00:00:00 --- log: started forth/19.12.08 00:05:58 --- join: dys joined #forth 00:27:43 --- quit: dddddd (Remote host closed the connection) 00:36:21 --- quit: WickedShell (Remote host closed the connection) 00:41:48 --- quit: gravicappa (Ping timeout: 265 seconds) 01:21:49 --- quit: f-a (Quit: leaving) 03:00:25 --- join: f-a joined #forth 03:27:56 --- quit: Jookia (Remote host closed the connection) 03:28:25 --- join: Jookia joined #forth 03:30:15 --- quit: f-a (Ping timeout: 245 seconds) 03:32:20 --- join: f-a joined #forth 03:48:47 --- quit: jsoft (Ping timeout: 268 seconds) 04:02:14 --- quit: f-a (Quit: leaving) 04:09:03 --- quit: iyzsong (Quit: ZNC 1.7.1 - https://znc.in) 04:47:57 --- quit: chunkypuffs (Read error: Connection reset by peer) 05:04:14 --- join: chunkypuffs joined #forth 06:26:52 --- quit: dave0 (Quit: dave's not here) 06:45:15 --- join: gravicappa joined #forth 08:11:16 --- join: dddddd joined #forth 08:21:15 --- quit: X-Scale (Ping timeout: 276 seconds) 08:24:12 --- join: X-Scale joined #forth 08:37:53 --- quit: X-Scale (Ping timeout: 265 seconds) 08:41:33 --- join: X-Scale` joined #forth 08:42:39 --- nick: X-Scale` -> X-Scale 09:24:57 --- quit: tabemann (Ping timeout: 250 seconds) 11:28:30 --- quit: gravicappa (Ping timeout: 265 seconds) 12:19:05 --- join: f-a joined #forth 14:19:47 --- join: jedb_ joined #forth 14:21:59 --- quit: jedb (Ping timeout: 240 seconds) 14:58:20 greetings Forthlings 14:58:44 ciao tp 14:58:57 heya f-a :) 14:59:49 f-a, I finally, kinda by accident fixed a major bug caused by a typo 15:00:20 ouch, but good! 15:00:30 how do you debug? word by word? 15:01:13 the program was fairly complex (for me) and I had sorted it out, saved to the SCM all was working then later on the program totally failed! I checked the SCM, nothing had changed, it was very confusing 15:02:05 yeah, word by word, but I can run GDB on the target or simulators, plenty of aids 15:03:03 turned out I had made a minor change to a pretty print word for something else and this was a dependency in the program that had stopped working 15:03:28 o 15:03:29 u 15:03:31 c 15:03:32 h 15:03:47 do you version control stuff? Seems very frustrating of a bug 15:04:00 the dependency wasnt in the program SCM, but in my library, so didnt show up. Whats more the error in the pretty print was a extra space typo! 15:04:22 I use something called "Fossil" 15:04:45 it's made by the author of Sqlite for him to manage sqlite 15:04:46 yeah fossil, good 15:05:23 so this was my most perplexing bug yet 15:06:32 yesterday rdrop-exit asked me to do a memory dump and the dump looked very odd, this was a artifact of my pretty print typo, tho I hadnt realised it at the time 15:07:12 I am imagining tp interactive-debugging a pretty print bug on , say, the ISS 15:07:30 and the pretty print typo bug caused problems switching between bases in the program that had stopped working ... 15:07:49 hahah, non I'm much more 'down to earth' 15:08:21 well if you do *not* fix bugs on the ISS, there is a good chance of things going "down to earth" fast 15:08:43 and of course all these bugs held up other development projects 15:09:24 so Im now considering SCM strategies for multiple independent projects and libraries 15:09:35 it's 'live and learn' ! 15:10:01 heheh, yeah what goes up must eventually come down 15:10:50 how is life at your end, cold, winter, eternal darkness ? 15:11:02 pretty cosy today 15:11:04 --- join: tabemann joined #forth 15:11:19 released a game, ate a exquisite veggie soup, also being sunday 15:11:48 * f-a searches what scm means 15:11:48 I imagine men in bear skins sitting around a hole in the ice fishing for Scandinavian Eels for breakfast 15:12:13 Software Control Method ? 