aw- has quit [Read error: Connection reset by peer]
aw- has joined #picolisp
_whitelogger has joined #picolisp
_whitelogger has joined #picolisp
_whitelogger has joined #picolisp
_whitelogger has joined #picolisp
_whitelogger has joined #picolisp
orivej has joined #picolisp
razzy has joined #picolisp
<razzy>
hi all, for the record: i like simple understandable tests where you can analyse results, complemented with one automated bundle of everything that says "everything is OK" or "there was a problem, start making simple tests to understand"
<razzy>
also maintenance costs of said bundle are too high for picolisp imho
<Regenaxer>
Hi razzy, which bundle are you talking of?
<razzy>
one that i do not know exists in picolisp
<razzy>
Regenaxer: do tests.l cover everything in picolisp?
<Regenaxer>
you "do not know exists"?
<Regenaxer>
what does that mean?
<Regenaxer>
Anyway. How can a test cover everything?
<Regenaxer>
What *is* everything?
<razzy>
everything-test could cover all basics building blocks or most used functions in software.
<Regenaxer>
No fuzzy tests there. Please implement!
<razzy>
do you mean it?
<Regenaxer>
sure
<Regenaxer>
Don't talk, do!
<Regenaxer>
a good exercise
<Regenaxer>
You could contribute something useful instead of just complaining and demanding all the time
<razzy>
:D T
razzy has quit [Quit: ERC (IRC client for Emacs 26.1)]
rob_w has joined #picolisp
razzy has joined #picolisp
orivej has quit [Ping timeout: 272 seconds]
osmano807 has joined #picolisp
<osmano807>
Hello, can someone help debug why the internal http server doesn't start even with the sample programs?
<osmano807>
It doesn't show any errors, but the server doesn't start (no tcp port open, but the picolisp shell starts)
<Regenaxer>
Hi osmano807
<Regenaxer>
You mean the (server ..) function, or the httpGate utility?
<Regenaxer>
and what did you start, and how?
<osmano807>
Nevermind. The problem was with my distro (NixOS) and wrong packaging (It wasn't passing shell arguments to the picolisp binary)
<Regenaxer>
All right :)
osmano807 has quit [Ping timeout: 256 seconds]
orivej has joined #picolisp
<tankf33der>
picolisp build and pass tests on voidlinux. ok.
<Regenaxer>
Thanks
<tankf33der>
and pil32 via multilib mode.
<Regenaxer>
ok
freemint has joined #picolisp
<freemint>
razzy how would you go about implementing fuzzing? what would you like to fuzz?
<razzy>
freemint: you put randomized inputs to every basic function. record outputs. use pairs input-output as benchmark. in future use same inputs. if outputs change, return failure
<freemint>
the same randomized inputs for everyone? It's not fuzzing then.
<razzy>
freemint: i mean multiple randomised inputs->output per function. it count as fuzzing imho
<razzy>
did i get my idea across?
<freemint>
so you got different inputs each time?
<razzy>
freemint: in first batch you have different inputs for one tested function.
<razzy>
in seconds and after you have same batch of different inputs so you can run diff on them
<razzy>
you can run diff between first which you know is OK. and everyafter which might be changed
<freemint>
ok i would go the "compare equivalent implentations" route with the other implementations being more audible
<freemint>
razzy so you can only catch new errors?
<freemint>
afk?
<razzy>
freemint: yes, new errors and hardbugs like segfault.
<razzy>
freemint: also, if you have different versions of same function, you can compare those
<razzy>
but ussually you do not program everything 3 times
<razzy>
freemint: it is very usefull, if you do some core changes, and want to be reasonably sure that everything else works ok.
<razzy>
freemint: did i explain my idea well?
orivej has quit [Ping timeout: 272 seconds]
lodsw has quit [Read error: Connection reset by peer]
<razzy>
btw, i think something(propably GC) restructuralise adresses of cells. adresses seem different each time i look. is it true?
lodsw has joined #picolisp
<Regenaxer>
The address of a cell never changes
<freemint>
razzy the addresses of a list does not change but (== (1 2) (1 2)) are not pointer equal because they are technically not the same list
<Regenaxer>
only a new cell may appear in that place
<Regenaxer>
freemint, right
<Regenaxer>
the reader creates two lists
<freemint>
razzy so your idea was vaguealy explained and your english was a little lacking, like my english gets bad when i am tired. Because of the vagueness i had to ask and i am not sure that your approach is the best you could take but since you implement it you can do your own thing
<freemint>
I would go with multiple equivalent implementations
<freemint>
but i would not include it in the default unit test and write a seperate fuzz.l instead which generate reports and you can leave running for some time and it will try to find errors.
