ray_work (Read error: 110 (Connection timed out))
08:35:00 --- join: Ray_work (n=Raystm2@199.227.227.26) joined #forth
08:42:11 <TreyB> Leave the names and add wordlist support :-)
08:44:46 <k4jcw> Yea.  I really should.
08:46:34 --- quit: PoppaVic ("Pulls the pin...")
08:50:18 --- quit: Ray-work (Read error: 110 (Connection timed out))
08:57:41 --- nick: Raystm2- -> nansmt
09:27:35 <JasonWoof> k4jcw: cool stuff
09:27:39 <JasonWoof> k4jcw: what forth?
09:28:26 --- quit: madwork ("?OUT OF DATA  ERROR")
09:34:20 --- join: madwork (n=foo@derby.metrics.com) joined #forth
09:37:46 <k4jcw> JasonWoof,  my own.  It's a pseudo-forth.
09:38:16 <k4jcw> The stack supports ints, floats, strings, and what I call mptrs.  The words all have type checking for their arguments.
09:38:52 <k4jcw> Gotta run out, I'll tell you more about it later.
09:51:09 --- join: virl (n=virl@chello062178085149.1.12.vie.surfer.at) joined #forth
09:53:04 --- join: Quartus_ (n=Quartus_@209.167.5.1) joined #forth
10:10:40 --- join: Ray-work (n=Raystm2@199.227.227.26) joined #forth
10:10:47 <Quiznos> rays
10:15:23 --- join: tgunr (n=davec@70-41-218-134.cust.wildblue.net) joined #forth
10:28:03 --- quit: Ray_work (Read error: 110 (Connection timed out))
10:40:20 <JasonWoof> *(@#&$^(!@*#^&$(*@#^&$(*@#&^$(@*#&^$
10:40:30 <JasonWoof> I can't get the gimp to let me move pixels around
10:40:42 <JasonWoof> this has got to be the most annoying thing in the universe
10:40:48 <JasonWoof> I select a rectangle
10:41:02 <JasonWoof> then try to drag it, and instead it makes a new rectangle
10:41:34 <JasonWoof> I try to move it with the move tool, and I can get it to either move the selection outline (but not the contents) or move the entire layer
10:42:45 <k4jcw> gimp ui sucks big time
10:52:21 <virl> ok..
10:52:57 <virl> it's interesting everybody seems to say that gimps ui sucks, except me. I'm quiet fine with it.
10:53:33 <JasonWoof> for the most part I've been pretty happy with the gimp
10:54:13 <virl> yeah :-)
10:54:21 <virl> another gimp loving soul
10:54:35 <JasonWoof> dunno why it was giving me trouble today
10:55:31 <virl> hmm? so a stupid behaviour ala squeak?
11:29:17 --- join: snoopy_1711 (i=snoopy_1@dslb-084-058-155-238.pools.arcor-ip.net) joined #forth
11:29:48 --- quit: Snoopy42 (Nick collision from services.)
11:29:57 --- nick: snoopy_1711 -> Snoopy42
12:26:44 --- quit: Cheery (Remote closed the connection)
12:28:40 --- join: EsotericMoniker (n=Esoteric@207.226.175.73) joined #forth
13:23:38 --- quit: Quartus_ (Read error: 104 (Connection reset by peer))
13:36:28 --- part: saon left #forth
13:46:26 --- nick: docl -> lukeparrish
14:32:36 --- join: vatic (n=chatzill@pool-162-84-156-148.ny5030.east.verizon.net) joined #forth
15:08:36 --- quit: Ray-work ("User pushed the X - because it's Xtra, baby")
15:56:34 --- nick: nansmt -> Raystm2
17:03:29 <virl> gnah, when I would know how a correct create_window request looks like for the xprotocol
17:07:31 <Quiznos> when? besides that statement doesnt parse properly in engrish? :)
17:11:17 --- quit: virl (Remote closed the connection)
17:27:06 --- quit: vatic ("*poof*")
17:36:10 --- nick: Quiznos -> GoogleMaster
17:39:50 --- nick: GoogleMaster -> Quiznos
17:52:40 --- join: nighty (n=nighty@CPE00119576a9c5-CM0012c90d36fc.cpe.net.cable.rogers.com) joined #forth
18:23:27 --- nick: Quiznos -> ElmerJFudd
18:49:32 --- nick: ElmerJFudd -> Quiznos
19:27:08 --- quit: nighty (Read error: 104 (Connection reset by peer))
19:29:59 --- quit: EsotericMoniker ("Trillian (http://www.ceruleanstudios.com")
20:19:18 <Quartus> Hi k4jcw.