15:12:16 one of the fine pleasures of the dark season :P 15:12:24 oh, oh! thanks 15:12:33 I am bad at acroing :P 15:12:42 I've always typed "SCM" without bothering to really think about it 15:14:14 http://fossil-scm.org 15:14:20 it's even in the url 15:14:32 true! 15:14:37 software configuration management ! 15:14:54 hey 15:14:59 hey tabemann ! 15:17:13 --- join: jedb__ joined #forth 15:19:06 so Im now thinking that instead of fetching words from my library in new projects, perhaps I'll make a symlink from the library word to the new project directory so all dependencies are versioned with the new project files 15:19:41 which would have found my dependency typo issue I think 15:20:03 ohh now I see the problem 15:20:10 yeah seem solid 15:20:30 --- quit: jedb_ (Ping timeout: 276 seconds) 15:20:31 plus the library is in it's own SCM 15:20:38 also test the crap out your libraries, which I unfortunately never do enough 15:20:55 so they will kindof overlap but I cant see any issue with that 15:21:13 well my library words all come from a project initially 15:21:35 makes programming in forth like being a magician 15:21:43 with spells written in a scroll 15:22:10 I'm guessing that all programming languages have similar issues 15:22:47 my bug was further disguised by the fact that the typo didnt cause a compilation error 15:23:15 it just silently caused base havoc 15:23:19 did you have any test running on the library? 15:23:30 no 15:24:04 do you mean a generic test of some kind ? 15:24:43 I presume f-a means unit testing in the junit sense 15:25:23 mhh yeah unit tests 15:25:34 In computer programming, unit testing is a software testing method by which individual units of source code, sets of one or more computer program modules together with associated control data, usage procedures, and operating procedures, are tested to determine whether they are fit for use. 15:25:40 or whatever the best practice® in Forth is 15:26:06 that would be hard for me to do as a lot of my libraries are hardware oriented 15:26:17 ah! yeah, true 15:26:30 for instance, one measures temperature using a specific sensor 15:26:49 and a specific chip etc, etc 15:27:05 Libraries I think are for software people 15:27:29 oops I meant to say "unit tests I think are for software people" 15:27:59 where all their software plugs into a standard fitting like a jigsaw 15:29:08 hmm, one issue with my idea of using a symlink from the library is what happens during a 'revert' ? 15:29:57 tp: only one way to find out, with the vcs I use, it will not modify the link-- linked folder (original) 15:30:03 especially if I changed the library dependency in some way in a project that made it incompatible with another project using the same dependency 15:30:36 f-a, ahh thats a good thing 15:31:04 --- join: dave0 joined #forth 15:31:07 if I want to modify a library part, I should just copy it to the project dir and changes it's name 15:31:49 f-a, as you say, esy to find out, and fossil has configs for every behaviour anyway 16:02:37 back 16:06:25 --- quit: f-a (Ping timeout: 265 seconds) 16:07:52 --- join: f-a joined #forth 16:54:19 --- quit: tabemann (Ping timeout: 250 seconds) 17:25:22 --- join: tabemann joined #forth 17:43:05 --- join: ryke joined #forth 17:51:06 hey guys 18:32:15 --- quit: ryke (Ping timeout: 276 seconds) 18:59:19 --- quit: dddddd (Ping timeout: 246 seconds) 19:08:39 --- quit: tabemann (Ping timeout: 250 seconds) 19:17:20 --- quit: f-a (Remote host closed the connection) 19:26:44 --- join: jsoft joined #forth 19:59:10 --- join: tabemann joined #forth 20:02:23 --- quit: dave0 (Quit: dave's not here) 20:13:27 --- join: gravicappa joined #forth 20:29:39 All rumours 20:30:39 well I tried to tell em that you being a nice guy was just rumours ..... 20:30:48 it's never been proven 20:31:28 however for 1 $billion I'm willing to lie and say that you are in fact a good guy :) 20:35:42 hey guys 20:35:57 hey tabemann 20:36:06 whats new ? 