00:00:00 --- log: started forth/19.01.08
00:15:28 --- quit: Keshl (Read error: Connection reset by peer)
00:16:05 --- join: Keshl (~Purple@24.115.185.149.res-cmts.gld.ptd.net) joined #forth
00:18:58 --- quit: jedb (Remote host closed the connection)
00:19:19 --- join: jedb (jedb@gateway/vpn/mullvad/x-yaspdfadzujyppkw) joined #forth
00:24:35 --- join: xek (~xek@apn-31-0-23-83.dynamic.gprs.plus.pl) joined #forth
00:37:51 --- join: pierpal (~pierpal@95.239.223.85) joined #forth
01:51:58 <WilhelmVonWeiner> clear
01:52:11 <WilhelmVonWeiner> ...wrong window - whoops
01:52:16 * Zarutian charges the defiblator again
02:02:18 <jackdaniel> - Doctor, patient runs!; - Charge defiblator! We'll catch him.
02:02:55 --- quit: ashirase (Ping timeout: 245 seconds)
02:15:45 <WilhelmVonWeiner> siraben: how have you defined DO?
02:18:02 --- join: ashirase (~ashirase@modemcable098.166-22-96.mc.videotron.ca) joined #forth
02:47:49 <siraben> WilhelmVonWeiner:  yes
02:53:24 <WilhelmVonWeiner> how have you defined yes?
03:00:31 --- join: proteusguy (~proteus-g@cm-58-10-154-246.revip7.asianet.co.th) joined #forth
03:00:31 --- mode: ChanServ set +v proteusguy
03:22:57 --- quit: proteusguy (Ping timeout: 250 seconds)
03:28:21 --- quit: Zarutian (Read error: Connection reset by peer)
03:28:27 --- join: Zarutian_2 (~zarutian@173-133-17-89.fiber.hringdu.is) joined #forth
03:28:39 --- nick: Zarutian_2 -> Zarutian
03:36:18 --- quit: `presiden (Quit: WeeChat 2.4-dev)
03:47:22 <siraben> WilhelmVonWeiner: it's implemented
03:51:44 <WilhelmVonWeiner> I think you misread my question
04:06:26 <corecode> siraben: why is it hard?
04:18:49 --- join: dddddd (~dddddd@unaffiliated/dddddd) joined #forth
04:27:00 <siraben> It's probably a weird bug I'm missing, hm
04:28:30 <corecode> i need an infrastructure to write tests
04:28:48 <corecode> right now i can run forth, but i can't get the results out
04:36:01 <corecode> so should POSTPONE some-immediate insert the whole thread of the immediate in the current location
04:36:08 <corecode> or should it insert a call
04:36:39 <corecode> the difference is what is on the return stack while the thread of the immediate is executing
04:44:01 <siraben> corecode: I use a test suite
04:48:26 --- quit: Zarutian (Ping timeout: 246 seconds)
04:49:29 --- join: Zarutian (~zarutian@173-133-17-89.fiber.hringdu.is) joined #forth
04:54:10 <corecode> any suggestions on the POSTPONE behavior?
04:56:58 <corecode> siraben: https://github.com/corecode/forth/blob/master/x86.fs#L339 ?DO
05:31:40 --- quit: pierpal (Quit: Poof)
05:31:59 --- join: pierpal (~pierpal@95.239.223.85) joined #forth
05:33:28 --- join: rdrop-exit (~markwilli@112.201.166.158) joined #forth
06:49:02 --- join: Kumool (~Khwerz@adsl-64-237-235-188.prtc.net) joined #forth
06:54:10 --- quit: Kumool (Ping timeout: 245 seconds)
06:55:14 <Zarutian> corecode: later.
07:01:39 <corecode> later what
07:01:51 <corecode> insert call?
07:03:12 <pointfree> siraben: Check out "compressed pattern matching" -- searching compressed data without decompression
07:04:10 <pointfree> antidictionary compression is particularly amenable to compressed pattern matching
07:08:23 <Zarutian> corecode: cant we just postpone this discussion?
07:08:54 <Zarutian> pointfree: why and where do you find these bizzarely interesting topics?
07:11:02 --- quit: rdrop-exit (Quit: Lost terminal)
07:11:07 <corecode> Zarutian: the reference to or the content of the discussion?
