00:00:00 --- log: started retro/06.08.25
05:51:57 --- join: timlarson_ (n=timlarso@65.116.199.19) joined #retro
06:26:39 --- join: Ray_work (n=Raystm2@199.227.227.26) joined #retro
06:40:03 --- nick: Raystm2 -> nanstm
10:55:35 <Quartus> crc, I've built the retrospect xt>length code into the disassembler, works quite well.
11:06:53 <Quartus> crc, question: why does retro's then write a no-op operation right afterward?
11:09:42 <docl> I think that might have been for the else feature
11:10:43 <Quartus> It doesn't appear to be needed even for else.
11:15:01 <Quartus> In fact definitely not, I've replaced then and all my tests run cleanly.
11:16:56 <Quartus> Now the ANS fib bench is more than twice the speed of the retro version.
11:17:55 <docl> heh, cool :)
11:20:09 <virl> -__-
11:20:31 <virl> hail ans, how perfect it is.
11:20:44 <docl> lol
11:21:20 <docl> don't get violent guys
11:21:21 <virl> everybody who is using retro should now stop to use it, because ans showed that it's faster! *sarcasm*
11:21:44 <docl> virl, ans and retro are focusing on two different things. it's apples and oranges.
11:25:05 * virl is only not in a good mood
11:25:30 <docl> sorry to hear that :(
11:27:35 --- join: snoopy_1711 (i=snoopy_1@dslb-084-058-096-142.pools.arcor-ip.net) joined #retro
11:27:37 <virl> nothing happens in retro, only fucking ans. I get sick.
11:27:58 <Quartus> Updated: http://quartus.net/retro/retro-ans.fs
11:28:04 <docl> well that's just lately virl
11:28:25 <virl> docl, ?
11:28:39 <docl> oh you meant in general
11:29:16 <docl> actually I think crc's been getting a lot done that I just haven't been keeping up on very well
11:29:32 <Quartus> I saw his notes for the new rx, he's working away.
11:30:55 <docl> yeah I have them too. need to take a closer look
11:34:43 <Quartus> what's the best x86 for testing if a register holds 0?
11:35:54 --- quit: Snoopy42 (Read error: 145 (Connection timed out))
11:36:19 --- nick: snoopy_1711 -> Snoopy42
11:37:01 <virl> so.. to bring a bit fresh air and clean this nice channel of the dust of ANS, what could be good ideas for retro as a webdev language?
11:40:25 * docl likes the RDML idea. perhaps it can be adapted to a good dynamic html generator
11:40:56 <virl> hmm..
11:42:38 <virl> then you have a document, but what's with the server scripts? do you mean it like some perl or php scripts, which produce the website?
11:43:31 <docl> yeah. it could feed to standard output instead of a file. then it would be the same as a perl script that produces html.
11:44:01 <virl> hmm.. 
11:46:31 --- join: neceve (n=claudiu@unaffiliated/neceve) joined #retro
11:56:06 <Quartus> Another update, hot on the heels of the last one: http://quartus.net/retro/retro-ans.fs
11:56:12 <Quartus> Just realized IF could also do with some optimization.
11:56:36 <Quartus> Very handy to have a SEE to browse around inside retro with.
12:07:32 --- quit: neceve ("Leaving")
12:11:59 --- join: neceve (n=claudiu@unaffiliated/neceve) joined #retro
12:31:50 --- quit: Cheery ("Download Gaim: http://gaim.sourceforge.net/")
13:31:28 --- quit: timlarson_ ("Leaving")
14:07:34 --- join: Quartus__ (n=Quartus_@209.167.5.1) joined #retro
14:09:48 --- quit: Ray_work ("User pushed the X - because it's Xtra, baby")
14:30:39 --- quit: Quartus__ ("used jmIrc")
14:31:42 --- join: Quartus__ (n=Quartus_@209.167.5.1) joined #retro
15:01:18 --- quit: neceve ("Leaving")
15:13:53 <crc> good evening
15:14:43 <crc> virl: actually a lot is happening with retro, just nothing public yet
15:16:07 <crc> I have been working on the core since 9.2.0 was released; I expect to start integrating the revamped core and retro together sometime next week
15:25:58 <Quartus__> hi crc.