20:36:30 I'm writing zeptoforth so it doesn't have the problem that mecrisp-stellaris has with allocating memory 20:36:36 I'm still partying after finding my bug and now looking at ways to prevent the same kind of thing in future 20:36:39 solution - two HERE pointers 20:37:02 and three sets of words for manipulating them 20:37:14 tabemann, does Mecrisp-Stellaris have a *problem* allocating memory ? 20:37:16 one set for manipulating them for RAM 20:37:26 one set for manipulating them for Flash 20:37:35 one set for either depending on the compilation mode 20:37:47 tp: like where you need that EVALUATE hack 20:37:56 are you going to submit them to mecrisp if they work all good ? 20:38:09 tabemann, ah yeah, but is it actually aproblem ? 20:38:29 Yeah what is this problem 20:38:58 jsoft, I doubt matthias would take them, hes very.very reluctant to change anything in Mecrisp-Stellaris 20:39:17 why's that 20:39:19 where to allocate memory from flash you have to execute : s" BUFFER: 1024" EVALUATE ; or like 20:39:47 is that common to be doing evaluate stuff is it ? 20:39:52 tabemann, let me paste part of a email from matthias on this subject 20:40:34 sure, you would have to do the same under zeptoforth if one wanted to *name* space in RAM, for instance 20:40:52 "Does he want to allocate buffers on-the-fly, and get rid of them with forgetram ? Mecrisp-Stellaris does not have anything beyond variables and statically allocated buffers for default. I usually either define a large buffer: once or change ramstart or ramend to reserve a big chunk of memory and do handling of this myself when I have special memory needs. " 20:40:54 except i bet I could come up with a different approach 20:41:42 I actually kinda agree with that. 20:41:46 i.e. a word that creates a new word on the fly programmatically 20:41:55 Dynamic allocations and what not on uC's Is a bit odd 20:42:02 the problem with : s" BUFFER: 1024" EVALUATE ; is that it cant be accesses from a flash based word later 20:42:35 jsoft, it's common to use 'evaluate' but not in this way 20:43:01 actually shouldn't that be : s" 1024 BUFFER: FOOBAR" EVALUATE ; ? 20:43:17 tabemann, yeah it should 20:43:41 tabemann, but it's useless as it cant be used later from flash 20:43:45 what i forsee is simply ALLOTing space and storing the pointer somewhere predefined 20:44:18 but of course this has to be done from an initialization word in flash 20:44:27 tabemann, it seems that what Mecrisp-Stellaris does is create a ram mirror of a buffer: when the mcu is rebooted 20:45:01 tabemann, is that similar to what matthias is doing now ? 20:45:13 donno 20:45:43 jsoft, matthias wont change anything because it reduces the reliability of Mecrisp-Stellaris and people are using it now 20:45:58 tp, well fair enough :) 20:46:00 I like reliable 20:46:24 jsoft, when I finally, after years of asking matthias to add something, got it done, it broke *MY* code right away 20:46:26 tp, is it quite popular ? 20:46:49 jsoft, hell yeah, up to 42 people are using it worldwide! 20:47:15 :D 20:47:38 jsoft, Mecrisp-Stellaris is the *only* FOSS reliable cortex-m Forth around atm 20:48:18 until tabemann releases his version 20:48:26 or rdrop-exit 20:48:44 or any of the other cortex-m forths under development 20:48:51 Hmm 20:49:18 I must get ahold of an MSP430 thing. That might be quite the bizzo for me sensor projects 20:49:38 the simple fact is that it's not fast to build a complete turnkey forth. It took Matthias 2 years to make "mecrisp" for MSP430 20:50:10 jsoft, you'll die of boredom, MSP430 is like taking candy from a baby after cortex-m 20:50:31 Perfect for being actually productive though 20:50:50 sure, you will find that projects will be a lot faster 20:51:05 Most things I find I do not need barely any horsepower for 20:51:08 but ... lets take your current 'on time' project 20:51:19 msp430 does NOT have a RTC 20:51:29 you'd have had to add one 20:51:59 Right, but its also high power anyway, so I would not be so fussed about using an MSP430 20:52:22 the MSP430 peripherals are *tightly* integrated and that's possible because there are so few of them 20:52:45 The ones I saw all looked useful 20:52:48 GPIO drive/sink current you mean ? 