00:00:00 --- log: started retro/11.04.16
03:35:35 --- quit: erider (Ping timeout: 246 seconds)
04:01:32 --- join: roarde (~roarde@pdpc/supporter/active/roarde) joined #retro
04:13:34 --- quit: roarde (Quit: Leaving)
04:54:40 --- join: erider (~chatzilla@pool-108-3-149-46.bltmmd.fios.verizon.net) joined #retro
04:54:51 --- quit: erider (Changing host)
04:54:51 --- join: erider (~chatzilla@unaffiliated/erider) joined #retro
04:54:56 <erider> morning 
05:23:45 <crc> good morning
06:15:56 <erider> hey crc shouldn't this increment the value not the pointer *s++;
06:16:10 <erider> crc I was looking at you example 
06:16:26 <erider> *s = c;
06:16:28 <erider>     *s++;
06:16:30 <erider>     l++;
06:16:46 <erider> shouldn't it be s++ 
06:22:42 <erider> crc: while (isalpha(c = getc(in))) {
06:22:43 <erider>         temp[i++] = c;
06:22:45 <erider>     }
06:23:14 <erider> that should build a string right 
06:39:45 <erider> hey crc: http://pastebin.com/FQe5TZm6 for some reason this goes into a forever loop 
06:51:36 <crc> a few issues:
06:51:36 <crc> 1) a pointer to nowhere (char *temp;)
06:51:36 <crc> 2) since the pointer is incremented, the printf() will start from the current value, not the original one
06:55:42 <erider> crc so I need to set it to NULL 
06:58:46 <crc> no, you need to point it to an allocated memory area
06:59:30 <crc> e.g., http://sprunge.us/BfDQ
07:00:49 <erider> crc: I have my numbers just like that and I get no issues 
07:01:47 <erider> while (isdigit(c = getc(in))) {
07:01:49 <erider>             temp[i++] = c;
07:01:50 <erider>         }
07:01:52 <erider> 		num = atol(temp);
07:02:09 <crc> that's not incrementing the pointer though
07:02:47 <crc> and how is temp declared for that function?
07:02:56 <erider> char *temp; 
07:03:25 <crc> so still problematic in that it's not specifically pointing anywhere
07:03:54 <erider> ok 
07:04:00 * erider is fixing it now
07:04:13 <crc> but here you are indirectly accessing it (via an index variable) rather than incrementing, so the original address is not lost
07:06:00 <erider> I tried that with the chars with no luck 
07:09:18 <crc> http://sprunge.us/CTbg
07:13:47 <erider> I think I fixed it 
07:13:54 <erider> crc: question 
07:14:01 <crc> yes?
07:14:27 <erider> doesn't *temp++ increment the value not the pointer 
07:15:03 <erider> why can't I do temp++ 
07:15:31 <erider> or temp[i++] 
07:22:35 <erider> crc: I wonder if I could have used isalpha like in my original 
07:23:09 <crc> isalpha matches any alphabetic char
07:23:17 <crc> temp[i++] will work
07:31:36 <erider> I think I have something that will work 
07:36:36 <erider> crc: http://pastebin.com/L2XgH166
07:43:57 <erider> crc: alright on to strings 
07:44:39 * erider has fixnum, Boolean and char literals
08:07:42 <erider> hmm not quite finish with the literals :( 
08:45:06 <erider> ok good now 
09:05:02 <crc> I wonder how hard it'd be to write a simple lisp-like parser in retro...
09:05:26 <crc> I have some code already for dealing with cons cells
09:58:51 <erider> crc: unknown :) 
10:23:05 --- join: aisa (~aisa@c-68-35-164-105.hsd1.nm.comcast.net) joined #retro
10:33:51 <erider> crc: is it always a good ideal to set a pointer to null 
13:52:44 <crc> why would you want it set to null?
