#avr Logs

Oct 30 2019

12:21 AM day__ is now known as day
12:42 AM rue_bed2: --
02:00 AM flip214: I got simulavr from git... but the timestamp column mentioned in the manual for tracing isn't implemented! It was easy enough, but other parts of the output (eg register values) are missing, too...
02:00 AM flip214: so I'm wondering what the canonical version of simulavr is - perhaps I got the wrong one?
04:32 AM twnqx: egy: let's start over. what do you want to achieve, what do you think you achieve, and how do you come to the conclusion of what you achieve.
04:33 AM twnqx: ucsrc and ubrrh are the same register. if bit 7 is 1, bits 0-6 get written to ucsrc; if bit 7 is 0, bits 0-6 get written to ubrrh
04:33 AM twnqx: to if you write 128 to ubrrh, what you actually do is setting ucsrc all-zero
04:33 AM twnqx: so if*
04:45 AM twnqx: also given the special read access rules for this register (first consecutive read yields ubrrh; second yields ucsrc, if exactly the previous instruction read ubrrh), i doubt a debugger can show the actual value of ucsrc
09:56 AM skz81: twnqx, he's using the AS7 simulator here I guess (not on-chip debuging)
09:57 AM twnqx: oh
09:57 AM twnqx: well, the pics expired before i could open them
09:57 AM twnqx: and i never did simulation, so no idea about that :P
10:34 AM skz81: 2.png is not found, but 1.png work for me... Anyway :)
10:34 AM skz81: workS even
10:36 AM twnqx: yeah, but that just has some code :PÜ
10:38 AM skz81: twnqx, Ok i beleive you, now : you're REALLY not used to that simulator and its UI ^^
10:39 AM twnqx: i don't even have winblows, and last i checked AS was winblows-only
10:50 AM nohit: i think one of atmel's mistake was that they didnt make cross-platform IDE
10:51 AM nohit: and they didnt make cheap development tools. ST just killed atmel with that
10:51 AM nohit: but AS and the simulator is great
10:52 AM nohit: i have used the simulator since avr studio 4
11:36 AM twnqx: why not just test on hardware?
11:42 AM nohit: why people use simulators in the first place
11:46 AM twnqx: dunno
11:46 AM PoppaVic: simulators are... wonky.. It's like pretending "GDB will save me!" - but, yes: if you test on an actual chip you are going to burn thru at least one and maybe more - and you can't really single-step or whatever like a typical processor
11:46 AM twnqx: they can't interface with anything, and you can't really simulate an arbitrary piece of hardware
11:47 AM PoppaVic: So, I eschew simulators AND gdb.. But, that's me - Others might get paid to be silly.
11:47 AM polprog: gdb + avarice is the bees knees
11:47 AM twnqx: gdb is kind of ok, if you have a jtag interface
11:48 AM polprog: best of both worlds
11:48 AM twnqx: (or one wire debug or whatever)
11:48 AM PoppaVic: there are no bees-knees <shrug> Lotsa' wishful thinking.
11:49 AM twnqx: but i kind of agree PoppaVic, gdb never really helped me with anythiing i couldn't have seen with just reading code :P
11:49 AM PoppaVic: in the Long Ago, I found those tools useful for memory-leaks on linux.. BUT, the thing always was: it wasn't my code leaking.
11:50 AM PoppaVic: (the exception was always the early stages of a list or tree I wrote from scratch. Thereafter, it was ALWAYS some stupid lib allocating and not documenting)
11:52 AM polprog: lol, true, painfully
11:52 AM polprog: the only time ive extensively used gdb was when playing with memory leaks :v
11:53 AM polprog: and buffer overflows
11:53 AM twnqx: my current software reboots occasionally
11:53 AM twnqx: i should try to figure out why
11:54 AM twnqx: but is has lufa in it, and it's easy to just blame that
11:55 AM PoppaVic: polprog: yes, the overflow thing is really, really common in learning offset 0 and <N stuff
11:55 AM polprog: "learning" :v
11:55 AM PoppaVic: yup
11:56 AM PoppaVic: well, ok - I might accept also "when yer exhausted, drunk and on hour 36" - but..
11:56 AM polprog: abusing the c programming language
11:56 AM polprog: haha
11:57 AM PoppaVic: C is simple... The standards make it hard.. Also, I usually test stuff in forth as well <shrug> I am more Get 'Er DONE than "Holy Standard".
11:57 AM PoppaVic: if yer gonna' fuck with "standard" - you might as well switch to asm's as fast as possible
12:26 PM nohit: debugger isnt something that you come up with use-cases, its something that you use in your everyday life, just like a text editor, its just a tool. "hmm i wanna see if my code ever reaches this function... *click* *click* *bang* ...it does, hmm how many milliseconds it takes to reach that function *click* *click* *bang* ...hmm i wanna see what the function actually writes in to memory, lets see...*click* *click* *bang*"
12:26 PM nohit: etc.
12:39 PM Fuchikoma: When you do something, never make life hard for the next guy who might have to deal with it later
12:39 PM Fuchikoma: 'cause chances are pretty good that guy is gonna be you
08:09 PM rue_shop1: polprog, "abusing C" you gonna show that switch trick?