20:53:00 no as in the board as a whole was high power 20:53:06 oh yeah, MSP430 is not trailing edge 20:53:19 mains powered .. I understand 20:53:54 the FRAM (flash) of later MSP430 models is as fast as ram and lasts many times longer than flash 20:54:10 yeah that weirdo ferro magnetic shenanigans eh 20:54:12 but it is worn out a bit, even on read cycles 20:54:58 oh 20:55:00 for instance, the lowest power i can get out of a stm32f051 (without being totally anal) is 4 uA in stop mode 20:55:19 MSp430 does the same at 0.4 uA 20:55:31 Yeah 20:55:39 what about l0's though 20:56:00 the other thing is that STM32F051 is VERY SLOW to restart from stop, MSP430 is FAST 20:56:54 I suspect that STM32L0xxx is similar to MSP430 in low power usage, but it's *very* comples to configure 20:57:07 what about IO ? 20:57:13 How much does this FRAM stuff run out 20:57:23 ? 20:57:27 like how many read cycles 20:57:41 10x more than flash I read 20:57:48 so LOT 20:57:50 S 20:58:17 and the latest FRAM units have a ton more peripherals and arent that expensive 20:58:18 Riiight, but read cycles.. that might clock up pretty quick 20:58:35 yeah, it's debated a lot 20:59:04 but you know flash is only guaranteed for 10 years regarding holding data 20:59:12 nothing lasts forever 20:59:30 except you and me pal 20:59:51 and Ive used up half of forever so far 21:00:00 ... ok, 3/4 of forever 21:01:39 tabemann, so Mecrisp-Stellaris has buffer: for uninitialised data but they must be preallocated. Same with create for flash based arrays etc 21:02:29 bacl 21:02:39 tabemann, however if special ram storage is needed, it can be made using assembly and moving the Forth Ram start area for space 21:03:50 tabemann, so while Mecrisp-Stellaris may appear to have memory allocation problems, does it really ? 21:04:22 I just like ALLOT too much 21:05:23 tabemann, and how often is it required for a small embedded device to dynamically create/destroy ram buffers/arrays during normal operation ? 21:05:57 I'm working with the cortex-M4, and I doubt that counts as a "small embedded device"! 21:05:59 tabemann, as opposed to just using predefined buffers for the task as programmed ? 21:07:05 tabemann, M4 has a limit of 1MB flash so it probably is "medium" small embedded ;-) 21:07:36 the GDf103 has a limit of 3MB flash I believe 21:08:01 because it uses serial flash on the die that overlays into ram on boot up 21:08:09 something like that 21:09:45 okay, I'm gonna hit the sack now 21:09:49 g'night guys 21:18:29 night tabemann 21:37:56 jsoft, how goes the xtal mod ? 21:41:45 tp, its not going, im going to do it in the next board 21:42:02 I don't ahve the assembled clock thingy with me, its been given :) 21:42:03 ah 21:42:20 wow, that was a FAST job! 21:42:40 I thought you already knew that, I gave it last friday, hence why I was up till 5am 21:42:42 no way I could develop something that fast. It would take me a couple more years 21:42:57 no I dint know, very well done! 21:43:08 so worlds most inaccurate clock ? 