08:53 <Nistur> mornin'

10:55 <Regenaxer> tankf33der here?

11:03 <tankf33der> here

11:04 <Regenaxer> moin!

11:05 <Regenaxer> What do I need to install to compile pil32 on an x86-64 system?

11:05 <tankf33der> hm

11:05 <tankf33der> i dont know

11:05 <Regenaxer> Must do some test, and have no more 32 bit system

11:05 <tankf33der> what distro?

11:05 <Regenaxer> Debian

11:06 <tankf33der> didnt do in years

11:06 <Regenaxer> :)

11:06 <tankf33der> i can try in hour.

11:06 <tankf33der> doesnt work out of box?

11:06 <Regenaxer> If it is ok for you

11:06 <Regenaxer> no

11:07 <Regenaxer> complains about includes

11:07 <tankf33der> archlinux and centos both x64 works both!

11:07 <Regenaxer> I see

11:07 <Regenaxer> OK, then don't worry, I ask Mr. Google

11:07 <tankf33der> paste error log

11:08 <Regenaxer> from main.c:5:

11:09 <Regenaxer> I think I just need to install one or two packages

11:09 <Regenaxer> Was too lazy to search around and thought you have it easily

11:09 <Regenaxer> no worry!

11:12 <Regenaxer> I try later, first make the pil64 and ersatz versions

11:12 <Regenaxer> But in the long range it would be good if I can test pil32 again

11:19 <tankf33der> you can, i belive

11:19 <Regenaxer> yeah

11:19 <Regenaxer> np

13:15 <Regenaxer> tankf33der: Just doing "apt install g++-multilib libc6-dev-i386" was enough

13:16 <tankf33der> yea

13:16 <tankf33der> as expected.

13:16 <Regenaxer> :)

17:41 <razzy> hi, i think i have bug in my version of picolisp. i write some code, when i (load) it, it changes. weird. https://ptpb.pw/_Nix/text any ideas pls?

17:44 <Regenaxer> Hi razzy. I can't see a difference at a first glance

17:44 <Regenaxer> ah, a comment

17:44 <Regenaxer> sorry

17:45 <Regenaxer> What else is running? Something modified it as a side effect (destructive function)

17:46 <Regenaxer> I think I know

17:46 <Regenaxer> This call is wrong: (setq Fld ( Dir Fld ) (n0 N))))

17:47 <Regenaxer> Needs (setq var 'val var 'val ...

17:47 <Regenaxer> So (n0 N) is taken as a "var" (settable place)

17:48 <Regenaxer> and its CAR is replaced with the next arg (which is not present and thus NIL)

17:48 <Regenaxer> So the code is modified the first time you run it

17:48 <Regenaxer> :)

17:49 <Regenaxer> So it is no bug in PicoLisp but in your code

17:51 <razzy> aaaah, thank you :] i feel safer :] i will meditate on it

17:51 <Regenaxer> great :)

17:51 <Regenaxer> Sometimes such self-modifying code is in fact useful

17:52 <razzy> i like self modifying code. if it modifies as i want :]

17:52 <Regenaxer> true

17:52 <Regenaxer> I often do such things:

17:52 <Regenaxer> : (task -1000 0 (msg (inc (0))))

17:52 <Regenaxer> -> (-1000 0 (msg (inc (0))))

17:52 <Regenaxer> 3

17:52 <Regenaxer> 2

17:52 <Regenaxer> : 1

17:53 <Regenaxer> 4

17:53 <Regenaxer> ...

17:53 <beneroth> hi Regenaxer :)

17:53 <beneroth> heya razzy

17:53 <Regenaxer> Hi beneroth! Welcome back!

17:53 <beneroth> thanks!

17:54 <razzy> i am happy i do not need to reinstall picolisp or OS :]

17:55 <beneroth> in 95% (guesstimate) of issues with picolisp, it's a BCK-cause (between chair and keyboard)

17:56 <razzy> but looking for it from that spot is sometimes near impossible

17:56 <beneroth> well that issue is caused by too low humbleness value in the developer :P

17:58 <beneroth> bad probability thinking too. for non-unusual use cases the probability is pretty low to find a bug in a system which is over 30 years old and in productive use in real companies for all this time

18:01 <razzy> i should of said, "i checked everything i could thinked off, i have only painfull options left, please help me"

18:02 <Regenaxer> No, this was a valid question

18:03 <Regenaxer> I think 'lint' would have complained

18:04 <Regenaxer> hmm, no, it won't

18:04 <beneroth> the question as completely valid. I just would have said "I think I have a bug in my picolisp code. I write ..." instead of "I think I have a bug in my version of picolisp" ;-)

18:04 <Regenaxer> A list is a legal 'var'

18:04 <razzy> there are optical illusions, mathematical illusions, that are near impossible to look through. and in programming there are illusions too. and only practical defence is to have other person look at it (not that this was the case)

18:04 <beneroth> Regenaxer, yeah, there are quite some use cases for a list var :)

18:05 <Regenaxer> T

18:05 <beneroth> razzy, T

18:05 <razzy> or taking heavy drugs

18:06 <razzy> which is not practical

18:06 <beneroth> you made that experience? most people don't have better logical thinking while on drugs, I believe.

