beneroth changed the topic of #picolisp to: PicoLisp language | The scalpel of software development | Channel Log: https://libera.irclog.whitequark.org/picolisp | Check www.picolisp.com for more information
razzy has quit [Ping timeout: 260 seconds]
razzy has joined #picolisp
razzy has quit [Ping timeout: 260 seconds]
razzy has joined #picolisp
rob_w has joined #picolisp
razzy has quit [Ping timeout: 268 seconds]
razzy has joined #picolisp
razzy has quit [Ping timeout: 260 seconds]
razzy has joined #picolisp
razzy has quit [Ping timeout: 260 seconds]
razzy has joined #picolisp
beneroth has joined #picolisp
razzy has quit [Ping timeout: 260 seconds]
razzy has joined #picolisp
<Regenaxer> tankf33der here?
razzy has quit [Ping timeout: 268 seconds]
razzy has joined #picolisp
<tankf33der> Here
<Regenaxer> Do you remember when you optimized SSL_CTX_set_options() in httpGate and ssl?
<Regenaxer> Hi, btw :)
<Regenaxer> ssl gives problems in PilBox
<Regenaxer> tls_process_server_certificate:certificate verify failed
<Regenaxer> The certs are ok, so do you have an idea?
<Regenaxer> It *might* be related to Android 11
<Regenaxer> the exactly same 'ssl' binary works in Termux
<Regenaxer> but fails on *some* sites in PilBox
<tankf33der> i remember nothing.
<Regenaxer> oh :)
<tankf33der> what sites fails?
<Regenaxer> you configured SSL_OP_NO_SSLv2 | SSL_OP_NO_SSLv3 | ...
<tankf33der> options above should not be a problem
<Regenaxer> picolisp.com, heise.de
<tankf33der> options above should not be a problem at all.
<Regenaxer> software-lab.de is ok
<Regenaxer> The certs are all ok
<Regenaxer> I checked with other tools
aw- has quit [Quit: Leaving.]
razzy has quit [Ping timeout: 260 seconds]
razzy has joined #picolisp
<tankf33der> android 11 is latest version, right?
<Regenaxer> yeah
<Regenaxer> And Termux is still built for an older version
<tankf33der> does pilbox uses android routines or API?
<Regenaxer> No idea
<Regenaxer> for example
<Regenaxer> (ssl "picolisp.com" "wiki" (line T))
<Regenaxer> this works in Termux or Debian
<Regenaxer> But in PilBox REPL you get a cert error
<Regenaxer> it is the same @bin/ssl binary
<Regenaxer> Arm64 and openssl libs
<tankf33der> sounds like android 11 added some restrictions you dont handle.
<Regenaxer> yes, looks like
<Regenaxer> but other sites work fine
<Regenaxer> (ssl "www.heise.de" "index.html" (line T))
<Regenaxer> So it has to do with cert checking
<Regenaxer> something more restrictive
<Regenaxer> Or it is something in my OnePlus 6Y
<Regenaxer> 6T
<tankf33der> i know what could be a problem
<tankf33der> This is chain of fails site
<tankf33der> one of cert expired.
<tankf33der> Next:
<Regenaxer> hmm
<tankf33der> This is chain for software-lab and it is fine
<tankf33der> the same for heise.de
<tankf33der> One of cert expired.
<Regenaxer> Strange, I just use Let's Encrypt
<tankf33der> Ssllabs.com site is best tool for sites comparison
<Regenaxer> ok
<Regenaxer> this gives no error as far as I can see
<tankf33der> Me too, but expand chain lists
<Regenaxer> ?
<Regenaxer> And why does Chrome work?
<Regenaxer> on the save device
<tankf33der> Expand this list by + clicking
<tankf33der> Maybe chrome could handle above.
<tankf33der> iphone could connect to picolisp.com
<tankf33der> latest ios.
<Regenaxer> but not with PilBox
<Regenaxer> Only PilBox gives a problem
<Regenaxer> i.e. bin/ssl
<tankf33der> Different openssl library?
<Regenaxer> no, same files
<Regenaxer> copied to PilBox
<tankf33der> this article about expired DST cert on September 2021
<tankf33der> is it on your side or your provider?
<Regenaxer> the cert?
<tankf33der> check pictures above
<Regenaxer> The article seems to the point
<Regenaxer> though I don't understand fully
<Regenaxer> Just wait?
<Regenaxer> "What should you do? For most people, nothing at all!"