21:43:12 lol yep :) 21:43:33 a cheap mechanical watch can be out 5 minutes a day 21:43:33 I mean, I could potentially have stayed up till 6am or something and risked ballsing some soldering up to put the xtal on 21:43:56 yeah, no time left is never a good time to do something like that 21:44:28 But yeah, I want to make Rev B over the holidays 21:44:38 With better code and better layout and an xtal 21:44:47 and easter eggs! 21:44:50 Hmm 21:44:55 Thats an idea :) 21:45:14 if you sen me one I'll be checking for a thin layer of C4 on the back ;-) 21:45:27 painted over 21:45:33 lol why would I do that 21:45:54 kiwis hate ausies, theyre jealous of out startling good looks 21:46:35 No they are just jealous of the delusional bliss they appear to be in, as demonstrated by your last comment :P 21:46:42 andand our intelligence, and our bank accounts which have $1 billions in them! 21:46:59 oh yeah, and that :) 21:47:10 --- join: smokeink joined #forth 21:47:50 But yeah I think I like your hackaday idea. It has some kicad stuff that might be usefull to someone too, python scripting to lay out the round tracks, and place the leds 21:47:56 so you'll swap it for one with a xtal when done ? 21:48:09 Probably yeah 21:48:17 If they guy wants it swapped out 21:48:21 I think your project could be the premier Mecrisp-Stellaris project ever! 21:48:59 Perhaps :) 21:49:02 we have plenty of bling-less fubctional scientific projects that the Slackaday crowd wouldn't look twice at 21:49:34 but youur clock will send them into a blinged out frenzy! 21:50:11 Perhaps a USB usart on there, for ease of shenanigans 21:50:43 look at the love they have for fake 'nixie' clocks ? 21:50:52 Not really 21:51:00 Nixie clocks are cool though. 21:51:04 they do, those projects are top scoreres 21:51:25 well they dont do anything for me, I grew up with nixies 21:51:58 so perhaps stick in a stm32f103 ? 21:52:10 Whys that ? 21:52:13 USb 21:52:26 plus we have a USB driver for it 21:52:40 Well I have spare whatever chips which do USB->TTL 21:52:52 ah, same thing really 21:53:07 just some extra soldering and layout 21:53:32 lol, why not just make the new clock take a blue pill board ? 21:53:38 haha, $2 21:53:49 Well stm32's are hardly expensive as is 21:53:54 Compared to the actual PCB cost 21:53:55 some are 21:54:06 Of the clockosaurus-rex 21:54:16 It cost like, $140 for 5 pcb's 21:54:20 a new factory stm32f103 is about $3.80 usd at arrow 21:55:04 a 'new' bluepill with usb connector, reg, rtc xtal and 8mhz system clock xtal us generally under $2 usd 21:55:39 Yeah I kinda want to do 'discrete' non module board thingy 21:55:54 I do think the STm32F103 on them is recycled tho, and they may come with a few clones instead of the stm32 21:56:14 kudos, you have a real designer in you 21:56:47 but utilising a blue pill would get hackaday going mental 21:57:03 there must ne millions of unused bluepills waiting for a home 21:57:04 lol really? Are they weirdos over there ? 21:57:08 yeah 21:57:16 all emo sjw's 21:57:21 Hah :D 21:57:43 many are arduino fanboys 21:57:58 Ahh arduino. 21:58:01 Those things. 21:58:15 I've only ever had one Forth project accepted there 21:58:17 I've not used one in some time 21:58:25 What was that ? 21:58:36 they have rejected others which I know were far more deserving 21:59:45 Got a link to it ? 22:01:17 trying to find it 22:03:02 https://hackaday.com/2017/05/13/microcontroller-load-meter-tells-you-how-hard-its-currently-working/ 22:06:40 this is my project: https://mecrisp-stellaris-folkdoc.sourceforge.io/instrument-mcu.html?highlight=loadmeter 22:09:26 it was my old vintage dB meter that got the project accepted, youll notice they make no mention of Forth and never updated the link 22:10:47 Forth is unfashionable on hackaday, just mentioning the word probably costs them readers 22:11:15 yet ... hackaday have had a long term lurker on my Forth channel for many months 22:11:39 obviously on the lookout for something that will increase readership 22:23:14 Hmm 22:23:27 What was it, some sort of RC thing into the meter ? 22:24:08 so just dont mention Forth as your project OS initially, just give the binary ;-) 22:24:15 thats all yeah 22:24:32 just a resistor iirc 22:25:01 tp, hmm good point :) 22:25:03 --- quit: smokeink (Remote host closed the connection) 22:25:33 it's enough for 99% of builders to have the binary 22:25:49 they dont care what it is as long as it works 22:26:23 you could even offer pre built and programmed boards for 1$billion 22:27:26 lol 22:27:34 to most kids thesedays, building something is buying a prebuilt and tested board and putting it in a case, or for the more adventurous, buying the parts (inc a pre programmed chip) and soldering them 22:27:55 you could offer a 'free' custom birthday alarm! 