07:11:16 --- join: Kumool (~Khwerz@adsl-64-237-235-188.prtc.net) joined #forth
07:11:38 <Zarutian> corecode: sorry, I was just pulling your leg.
07:13:02 <Zarutian> corecode: What POSTPONE does in most Forths I have seen is take the next word in the input stream, look it up in the dictionary and COMPILEing that xt into the current word being defined.
07:13:31 <corecode> yea
07:13:43 <corecode> that;s what i have
07:14:35 <pointfree> Zarutian: I came across compressed pattern matching back in high school while working on a scheme for similarity enhanced transfer + named-data networking https://www.cs.cmu.edu/~dga/papers/nsdi2007-set/ https://en.wikipedia.org/wiki/Named_data_networking
07:14:57 <Zarutian> this is basically to circumvent running the looked up word at that time where it is being used in a composition of a new word definition even when the word is an IMMEDIATE word
07:17:24 <corecode> yea
07:18:07 <pointfree> PoppaVic: What I don't like about the standard loop constructs is the use of the I J K registers. They make it tricky to factor the body of a loop. I just use the dictionary pointer as the loop index. 
07:18:07 <pointfree> All the dictionary words  such as HERE ALLOT C, , and anything using them work and it can all be easily factored.
07:18:15 <corecode> well if it isn't an immediate, you don't compile in the XT, but you compile in the equivalent of ['] word COMPILE, , i.e. compile in code that will compile in the XT
07:19:30 <Zarutian> right, forgot an level of meta here
07:19:33 <corecode> (:
07:22:07 <Zarutian> pointfree: re that paper you just linked to. Pretty interesting. But so far I had looked at named data networking to be a bit of a superset of how freenet works.
07:23:50 <Zarutian> pointfree: for instance I had sketched out an extremely simple TFTP like udp based protocol using merkel tree cryptographic hashes of 1 KibiByte chunks
07:27:01 --- quit: tabemann (Ping timeout: 252 seconds)
07:27:21 <pointfree> Zarutian: with ndn, routing is done by matching prefixes of actual data (not location), even at the link layer sometimes.
07:28:15 <pointfree> (it's not an overlay)
07:29:10 <Zarutian> pointfree: basically, udp packets directed to a wellknown port and wellknown multicast address. If the first byte in the payload is 0x00 then the next 1024 bytes is a data chunk. If the first byte is 0x01 then the sender is soliciating for the data chunk matching the sha256 hash (next 64 bytes)
07:30:28 <Zarutian> pointfree: had not decided what to do with the rest of the possible values of the first byte other than reserving 0xFF for future expansions for when and if the rest of the possible values had been used up for something else.
07:34:13 <Zarutian> pointfree: main starting motivation for it was finding an utility for the ginormous memory many routers have other than bloating packet buffers
07:35:49 <Zarutian> pointfree: another use was for diskless systems or other such to use the LAN as fast (possibly encrypted) swap.
07:39:34 * Zarutian continues to read that paper
07:43:07 --- quit: pierpal (Read error: Connection reset by peer)
07:43:25 --- join: pierpal (~pierpal@95.239.223.85) joined #forth
07:47:31 --- quit: pierpal (Read error: Connection reset by peer)
08:00:36 <john_cephalopoda> Heh, work on my matrix lib goes well :D
08:06:35 * Zarutian fiddles: 0x7B CONSTANT openCurly 0x7D CONSTANT closeCurly 0x5C CONSTANT backslash 
08:09:36 <john_cephalopoda> https://bpaste.net/show/87d9c63afa66
08:09:50 <john_cephalopoda> I can already do matrix addition, now working on matrix multiplication.
08:50:49 <corecode> yey now i can parse negative numbers
08:52:16 <Zarutian> corecode: like your bank balance? (<- traditional jib regarding parsing negative numbers)
08:54:36 <corecode> nah, i'm in IT, my balance is positive
08:56:02 <corecode> (still)
10:16:04 --- join: dys (~dys@tmo-102-55.customers.d1-online.com) joined #forth
10:25:13 --- join: proteusguy (~proteus-g@cm-58-10-154-246.revip7.asianet.co.th) joined #forth
10:25:13 --- mode: ChanServ set +v proteusguy
10:28:21 --- join: pierpal (~pierpal@95.239.223.85) joined #forth
10:33:51 --- quit: X-Scale (Ping timeout: 244 seconds)
10:35:24 <PoppaVic> pointfree: yes, I like to use HERE when possible.