13:55:12 <erider> crc: just wondering is good practice to set it to something, because of junk 
14:19:19 <crc> char *ptr;  <---- this is fine, just point it somewhere valid before use
14:56:19 --- quit: aisa (Quit: aisa)
16:02:48 <erider> crc: ok, hey question 
16:03:42 <crc> yes?
16:03:52 <erider> crc: is it better to parse char by char of just suck the whole thing into a buffer and parse that
16:04:17 <erider> the reason way I asks is because I just found a bug 
16:05:02 <crc> it depends on the application
16:05:03 <erider> my fixnum routine is taking one more char than it should 
16:06:00 <crc> I generally read by tokens (e.g., with retro), breaking when a particular character is encountered. For some things though (such as rancid), I read a line, then parse it into various parts
16:06:05 <erider> so with I have this (2 . 2) it keeps 2) but when I do (2 . 2 ) its good 
16:06:18 <erider> I am working on pairs :) 
16:06:27 <crc> pair = cons?
16:06:42 <erider> yeah car, cdr and con 
16:07:29 <crc> can you pastebin the entire code?
16:08:04 <erider> crc: sure it doesn't do much be echo right now 
16:08:22 <crc> easier to debug if I have the full code :)
16:15:53 <erider> http://pastebin.com/VcRXgdwz
16:18:25 <erider> I need to look at how I am handling fixnums 
16:22:58 <erider> crc: I am wondering if atol is dong something weird 
16:25:32 <erider> crc: wonder if this would fix it? (isdigit(c = getc(in)) && (!is_delimiter(c)))
16:28:19 <crc> nope
16:28:49 <erider> no? 
16:29:21 <erider> the num buffer must be getting an extra char 
16:30:20 <crc> doesn't appear so
16:33:02 <crc> read() is consuming the )
16:35:38 <erider> really!?
16:36:14 <crc> I think so, trying to confirm now
16:36:56 <erider> I am going to try an ungetc 
16:38:32 <crc> http://sprunge.us/QTRN
16:38:45 <crc> a patch that lets (2 . 2) parse properly
16:40:20 <erider> yeah the ungetc worked for one thing 
16:40:48 <erider> crc: hey that is the problem with reading one char at a time 
16:42:11 <crc> reading lines into a buffer has its own issues. what if you fill the buffer before the input is completed? token by token is easier, but not always possible.
16:45:26 <erider> crc: well I guess I will keep moving with token by token method :) 
16:49:42 <erider> crc: you said you never used ungetc? then how did you put token back on the stream?
16:50:01 <crc> I don't use ungetc in my code
16:52:05 <erider> crc: how do you rewind 
16:52:20 <crc> I don't
16:52:49 <crc> retro's standard parser can't rewind the input source
16:56:11 <erider> oh
16:58:07 <crc> even if I wanted to, the vm does not expose the input stack directly, so there's no way to rewind
16:58:29 <erider> crc: I have learned with using ungetc you have to follow the parser char by char to follow the flow so you can add the ungetc where needed 
16:59:22 * erider wonders if you should try to implement it without ungetc
18:56:17 <erider> crc: any movement on the logger today?
19:01:24 --- join: edrx (~Eduardo@186.205.236.21) joined #retro
19:01:38 <crc> no code changes
19:02:07 <crc> hi edrx
19:02:14 <edrx> hi crc and cfa
19:02:42 <edrx> I am trying to write the lua functions that call libretro
19:03:21 <edrx> I connected to ask a few questions on #lua, I was not expecting to meet you so soon 8-)
19:04:08 <crc> I normally lurk here in the evenings
19:07:32 <edrx> I have been living mostly offline
19:18:21 <erider> hmm libretro 
20:06:30 * crc is having connection problems (wimax doesn't do well during thunderstorms with heavy rain)
20:07:10 <crc> erider: the C implementation of the vm can be built as a library now, and (at least potentially) be used/embedded into other applications
20:08:02 <erider> that is interesting
20:09:14 <crc> it'll be more interesting in a few weeks, when I have time to add functions to allow finer control over running retro code, and extending the I/O ports at runtime
20:10:22 <erider> I wonder how he is going to use it in lua 
20:11:07 <edrx> crc: it works!!!!!! 8-)
20:11:27 <edrx> lemme clean up the makefile and build a tarball
20:11:28 <crc> yay!