20:19:58 <Quartus> Just saw your note.
20:21:13 <k4jcw> ah
20:21:39 <k4jcw> I was going to ask about a typical implementation of case/of/endof/encase, I think.
20:21:52 <Quartus> There's on in the Standard doc that works nicely.  
20:22:22 <k4jcw> Do you happen to have a link to that?  Or at least the root?
20:23:19 <Quartus> http://quartus.net/dpans
20:24:46 <Quartus> http://www.quartus.net/files/PalmOS/Forth125/Library125/case.txt
20:25:44 <k4jcw> So basically it's building a list of fixup addresses that endcase patches?
20:26:39 <Quartus> Sure, more abstract to say it compiles a sequences of IFs that are resolved by multiple THENs via endcase.
20:27:10 <k4jcw> Yea, ok.
20:27:27 <k4jcw> I got all my network words tested today.  I have a stupid little echo app that works quite well.
20:27:32 <Quartus> Good deal.
20:28:33 <Quartus> CASE/OF/ENDOF/ENDCASE is not my favourite construct.  It's sometimes necessary, but it should be avoided.  Similar in that way to PICK and ROLL.
20:29:11 <k4jcw> OK, so the top of the text file says "case 97.8.27 (from dpANS docs)", but I don't see a section 97 in the index.
20:29:24 <Quartus> That's the date I wrote the file. :)
20:29:37 <k4jcw> argh
20:29:49 <Quartus> It's in A3.2.3
20:30:29 <k4jcw> A.3.2.3 Stacks
20:30:36 <Quartus> A3.2.3.2 actually, scroll down a bit.
20:30:40 <k4jcw> oops.
20:30:43 <k4jcw> I didn't mean to paste that.
20:31:17 <k4jcw> Why don't you care for it?  And what would you suggest instead?
20:31:33 <Quartus> http://quartus.net/dpans/dpansa3.htm#A.3.2.3.2      scroll down a bit
20:32:02 <k4jcw> Yah, I found it.
20:32:09 <Quartus> CASE is syntactic sugar for a bunch of nested IF/ELSE/THEN conditionals, which signal bad Forth the way the use of PICK, ROLL, a multiplicity of variables, or locals does.
20:32:47 <Quartus> Lookup tables are a far better solution, generally.
20:32:54 <k4jcw> Well, I would agree, except in certain cases.  If you use system calls, many return -1, 0 and >0.  -1 means error, 0 means file closed, and >0 is length.
20:33:16 <k4jcw> And I much prefer the readability of case over if blocks.
20:33:23 <Quartus> Try it with neither one.
20:33:49 <k4jcw> Not sure what you mean.
20:33:57 <Quartus> What function are you talking about that returns those three possibilities?
20:34:28 <k4jcw> Off hand, specifically read()
20:34:54 <k4jcw> Also, select().
20:35:07 <Quartus> Ok. I'm out of touch, I'll have to check the args for read().  I assume you want to wrap it to make it more Forth-like, and then use that wrap.
20:35:09 <k4jcw> -1 means error, 0 means timeout, >0 indicates files set.
20:35:15 <k4jcw> Yes.
20:35:52 <Quartus> You'd want to THROW the error.       So       ...      ?dup -1 = -whatever throw    takes care of that cas.
20:35:54 <Quartus> case.
20:37:46 <Quartus> So you're down to dealing with the 0 (read() read end-of-file)  or >0 (bytes read).
20:38:04 <Quartus> Which I'd just return.
20:38:24 <k4jcw> What would be nice (IMHO) is a ?= word that does not duplicate the TOS-1 value if they match.
20:38:32 <k4jcw> -1 ?= -xx throw
20:38:41 <Quartus> If you make errno visible in your Forth, it could as easily be         ?dup -1 = errno and throw
20:38:55 <k4jcw> Yes.  errno is already visible.
20:39:00 <Quartus> I left out an 'and' up there.  That's the idiom.  Make the comparison, and the value, do the throw.
20:39:27 <k4jcw> I need to fix my value for TRUE, I think.  It's 1, instead of all bits set.
20:39:35 <Quartus> Yes, you need all bit sets.
20:39:41 <Quartus> er, all bits set.
20:40:31 <Quartus> : errno-throw  = errno and throw ;
20:40:35 <k4jcw> I think I am going to be ready to implement word lists here shortly.  I may (read: will) have more questions about that.
20:40:45 <Quartus> Then       ?dup -1 errno-throw
20:41:10 <Quartus> Ok.
20:41:13 <k4jcw> Why use ?dup instead of dup?
20:41:33 <JasonWoof> question before means maybe do that
20:41:33 <Quartus> Oh, I don't know, just an example.