<freemint>
razzy my approach is huge undertaking but it will end up generating a powerfull documentation
<razzy>
your aproach? to fuzzying?
<Regenaxer>
razzy: Counterexample for what?
<Regenaxer>
I see
<Regenaxer>
hmm, you still did not grasp Lisp
<Regenaxer>
(adr '(cdr a)) is *read* newly each time
<Regenaxer>
(adr (cdr a)) is the CDR
<Regenaxer>
As I said, an existing cell is *never* moved
<Regenaxer>
this would create havoc, as it might be referred to from many places
mtsd has joined #picolisp
<freemint>
Oh Regenaxer i did not see that either
<razzy>
Regenaxer: thx, i will meditate on it
<Regenaxer>
Good :)
<mtsd>
Good evening :)
<Regenaxer>
Good evening mtsd!
<mtsd>
Hello Regenaxer!
<Regenaxer>
:)
<mtsd>
Last few days of 2018 :)
<Regenaxer>
yeah
<mtsd>
We are going to celebrate new years eve with some friends. They have small children too, so it will be quite peaceful
<mtsd>
Well, as peaceful as it can get, with many small children around ;)
<Regenaxer>
Same here (though without small children)
<Regenaxer>
They come here, and we make Sushi-Party :)
<mtsd>
Sounds great! I really like sushi :)
<Regenaxer>
:)
<mtsd>
Tried making some at home once
<freemint>
My '+ idea
<freemint>
(de add (A B) (+ (x| (A B)) (>> -1 (& (A B))))) (de addlist @ (if (cdr (rest)) (add (car (rest)) (apply addlist (cdr (rest)))) (car (rest))))
<Regenaxer>
'A' is a function?
<Regenaxer>
(+ (x| (A B))
<freemint>
sorry misstype
<freemint>
(de add (A B) (+ (x| A B) (>> -1 (& A B))))
<Regenaxer>
What kind of addition is this?
<freemint>
Addition by binary shifting how it is done in hardware
<Regenaxer>
(if (cdr (rest)) can be done much simpler
<Regenaxer>
(rest) is expensive
<freemint>
I was curious if there was an better way
<Regenaxer>
you iterate with (next)
<Regenaxer>
and a loop is much better than recursion
<Regenaxer>
I did not completely understand what you do
<freemint>
(de addl @ (let N 0 (loop (NIL (rest)) (setq N (add (next) N))) N )) but i think i could make it more beautifull
<Regenaxer>
No, not (rest)
<Regenaxer>
it builds a new list on each call
<Regenaxer>
(args) is better
<Regenaxer>
checks *if* there are more
<Regenaxer>
(de addl @ (let N 0 (while (next) (setq N (add @ N))) N]
<Regenaxer>
or
<Regenaxer>
(de addl @ (let N 0 (while (args) (setq N (add (next) N))) N]
<freemint>
what does (setq N (add @ N)) do?
<Regenaxer>
'@' is bound to the result of (while (next)
<freemint>
ahh
<Regenaxer>
(doc '@)
<Regenaxer>
or better picolisp/doc/ref.h
<Regenaxer>
oops
<Regenaxer>
picolisp/doc/ref.h#atres
<Regenaxer>
@ Result
<Regenaxer>
-> Flow functions
<freemint>
why won't you make it picolisp/doc/ref.h#aters ;)
<Regenaxer>
the path depends on the installation
<Regenaxer>
Here it is e.g. file:///data/data/com.termux/files/usr/lib/picolisp/doc/ref.html#atres
<freemint>
you did not catch the joke?