18:07 <beneroth> well of course there is the Ballmer Peak

18:07 <beneroth> https://www.xkcd.com/323/

18:07 <razzy> important thing is you have different logical thinking. which make it easier to find blind spots. LSD experience :]. not very practical :]

18:11 <beneroth> drugs are just a (risky) short-cut. there are many ways to achieve different logical thinking and to change ones mind.

18:12 <razzy> T

19:53 <beneroth> fun story (the stories of this authors are just epic): https://www.reddit.com/r/talesfromtechsupport/comments/4yhmaw/the_four_second_rule/

19:53 <beneroth> "Let me tell how web works. Everything with backend takes at least four seconds. One second for browser to talk to server, one second for server to talk to backend, one for backend to send response to server, and one for server to send response to browser. Four seconds, no faster"

19:54 <beneroth> for the people not working with business IT: yes, that's about the state of the industry...

20:00 <Regenaxer> They have not tried pil yet ... :)

20:02 <beneroth> that I think whenever I read about other database products, especially the new revolutionary ones...

20:02 <beneroth> on the other hand, competition being ignorant of pilDB might be our fortune :)

20:05 <beneroth> Regenaxer, nearly all graph databases out there are In-Memory-databases, sometimes without any disk storage, sometimes with traditional relational DB as storage, a few with a custom storage

20:07 <beneroth> and graph seems to usually mean you cannot query by index, but by graph-path. seems to be an unnecessary restriction for purities sake. pilDB can do both pretty well.

20:09 <beneroth> razzy, HP is working on this (the hardware HP, not HPE which becomes a pure service/consulting company). they're researching a new form of RAM which ought to be persistent, they hope to bring it to market within the next few years

20:10 <razzy> there should be software layer above physical media managing multiple plug and play devices

20:10 <beneroth> https://en.wikipedia.org/wiki/Memristor

20:11 <razzy> as a programmer you should not care about memory or access time :]

20:11 <beneroth> razzy, you can have such layers, but then the application cannot longer chose to use fast but not persistent memory (RAM) or slow but persistent memory (disk)

20:12 <beneroth> as a programmer you should care about performance observable by the user (mainly responsiveness)

20:13 <beneroth> https://stackoverflow.com/questions/164175/what-is-considered-a-good-response-time-for-a-dynamic-personalized-web-applicat

20:13 <razzy> my point is software should be better at choosing than programmers. but it is for longtalk

20:15 <beneroth> I agree in principle, but my point is: software can only do it insofar it has enough knowledge about the specific use case to make meaningful decisions. and usually the application does not have this knowledge, while the programmer ought to have it.

20:17 <razzy> as a programmer you should specify only when you want it done. let software optimize computing and data access through the system

20:19 <razzy> and i think it is possible :] but longtalk.

20:27 <beneroth> sure. this was tried a lot. gives some result, especially for giving a lower limit of productivity for bad programmers. an example where it after decades worked out well is compilers - they do optimization, and in the usual cases better than a programmer. but some stuff is still written in ASM. a capable programmer doing manual optimization will in nearly all cases produce better results than a preprogrammed generic general optimization heuristic.

20:29 <beneroth> also, you aware that you just said "more control to the software than the programmer, at least as long as the programmer says nothing else" in a community having "more control to programmers" as their main principles, going so far as stating "sometimes programmer should be forced to do control, so programmer feels the pain when something is stupid?" :)

20:30 <beneroth> https://picolisp.com/wiki/?pros-and-cons

20:32 <beneroth> https://software-lab.de/doc/faq.html#who

20:39 <razzy> beneroth: well, i am aware. and i think those two only seem mutually exclusive :] it is for longtalk. pls forget i said something :]

20:46 <beneroth> I fully agree that they are not mutually exclusive. actually, the current core of my work with picolisp is the goal to put more automatism into database work.

20:50 <razzy> i am thinking of chess board evaluation function :]

20:52 <beneroth> oh well, agreed, there the whole goal/achievement is to get as much as possible into the automatism.

20:53 <beneroth> but this is also a topic rather far away from the general "should this stored in memory or on disk" topic it seemed to be about at the beginning :3

20:55 <beneroth> btw. for your use case the pilDB might be very useful, if you might want to store some stuff for longer within the db. everything fetched from pilDB stays in memory (cached) by default, so its not much overhead compared to working purely with memory variables.

20:56 <beneroth> though you have to tell it when it should write stuff to disk (commit 'upd), but you don't have to tell it when it needs to load it from disk - everything gets loaded automatically on first access.

21:01 <beneroth> good night, bbl

21:06 <razzy> good night

21:15 <razzy> i do not know how to work with pilDB yet.

21:53 <razzy> btw, in general i like clearly marked destructive functions, and clearly marked self-modifying code. setq seem ok, but the posibility of other magical functions scares me.