10:35:31 --- quit: gravicappa (Ping timeout: 246 seconds)
10:56:35 <PoppaVic> pointfree: I tend to think that the loop-vars should really exist on their own stack - just to get them the hell out of the way.
10:59:57 <pointfree> https://www.reddit.com/r/Forth/comments/6loglg/loops_that_can_be_factored_use_the_dictionary/
11:01:38 <pointfree> PoppaVic: I don't use any loop variable words at all.
11:11:23 * pointfree would write a few things differently since writing that reddit post. 
11:13:23 <WilhelmVonWeiner> john_cephalopoda: your matrix multiplication only works on matrices of equal size?
11:17:46 <corecode> so what would be a good command set for a forth cpu, for small code
11:18:07 <corecode> my feeling is that 16 bit stack cells would be useful
11:18:23 <corecode> or even 32 bit
11:18:37 <corecode> but then maybe 8 bit tokens?
11:18:44 <WilhelmVonWeiner> corecode: look up the F18A, SeaForth, Sh-Boom, Nga, Ngaro, RTX2000, and zero-operand CPUs in general
11:19:26 <corecode> yea i read about the willow garage one
11:19:35 <corecode> thanks for the pointers
11:21:36 <WilhelmVonWeiner> This was the SeaForth instruction set https://m.eet.com/media/1066227/intellasysfigure2.gif
11:23:56 <WilhelmVonWeiner> corecode: F21, too
11:24:04 <WilhelmVonWeiner> that's well documented on Ultratechnology
11:25:07 <corecode> interesting
11:31:24 <john_cephalopoda> WilhelmVonWeiner: Nope, it works for all matrices that got the right size.
11:35:42 <john_cephalopoda> WilhelmVonWeiner: The number of columns of the second matrix must be equal to the number of rows of the first matrix. That's a requirement that rises from the definition of matrix multiplication .
11:36:11 <WilhelmVonWeiner> I know how matrices work, I must've misread a comment
11:53:16 <john_cephalopoda> WilhelmVonWeiner: I got a function for checking if matrices are of equal sizes, which is used in the addition function.
11:55:45 <corecode> WilhelmVonWeiner: thanks for your input
11:56:00 <corecode> so seaForth and F18A have a very similar code set
11:59:34 --- join: [1]MrMobius (~default@c-73-134-82-217.hsd1.va.comcast.net) joined #forth
12:00:45 --- join: mtsd (~mtsd@94-137-100-130.customers.ownit.se) joined #forth
12:02:50 --- quit: MrMobius (Ping timeout: 268 seconds)
12:02:50 --- nick: [1]MrMobius -> MrMobius
12:19:42 --- quit: pierpal (Read error: Connection reset by peer)
12:22:41 --- join: X-Scale (~ARM@83.223.243.180) joined #forth
12:23:02 --- join: pierpal (~pierpal@95.239.223.85) joined #forth
12:26:17 --- quit: pierpal (Read error: Connection reset by peer)
12:28:36 --- join: pierpal (~pierpal@95.239.223.85) joined #forth
12:31:28 --- quit: pierpal (Read error: Connection reset by peer)
12:31:44 --- join: pierpal (~pierpal@95.239.223.85) joined #forth
12:38:08 --- quit: pierpal (Read error: Connection reset by peer)
12:55:01 --- quit: mtsd (Quit: WeeChat 1.6)
13:02:46 --- quit: xek (Remote host closed the connection)
13:03:11 --- join: xek (~xek@apn-31-0-23-83.dynamic.gprs.plus.pl) joined #forth
13:13:25 <john_cephalopoda> Yay!
13:13:28 <john_cephalopoda> Matrix multiplication works!
13:14:10 <john_cephalopoda> https://thecutecuttlefish.org/tmp/matrix.fth
13:40:12 <the_cuckoo> nice :)
13:50:59 <the_cuckoo> wasn't planning to do anything myself, but the parser logic was bugging me - had to test my theory and it's panning out - nested do/loop working as expected, plenty of room for trivial optimisation (but reasonably fast as is)
13:52:16 <the_cuckoo> and none of this "Interpreting a compile-only word" stuff :)
13:53:25 <the_cuckoo> also thinking - i don't think i'll ever need a return stack... have a different approach for that one
13:55:00 <corecode> the_cuckoo: what do you do?