22:28:07 they just supply a birthdate with the order 22:28:49 make up a website and advertise that *one* product, I bet youll sell a minimum 100 units a year and make money on the boards and postage 22:29:41 Ive personally sold over $4 million in WiFI gear on my old website over 15 years 22:31:49 Really ? 22:32:13 yep 22:32:23 really 22:32:27 Much of a margin ? 22:32:42 nah, 15% average 22:32:54 it was a living tho, pretty good at times 22:33:05 I closed that biz at the start of this year 22:33:18 started it in 2004 22:34:46 in the end overseas biz running massive warehouses here sold at lower prices than I could buy so that was that, but it was good for 10 of the 15 years 22:35:04 and I was only reselling at low margins 22:36:18 but you have a actual product that someone would have to copy to make cheaper, so even if you only made $10 a unit as a kit, all youd have to do is pack n ship, or get someone else to do it 22:36:37 you could operate out of a digital ocean account for $5USD a month 22:36:45 Hmm 22:37:11 and anything else that you make that's useful could also be added to it 22:37:38 you just supply binaries, you dont even mention forth, people wont care 22:38:38 if you use a msp430 and mecrisp-across, the binaries are 100% yours you dont have to give the source 22:39:11 makes you wonder how many products actually run binaries produced by a tethered Forth 22:39:20 there ia no way of knowing 22:39:40 the robo vac, your washing machine ? 22:53:23 Perhaps even.. the light switch... 22:53:46 one never knows 22:54:21 popular belief is that C or C++ builds everything, and the fanboys are 100% certain 22:55:09 they even think that C/C++ is so awesome that it's use in space is undeniable and inevitable 22:56:32 but they dont understand that when it comes to space, the costs are so unbelievably high and the environment so hazardous that decades old but proven gear is used in preference 22:56:50 and that usually means rad hardened Forth gear 22:57:24 jsoft, you know where the slowest computer response I ever saw was ? 22:57:36 ... military 22:57:44 Prolly some RF speed of light based thing? 22:58:12 Or do you mean response to technology 22:58:30 the system has 3 computers and a voting mechanism between them for the utmost in reliability 22:58:50 it was on a current au submarine 22:58:59 well a number of subs 22:59:35 want to read a manifold temperature ? enter your security code and wait.... and wait .... and wait 22:59:52 eventually you get a value 23:00:18 There is some outrageous 50 hz hum outside, as if HV transformer is on the fritz or something 23:00:27 slowest system I ever saw, no doubt I had the lowest of the low priorities but there were only a few people on board 23:00:30 * jsoft goes to investigate 23:00:52 probably your clock drawing megawats to run those serial leds! 23:01:55 That thing is not within hearing range of me :P 23:02:02 Nah its some motorized gizmo 23:02:14 T2 terminator ? 23:02:19 Perhaps 23:02:57 a Tesla limping home ? 23:03:12 seen the Tesla "cybertruck" ? 23:03:39 Yeah some square thing 23:05:09 that 'Cybertrick' is a total gamechanger 23:05:50 many people think it's basic or of not much use, but theyre so wrong 23:06:26 Musk is a utter genius 23:19:57 did you know that the 'Cybertruck' stores 15 miles a day of energy in the sun, when it has the optional solar roof ? 23:20:44 --- join: smokeink joined #forth 23:29:06 --- quit: smokeink (Remote host closed the connection) 23:29:19 --- join: mtsd joined #forth 23:59:59 --- log: ended forth/19.12.08