20:41:45 <JasonWoof> question after means "should I do x?"
20:41:50 <k4jcw> Yes, I know that.
20:42:04 <k4jcw> But for a test for -1, it has no advantage that I saw.
20:42:16 <JasonWoof> oh, I was assuming different quoting: Why use "?dup" instead of "dup?"
20:42:18 <k4jcw> I thought perhaps he had some arcane reason.
20:42:26 <k4jcw> Stupid punctionation!
20:42:26 <Quartus> It would discard a zero result.  You get the idea, though.
20:42:34 <k4jcw> Yes
20:43:02 <Quartus> I spent some time writing a quicksort implementation over the last couple of days.
20:43:47 <k4jcw> Fixed record size with key fields, or arbitrary record sizes?
20:44:36 <Quartus> Sorts arrays of cells, you can use them as integers, or as pointers and set the comparison word to anything you like.
20:45:10 <k4jcw> cool
20:45:11 <Quartus> foo 5 ' < sort       sorts an array of cells, descending.
20:45:28 <Quartus> foo 5 ' scompare sort       sorts an array of pointers to counted strings.
20:46:17 <Quartus> etc.
20:47:07 <Quartus> I found one that Ertl is credited as having written, and it flat out didn't work.  I hope it was a transcription problem.
20:47:09 <k4jcw> Id'a not thunk of passing the xt to the sort routine.  I'd have set up a deferred word that the user revectored to his compare routine.  Your way is better.
20:47:29 <Quartus> I do set up a deferred word, but it's sort's business what that word is named.
20:51:05 <Quartus> Information-hiding, and all that. :)
20:51:31 --- join: futhin (n=wunderwa@bespin.org) joined #forth
20:51:31 --- mode: ChanServ set +o futhin
20:51:41 --- part: futhin left #forth
20:52:55 <Quartus> It makes a much more informative benchmark than the simple fib routines I was using before.
20:56:53 <k4jcw> I wonder if my pseudo-forth can run your sort.
20:57:19 <Quartus> I don't know, your typechecking stuff might get in the way.
20:57:39 <Quartus> Do you have RECURSE ?
20:57:45 <k4jcw> Yes.
20:58:43 <Quartus> Can you @ a cell, pull another cell from the return stack with R@ two to an arbitrarily revectored deferred word that compares the two? 
20:58:56 <Quartus> sorry, "and pass the two to an"
20:59:11 <Quartus> over-deleted, there. :)
20:59:16 <k4jcw> In theory.
20:59:22 <Quartus> Might work, then.
20:59:53 <Quartus> Be powerful slow in your system, though.  Better to call qsort() since you've got the library right htere.
21:00:39 <k4jcw> It would be interesting to run as a benchmark to see if the speed ratio remains constant compared to fib.
21:02:03 <Quartus> It'll likely be considerably worse.
21:02:22 <Quartus> fib mostly tests the speed of a call.
21:02:56 <k4jcw> For whatever interest it may be, I compiled mine under Cygwin and my test app continued to work.  I had only one issue, which was that cygwin does not have cfmakeraw in it's library.
21:03:06 <Quartus> What is that function?
21:03:25 <k4jcw> It's for setting all the right bits to make the console uncooked, non-block, etc.  
21:03:27 <Quartus> Ah, a termios thing.
21:03:30 <k4jcw> Yes.
21:05:47 <k4jcw> Is postpone anything more than "' <word> ," ?
21:05:56 <k4jcw> That's tick word comma
21:06:12 <Quartus> Yes, it's more than that.  It compiles the compilation-semantics of a word.
21:06:28 <k4jcw> I'm not sure what the means.
21:06:34 <Quartus> So  : foo  postpone dup ; immediate    makes foo a word that will compile DUP.       : bar  5 foo . . ;   bar -> 5 5 
21:07:38 <Quartus> If postponing a word that executes at compile-time, like IF or ( or something, it later executes it instead of laying down code.
21:07:52 <Quartus> So POSTPONE xxx  does whatever xxx would do if it were encountered at compile-time.
21:08:08 <Quartus> Only it does it later. The action is POSTPONEd.
21:10:40 <Quartus> so, in pseudo-code,     : postpone   bl parse (find) immediate? not if  literal,  ['] compile, then compile, ;
21:11:03 <Quartus> That pseudocode assumes a pseudoword pseudonamed (find) that returns an xt and the immediate status.
21:11:20 <Quartus> And a literal, that lays down code for a literal.
21:11:53 <Quartus> oh, and postpone itself would be immediate.
21:12:05 <k4jcw> I think I can implement that.