13:56:19 <the_cuckoo> well, i'm not implementing forth - it's a c++ api which implements a non-standard stack language, so i have the whole of c++ to play with
13:56:34 * PoppaVic hurls
13:56:38 <the_cuckoo> :D
13:56:53 <PoppaVic> evil (orange?) man!
13:57:13 <the_cuckoo> tee hee
13:57:57 <corecode> the_cuckoo: many operations seem to be easier with two stacks
13:59:28 <the_cuckoo> yes - but they're also easier still with many :) - in essence for the do/loop stuff, i have a specific 'variable' (associated to the stack object) which holds a stack of loop state (which is currently a pair - index and leave state)
14:00:14 <the_cuckoo> other stuff can have their own variables which hold their own state (stacked or not depending on context)
14:02:51 <the_cuckoo> but still playing with the idea - i could easily have a 'return stack' which you can dump arbitrary objects on, but think that'll be harder to manage - but still playing with it - may change direction yet
14:02:52 <WilhelmVonWeiner> stacks are simpler than variables, I wonder about the speed comparison though
14:03:16 <WilhelmVonWeiner> I'm not well learned how how memory and registers and variables are handled
14:04:34 <the_cuckoo> ah - this ain't going to be a speed demon - at the moment, without much optimisation going on, i'm like 10x slower than python - fairly sure i can compare with python though
14:05:52 <the_cuckoo> generally, it's more about 'domain specific languages' than it is about being a really generic language in itself 
14:07:56 <the_cuckoo> 10x slower than python is still like 100x faster than bash :p
14:08:38 <WilhelmVonWeiner> if a solution is fast in one language it's probably fast enough in any language
14:10:16 <the_cuckoo> sure - when i call a c++ function, it'll run as fast it would under normal circumances - and that's kinda the point - the c++ code does the heavy lifting - the stack language just provides the ability to invoke them with some branching and looping
14:11:00 <the_cuckoo> repl style rather than going through the normal edit/compile sequence iyswim
14:12:00 <corecode> so how do you do the gateway?
14:12:04 <WilhelmVonWeiner> I wasn't talking about your program in particular, just as a general concept.
14:12:25 <corecode> do you generate stubs for all your functions/methods?
14:12:44 <corecode> and how do you deal with complex data
14:13:19 <the_cuckoo> complex data can go on the stack - the stack is made up of boost::any objects which can wrap pretty much any c++ object
14:13:48 <the_cuckoo> so, basically, the 'binding' stuff is just fetching from the stack, casting for use, pushing results back
14:14:05 <the_cuckoo> currently just using lambdas
14:14:51 <the_cuckoo> https://gitlab.com/lilo_booter/rpany/blob/master/vocabs/foo.cpp <- v. basic class
14:15:18 <the_cuckoo> and similar, but consumes some args: https://gitlab.com/lilo_booter/rpany/blob/master/vocabs/rectangle.cpp
14:19:06 <the_cuckoo> dunno if it makes a lot of sense to do it this way, but the code is really very mechanical to write - and as it's a header only library (with a couple of non standard headers coming from boost), it's trivial to integrate into any system 
14:24:53 <WilhelmVonWeiner> http://sametwice.com/herkforth_words
14:25:11 <WilhelmVonWeiner> crc has mentioned these before (0; is an instruction in Nga) but they're so good.
14:25:41 <WilhelmVonWeiner> I wish `if;` was written in CODE, or at least I wish I knew how to define it so
14:26:30 <the_cuckoo> if is always ugly :(
14:27:10 <Zarutian> hmm
14:27:12 <WilhelmVonWeiner> most of the time I abstract any if statement into it's own word. 
14:27:23 <WilhelmVonWeiner> that way I can rdrop out and keep a sane control flow
14:27:49 <Zarutian> : IF COMPILE BRZ HERE ; IMMEDIATE
14:28:47 <WilhelmVonWeiner> BRZ?