20:13:01 <edrx> it would be great to have another way to call retro from lua, though
20:13:39 <crc> edrx: I'm looking into ways to allow more control over retro from the library functions
20:13:44 <edrx> I'd like to pass a zero-terminated string to retro and make it execute it as a line of input, then return
20:13:57 <edrx> ok
20:16:04 <crc> I've dabbled with controling retro from other languages (javascript) in the past, so should be able to draw at least a little from my old experiments
20:16:31 <edrx> great
20:29:55 <edrx> crc: is this supposed to work?
20:30:27 <edrx> echo '2 3 * putn cr bye' | ./test
20:31:33 <edrx> sorry, it does work. I wrote "test" instead of "./test" in the makefile...
20:31:56 <crc> ok
20:32:19 <crc> that should work, unless something is seriously screwed up in the vm
20:37:49 <edrx> crc: the makefile is ready, and all tests work. let me add targets for downloading and building lua inside the luaretro directory, and I'll pack everything and send it to you.
20:38:18 <crc> ok, sounds good
21:01:54 <edrx> crc: what os are you using? the makefile should run on any linux without changes, I guess
21:02:14 <crc> I use linux and os x
21:02:45 <edrx> ok, making the tarball 8-)
21:22:29 <edrx> crc: http://angg.twu.net/RETRO/luaretro.tgz
21:22:37 <edrx> I am not a C programmer
21:22:43 <edrx> *at all*
21:23:25 <edrx> so expect all kinds of ugly bad habits (in luaretro.c) 
21:24:08 <edrx> http://angg.twu.net/RETRO/Makefile.html#help
21:24:29 <edrx> "make runluatest" should work right away on a linux box
21:25:57 <edrx> "make LUAOS=osx runluatest" should download and build lua51 correctly on osx, but I don't expect the rest to work
21:28:53 <crc> on linux, I had to add a LUA51SRC to the makefile to get it to build. works ok otherwise.
21:29:38 <edrx> oops! fixing it...
21:30:19 <edrx> when you make retro callable from javascript
21:30:36 <edrx> how did you execute retro code from js?
21:30:49 <edrx> s/make/made/
21:31:42 <crc> two ways: a string evaluation function that mapped a string as input, and a set of functions for getting a pointer to a function (by name) and invoking the pointer
21:32:50 --- quit: erider (Ping timeout: 246 seconds)
21:33:07 <edrx> would it be easy to add a string evaluation function to libretro.c?
21:34:02 <crc> for osx, "make LUAOS=macosx runluatest" works after adding the LUA51SRC and -llua to the rule for retro.so
21:34:49 <edrx> fixing the makefile...
21:35:38 <crc> it'll require small changes to the input handling (dev_getc in the libretro.c you have, rxReadConsole in the latest source in my repo)
21:35:49 --- join: starlight (~tau@186-194-34-129.i-next.psi.br) joined #retro
21:35:57 <starlight> hi.
21:36:03 <crc> hello
21:41:09 <crc> edrx: I'm going to head to bed now (almost 1am here). this looks very promising, so I'll work with it a bit tomorrow. I should be able to get a string evaluation into place pretty quickly, and update things for the latest vm code.
21:42:03 <edrx> sleep well 8-) (almost 2am here, btw)
21:42:16 <edrx> I'll add a stub for a lua function that call the string evaluation function
21:42:46 <edrx> and I'll put the new version, with a README, at http://angg.twu.net/RETRO/
21:42:56 <crc> sounds good
23:59:59 --- log: ended retro/11.04.16