16:01:19 --- quit: Quartus__ ("used jmIrc")
16:15:51 <virl> hi crc, I'm happy to read you :-)
17:38:38 --- join: jas2o (n=jas2o@WNPP-p-203-54-32-52.prem.tmns.net.au) joined #retro
18:01:20 --- quit: jas2o ("jas2o.forthworks.com")
18:10:35 --- quit: virl (Remote closed the connection)
18:37:30 --- quit: Shain (Remote closed the connection)
19:18:24 --- join: Shain (i=steve@c-67-161-56-76.hsd1.ca.comcast.net) joined #retro
20:11:38 <Quartus> crc, if I were to want to get at the read-line you've got in retroforth, use it in an app, is that a possible thing?
20:13:47 <Quartus> The console input routine, that is.
20:39:07 <crc> not directly
20:39:32 <Quartus> Any indirect way?  I've been hunting through your sources unsuccessfully to find where it's called.
20:39:56 <crc> look in modules/interpret
20:40:31 <Quartus> I have 9.2.7 sources here, interpret is really, really short.
20:40:46 <crc> it is a short bit of code
20:41:05 <Quartus> But the console input lets you edit, scrollback, all that good stuff.
20:41:21 <crc> are you using windows?
20:41:23 <Quartus> Yes.
20:41:42 <crc> that is all provided transparently by the Win32 console subsystem
20:41:51 <Quartus> Through what api conduit?
20:42:31 <Quartus> Sorry, bit new to this process.  I'm wanting to build ACCEPT for the layer.
20:42:37 <crc> ReadFile, using the stdin 
20:42:51 <crc> basically the definition of 'key' in windows/windows.f
20:42:55 <Quartus> Ah, ok.
20:43:03 <Quartus> Thanks.
20:43:08 <crc> np
20:43:26 <Quartus> I take it the non-Windows versions don't have as sophisticated an input.
20:44:00 <crc> nope
20:44:15 <crc> I use rlwrap for that on Linux/BSD
20:45:16 <Quartus> That's a neat tool.
20:45:29 <crc> yeah
20:45:41 <crc> it's a very handy thing to keep on hand :)
20:46:12 <Quartus> I'll say.  I remember building readline into a Linux Forth I wrote many years ago.  This would save the trouble. :)
22:06:23 <Quartus> crc: :: d: key dup 13 = if d: key drop then ; is key
22:06:23 <Quartus> :: d: how ."  ok" cr ; is how
22:06:59 <Quartus> That discards the 10 after the 13, in Windows. It would likely have no effect in Linux, unless somebody wanted to pass ctrl-M via key.
22:09:07 <crc> cool
22:09:25 <crc> I'll integrate this into 9.3 (or 9.2.10, if I do a 9.2.10)
22:10:18 <Quartus> Just writing an 'accept' now.
22:12:03 <Quartus> Love for the carriage-return not to happen after input.
22:12:11 <Quartus> But this is progress, anyway.
22:55:39 <Quartus> crc, if you expose source and tp I can write a considerably more-robust EVALUATE.  Still not perfect, but an improvement.  Right now I'm getting at those two variables with  whitespace 5 cells +  and  whitespace 6 cells +   which is far from ideal :)
23:21:34 <Quartus> Updated, now with ACCEPT and an 'ok' prompt, and a cure for LF after CR in the Windows version: http://quartus.net/retro/retro-ans.fs
23:27:14 <crc> I'll expose them in 9.3
23:27:38 <Quartus> Ok. I suggest a whole internal save-input and restore-input setup as a more permanent solution.
23:32:22 <Quartus> That way external routines can call save-input and restore-input, and any given rx release can make sure it's storing & restoring whatever is necessary.
23:34:33 <Quartus> Though there's a minor issue with eval/EVALUATE, wherein you want to save the input specification on the return stack so it's out of the way, but that's manageable. 
23:35:04 <Quartus> The version I just wrote is : evaluate  >in @ >r  whitespace 5 cells + @ >r  whitespace 6 cells + @ >r  ['] eval catch  r> whitespace 6 cells + !  r> whitespace 5 cells + !  r> >in !  throw ;
23:35:34 <Quartus> Obviously not for inclusion in the official ANS layer just yet. :)
23:50:40 <Quartus> crc, weird stuff.  That key revectoring disrupts the stack underflow checking I built into 'how'.
23:51:29 <Quartus> Better leave that out of the 9.3 (or .10 or whatever) until I figure out why.
23:51:36 <crc> ok
23:59:37 <Quartus> Wish that was clearer to me.  It seems you do special things when a line parses out to 0 length.
23:59:59 --- log: ended retro/06.08.25