14:28:52 <Zarutian> branch if zero
14:29:19 <Zarutian> : BRZ R> DUP @ SWAP 1+ ROT ?: >R EXIT
14:29:44 <Zarutian> : ?: SKZ SWAP DROP ;
14:29:50 <Zarutian> s/EXIT/;/
14:30:22 <Zarutian> where SKZ is SKIP if zero (skips the next word-xt)
14:32:45 <Zarutian> SKZ is much easier to implement in VMs and microcode, just increment the PC again if the TOS is zero
14:59:34 --- join: grosged (~grosged@bop62-1-88-178-198-84.fbx.proxad.net) joined #forth
15:00:04 --- quit: grosged (Client Quit)
15:03:33 <corecode> 1+? not 1 CELLS +?
15:07:07 --- quit: dys (Ping timeout: 250 seconds)
15:08:42 <Zarutian> corecode: CELL+ is just aliased to 1+ in the forth I have as I have cell based addressing and not byte addressing
15:12:57 <PoppaVic> gonna' make it more painful to support/port w/o CELLS in there, in any case
15:16:35 <crc> WilhelmVonWeiner:   for retro, if; can be defined to inline the machine code:   :if; compiling? [ 'lixozr.. i #-1 d 'dr...... i ] if ; immediate 
15:18:08 <john_cephalopoda> Wtf am I doing?
15:18:27 <john_cephalopoda> I spent basically all day creating a matrix library and now I can render points in 3d.
15:18:39 <WilhelmVonWeiner> crc: would those not need to be in as{ }as?
15:19:10 <crc> not here, since I want this to assemble the instructions and data into the definition using if;
15:19:56 <john_cephalopoda> https://thecutecuttlefish.org/tmp/matrix.fth <-- The matrix library
15:20:01 <john_cephalopoda> https://thecutecuttlefish.org/tmp/sixel.fth <-- The sixels library
15:20:05 <Zarutian> PoppaVic: well : CELL+ 1 CELLS + ; would work
15:20:23 <john_cephalopoda> https://thecutecuttlefish.org/tmp/spatial.fth <-- A small demo, drawing points into a 3d scene.
15:20:38 <john_cephalopoda> The points on the left are perspectively transformed.
15:20:55 <WilhelmVonWeiner> crc: I will have to try and write some useful Nga assembler. 
15:21:29 <WilhelmVonWeiner> john_cephalopoda: you duplicate so much code
15:21:58 <WilhelmVonWeiner> two best refactoring methods in Forth: Moving things from runtime to compile time, and compile time to design time
15:23:18 <john_cephalopoda> The duplicates in spatial.fth are deliberate.
15:25:52 <john_cephalopoda> It's only there to give a short demo. Nothing in there is designed.
16:07:02 --- quit: john_cephalopoda (Ping timeout: 252 seconds)
16:09:00 --- join: john_cephalopoda (~john@unaffiliated/john-cephalopoda/x-6407167) joined #forth
16:45:59 --- join: dave0 (~dave0@47.44-27-211.dynamic.dsl.syd.iprimus.net.au) joined #forth
16:46:20 <dave0> hi
16:52:51 --- join: rdrop-exit (~markwilli@112.201.166.158) joined #forth
17:03:34 <rdrop-exit> Greetings Forthwrights
17:04:06 <dave0> hi rdrop-exit 
17:04:15 <rdrop-exit> Hi Dave0
17:12:50 --- quit: proteus-guy (Ping timeout: 250 seconds)
17:17:06 --- join: proteus-guy (~proteusgu@2403:6200:88a6:329f:55c1:9c9e:ed64:b2a8) joined #forth
17:17:55 --- quit: X-Scale (Ping timeout: 245 seconds)
17:22:20 --- quit: Kumool (Quit: EXIT)
17:28:09 --- join: X-Scale (~ARM@83.223.243.32) joined #forth
17:31:09 --- join: `presiden (~presiden@unaffiliated/matematikaadit) joined #forth
17:34:38 --- join: tabemann (~tabemann@h193.235.138.40.static.ip.windstream.net) joined #forth
17:52:01 --- quit: xek (Remote host closed the connection)
18:12:23 --- join: xek (~xek@apn-31-0-23-83.dynamic.gprs.plus.pl) joined #forth
18:12:51 --- quit: xek (Remote host closed the connection)