21:12:17 <Quartus> Here's the full pseudothing:       : postpone   bl parse (find) immediate? not if  literal,  ['] compile, then compile, ; immediate
21:13:32 <Quartus> Maybe more pseudocorrect:     : postpone   parse-name (find) immediate = not if  literal,  ['] compile, then compile, ; immediate
21:13:33 <k4jcw> However, I can't do it in Forth (I need name for my pseudo-Forth), I'd have to do it in C.  Right now, the tokenizer is not exposed.
21:13:58 <Quartus> er, scratch that last one, can't have the immediate constant named 'immediate', that's taken. :)
21:14:02 <Quartus> I'll pseudostop now.
21:14:05 <k4jcw> k
21:14:37 <k4jcw> What does literal, do?  Laydown (literal) ?
21:14:39 <Quartus> Oh, any number of Forth programs depend on having access to the input buffer, and expect certain behaviour from PARSE WORD and friends.
21:15:17 <k4jcw> Yes, and they won't run.  I don't have TIB, and don't really play to support it.
21:15:43 <Quartus> literal, is a normal word (non-immediate) that lays down code to push the provided literal value onto the stack, however that works in your system.
21:16:35 <Quartus> TIB isn't needed, but >IN and SOURCE are handy, as is REFILL, WORD and PARSE.  parse-name is another useful one.  Once built you can have your interpret loop use'em.
21:18:03 <Quartus> How do you include source without a way of saving and restoring the current input specification?  No way to nest one include inside another?
21:19:18 <k4jcw> It can nest includes.
21:19:35 <k4jcw> I have "fload", and it ends up being recursive.  Whatever state it needs is saved.
21:20:11 <Quartus> Another annoying arbitrarily-renamed word for a traditional function. :)
21:20:31 <k4jcw> In fact, it's somewhat intelligent.  If you're loading a nested file and it gets an error, you get the load stack, so you can see what included what.
21:20:42 <k4jcw> fload came from FPC, I think.
21:21:44 --- quit: tgunr (Read error: 104 (Connection reset by peer))
21:21:45 <Quartus> Yes, displaying the nested include list is the typical 'include' setup based on included, with save-input, restore-input, and catch/throw.
21:21:52 <k4jcw> I also like my error stack.  For instance (short paste):
21:21:55 <k4jcw> bind -- 98/Address already in use
21:21:55 <k4jcw> --> 0x8064710 (abort)
21:21:55 <k4jcw> --> 0x8067020 (prettyAbort)
21:21:55 <k4jcw> --> 0x8067088 (errchk)
21:21:55 <k4jcw> --> 0x8067130 (bind-socket)
21:21:55 <k4jcw> --> 0x8067378 (listen-test)
21:21:57 <k4jcw> --> 0x80673a8 (go)
21:22:01 <k4jcw> (0,0) > go
21:22:14 <k4jcw> Well, the (0,0) > go belongs at the top.
21:22:22 <Quartus> Sure. That's the usual sort of backtrace.
21:22:35 <k4jcw> gforths is hideous
21:22:57 <Quartus> What's hideous about it?  It lists addresses and names.
21:23:36 <Quartus> Quite similar to yours, actually, with $ instead of 0x.
21:24:39 <k4jcw> Oh, I remember.  It's that for any mistyped word you get one.  That's what I find annoying about it.
21:25:05 <Quartus> For any throw, but it's only a couple of lines for a mistyped word.
21:25:13 <Quartus> Backtrace:
21:25:13 <Quartus> $E850B4 throw
21:25:13 <Quartus> $E90B10 no.extensions
21:26:59 <k4jcw> Too verbose.  
21:27:36 <k4jcw> I'm off to bed.  Tomorrow (later today?) I shall implement implement case/of/endof/endcase, and try to implement postpone.
21:27:40 <Quartus> Ok.
21:27:49 <k4jcw> Not sure if I completely understand postpone well enough to do it right.  We shall see.
21:27:58 <k4jcw> Thanks for you help.  It's always educational.
21:28:05 <Quartus> Ok. There are coretests for postpone.
21:28:20 <k4jcw> ah, cool.  I should get a pointer to those, and try some of them.
21:28:33 <Quartus> Good idea. :)
21:29:16 <k4jcw> Are they in the gforth source distro?
21:29:22 <Quartus> Think so, yes.
21:29:25 <k4jcw> k
22:31:47 --- join: Cheery (n=Cheery@a81-197-19-23.elisa-laajakaista.fi) joined #forth
22:58:43 <Quartus> Update. Now with a backtrace facility.  http://quartus.net/retro/retro-ans.fs
23:59:59 --- log: ended forth/06.08.28