<Regenaxer>
oh, ups
<Regenaxer>
Sitting here with family after a few beers ;)
<Regenaxer>
But haters is too harsh
<freemint>
:)
<freemint>
is there a cheap? get the position of the most significant bit which is not 0?
<freemint>
i do not like that my add still uses +
<Regenaxer>
Unfortunately not
<freemint>
is there a chop for numbers?
<Regenaxer>
yes, the normal 'chop' chops the decimal digits
<freemint>
(chopn 9) (1 0 0 1)
<freemint>
to binary numbers
<Regenaxer>
(chop (bin 9))
<Regenaxer>
but expensive
<Regenaxer>
as 'bin' is expensive
<Regenaxer>
better do the equivalent directly
<freemint>
yeah i do not like that string conversion
<Regenaxer>
without building a symbol first
<freemint>
but nothing built in?
<Regenaxer>
exactly
<Regenaxer>
to find the highest bit?
<freemint>
mhh i had that pop up in an actual program.
<Regenaxer>
dup dup not and =
<Regenaxer>
this is the Forth solution iirc
<Regenaxer>
We should try the same in Lisp
<Regenaxer>
hmm, rather not
<freemint>
I had a use case for bin list to num recently
<Regenaxer>
So modify the 'bin' function
<Regenaxer>
can be easier, without sign
<freemint>
i implemented by own adhoc solution back then (list (/ (& 8 HW) 8) (/ (& 4 HW) 4) (/ (& 2 HW) 2) (& HW 1)) it was not very perfomant but it was only need when creating a few onjects on startup
<Regenaxer>
Shifts are probably better than divisions
<Regenaxer>
And a loop makes it more flexible for the size
<freemint>
but i do not know where to start loop i only know when to stop
<Regenaxer>
like in 'bin'?
<freemint>
I have not understood how bin solved that problem
<Regenaxer>
in the first 'cond' clause
<Regenaxer>
the sign stuff can be ignored
<Regenaxer>
'S'
<freemint>
Can you explain the L and A part
<freemint>
(or let me study a little)
<Regenaxer>
(let (L (& 1 X) A (cons 0 I))
<Regenaxer>
(until (=0 (setq X (>> 1 X)))
<Regenaxer>
(push 'L (& 1 X)) ) )
<Regenaxer>
That's the minimum
<Regenaxer>
giving a list of 0 and 1
<Regenaxer>
So you could also do something else
<Regenaxer>
instead of building a list
<freemint>
Why isn't the list in the wrong order?
<Regenaxer>
because of 'push'
<freemint>
ahh
<Regenaxer>
starts from lowest bit
<freemint>
of course
<Regenaxer>
'A' is also not needed
<freemint>
A is for chopping into groups of blocks
<Regenaxer>
right
<Regenaxer>
(at A
<freemint>
so much alien technology
<Regenaxer>
Which part is alien?
<freemint>
'at it is self modifying code.
<Regenaxer>
In this case not
<Regenaxer>
it modifies the value of the variable A
<Regenaxer>
(loop (msg (inc (0)))) *might* be called self modifying code, but I think it is *not*
<Regenaxer>
the (0) is data, not code ;)
<Regenaxer>
a kind of artificial distinction in picoLisp
<freemint>
(de H () (at '(0 . 3) (prinl "bumm"))
<Regenaxer>
same
<Regenaxer>
(0 . 3) is *data*
<freemint>
self modifying data?
<Regenaxer>
the quote is not needed BTW
<Regenaxer>
no, just normally modified data
<razzy>
um, how do i save anonymous symbol to file, to reload it later by (read)? is there special function?
<razzy>
or do i need to print everything about symbol, save it, and reaload with lots of pain
alexshendi has joined #picolisp
mtsd has quit [Quit: WeeChat 1.6]
orivej has joined #picolisp
<Regenaxer>
razzy, yes, you need to save the value and the property list
shpx has joined #picolisp
<Regenaxer>
or better use a database
<Regenaxer>
Not much pain
<Regenaxer>
(println (val Sym) (getl Sym))
<Regenaxer>
(set Sym (read)) (putl Sym (read))
<Regenaxer>
if no DB
<Regenaxer>
But anyway, a DB is exactly for that
alexshendi has quit [Read error: Connection reset by peer]