18:13:42 --- join: xek (~xek@apn-31-0-23-83.dynamic.gprs.plus.pl) joined #forth
18:14:51 --- quit: tabemann (Ping timeout: 272 seconds)
18:35:42 --- join: tabemann (~tabemann@2602:30a:c0d3:1890:7d04:1e92:6395:7db) joined #forth
19:12:39 --- join: Kumool (~Khwerz@adsl-64-237-235-188.prtc.net) joined #forth
19:35:17 --- quit: proteusguy (Remote host closed the connection)
19:49:32 --- join: [1]MrMobius (~default@c-73-134-82-217.hsd1.va.comcast.net) joined #forth
19:52:40 --- quit: MrMobius (Ping timeout: 250 seconds)
19:52:41 --- nick: [1]MrMobius -> MrMobius
20:06:31 --- join: gravicappa (~gravicapp@h109-187-38-226.dyn.bashtel.ru) joined #forth
20:44:14 --- quit: dddddd (Ping timeout: 250 seconds)
20:53:05 <tabemann> quiet is #forth tonight
20:53:10 --- join: Keshl_ (~Purple@24.115.185.149.res-cmts.gld.ptd.net) joined #forth
20:53:20 --- join: dave69 (~dave0@47.44-27-211.dynamic.dsl.syd.iprimus.net.au) joined #forth
20:53:31 <tabemann> hey
20:53:34 --- join: [1]MrMobius (~default@c-73-134-82-217.hsd1.va.comcast.net) joined #forth
20:53:39 <PoppaVic> oh. Hi, Yoda.. Yo-yo-yo-yoda...
20:55:04 --- quit: Kumool (Quit: EXIT)
20:56:01 --- join: X-Scale` (~ARM@83.223.243.32) joined #forth
20:57:27 <tabemann> finally got hashforth working
21:02:31 --- quit: MrMobius (*.net *.split)
21:02:31 --- quit: X-Scale (*.net *.split)
21:02:31 --- quit: dave0 (*.net *.split)
21:02:31 --- quit: Keshl (*.net *.split)
21:02:31 --- quit: gabc (*.net *.split)
21:02:32 --- nick: [1]MrMobius -> MrMobius
21:02:33 --- nick: X-Scale` -> X-Scale
21:04:29 --- join: gabc (~gabc@165.227.34.49) joined #forth
21:05:43 --- nick: dave69 -> dave0
21:14:40 --- quit: dave0 (Quit: dave's not here)
21:22:44 --- join: dys (~dys@tmo-108-204.customers.d1-online.com) joined #forth
21:45:51 <crc> tabemann: the repository is missing the Makefile
21:46:34 <tabemann> added
21:46:56 <tabemann> at this point it doesn't crash when lightly touched
21:47:44 <tabemann> it can't do a whole lot - in the end all it can do is read from and write to the terminal - but as written it could accept a simple cooperative multitasker
21:49:16 <tabemann> opinion question - should I add support for heap allocation, considering that this is a proof-of-concept for something that may not have a heap, as in on really small systems
21:50:00 <tabemann> also, with a small bit of modification I could add support for FORGET or MARKER
21:52:11 <crc> I'll try building it on my systems (OpenBSD & FreeBSD) later today
21:52:39 <tabemann> there's an included image so you don't need to download and build attoforth just to use it
21:53:07 <crc> I don't use heap allocation at all (assuming you mean allocate and free or similar)
21:53:14 <tabemann> yeah
22:04:50 <crc> Doesn't build on OpenBSD with that Makefile. https://forthworks.com/share/6cbee8543df57e408376e707e86ce122 with BSD make, https://forthworks.com/share/1c5a017a563c4385d9353a9a7226786e with GNU make
22:05:46 <crc> Getting late here (1am), so I won't be able to look further for a few hours
22:30:48 <tabemann> back
22:31:13 <tabemann> I made some fixes to the code (I forgot to add a file to the git repository)
22:34:19 --- quit: dys (Ping timeout: 250 seconds)
23:57:08 --- join: dave0 (~dave0@47.44-27-211.dynamic.dsl.syd.iprimus.net.au) joined #forth
23:57:20 <dave0> re
23:59:59 --- log: ended forth/19.01.08