<tankf33der> can you try to open monocypher.org site from pilbox
<tankf33der> should fail
<tankf33der> should fail as picolisp.com
<Regenaxer> yep. Same
<tankf33der> it have the same DST expired cert in chain
<Regenaxer> So this is a temporary problem with letsEncrypt?
<tankf33der> unknown.
<Regenaxer> I read the article more thoroughly
<Regenaxer> Not really clear
<Regenaxer> I will call ssl with negative port
<Regenaxer> Then ignores the error
<Regenaxer> Not sure if it will fix. My OnePlus got an OS update a few weeks ago
<Regenaxer> and still gets further
<Regenaxer> No idea why it does not work
<tankf33der> chrome works because it can handle when one of root cert expired, ssl library has a trust store
<Regenaxer> I see
<tankf33der> your bin/ssl is too raw and cant handle broken chains
<Regenaxer> that's ok
<tankf33der> this is it how i see it now.
<Regenaxer> It is used for dedicated accesses
<Regenaxer> yeah, understand
<tankf33der> This checker and chrome works because they even do not see expired cert in the chain.
<Regenaxer> yeah, like the two checkers above which I use
<Regenaxer> ok, let's not worry
<Regenaxer> I use -443 for now
<Regenaxer> I use it also to access my self-signed certs on some servers
<tankf33der> Good
<tankf33der> afk.
<Regenaxer> There is further authentication on the app level
<Regenaxer> thanks a lot!!!
<Regenaxer> Made it clear
<beneroth> some android/Google stuff does allowed invalid parent certs, I heard
<Regenaxer> It is all a mess ;)
<beneroth> yeah
<beneroth> and german state crypto is also a special kind of rabbit whole...here I read that about android...
<Regenaxer> uh, the health system
<Regenaxer> gematik
<beneroth> Seems to me that german state-run "digitalization projects" are even worse than Switzerlands.. here we have corruption and never finished/usable projects. but the creation of new bad pseudo-standards is mostly not happening (they try, everyone ignores it)
<Regenaxer> sigh
<beneroth> did you change bin/ssl in pil21? I guess there were updates?
<Regenaxer> yes, I think it changed meanwhile
<Regenaxer> 14apr21
rob_w has quit [Quit: Leaving]
<Regenaxer> Nothing critical though
<beneroth> ok, thanks for the info
razzy has quit [Ping timeout: 260 seconds]
razzy has joined #picolisp
<razzy> at least switzerland and germany still looks like democracies
<razzy> I am testing debugger, i think i can have contribution on PilCon.
<Regenaxer> Coool!
razzy has quit [Ping timeout: 268 seconds]
razzy has joined #picolisp
razzy has quit [Ping timeout: 260 seconds]
razzy has joined #picolisp
razzy has quit [Ping timeout: 268 seconds]
razzy has joined #picolisp
<tankf33der> One mode good explaination
<tankf33der> mode -> more
<Regenaxer> Ah, nice. Thx!
<razzy> I want to debug picolisp core system that is written in picolisp. i have learned that it is not possible now. I want next best thing, i do not know what it is.
<razzy> Is there good reason to not have possibility to compile picolisp and have self-hosting core?
<Regenaxer> PicoLisp is not compilable to have the interpreter more powerful and faster
<razzy> Regenaxer: could you explain more powerfull?
<Regenaxer> The language can express things not doable if the code is going to re compiled. For examqle FEXPRs, and dynamically bound symbols are not efficient
<Regenaxer> In Pil, each built-in function interprets its arguments individually
<Regenaxer> it decides for each arg what to do
<Regenaxer> A compiler cannot know this
<razzy> I want self-hosted interpretter|core and "normally" interpretted picolisp. i can avoid using powerfull features in compilig core|interpretter.
<Regenaxer> What is a "self-hosted" interpreter?
<razzy> Regenaxer: if I compile C compiler in older C compiler, it is self-hosted
<Regenaxer> So is PicoLisp. It compiles itself.
<razzy> Regenaxer: can i run ./pil ./src/* + and debug what picolisp interpretter is doing?
<Regenaxer> yes
<Regenaxer> Look into Makefile
<Regenaxer> src/*.l is all executed
<Regenaxer> $(PIL) lib/llvm.l main.l -bye > base.ll
<Regenaxer> so pil loads lib/llvm.l and yhen loads main.l
<razzy> omg, I want to debug it sooo much
<Regenaxer> that is, main.l *is* executed
<razzy> so in other words, I can interpret picolisp interpretter and in this interpretted interpretter i could interprett my code. yes? and I could debug my code and underlaying picolisp interpretter. uff :]
<Regenaxer> uh :)
<razzy> yes? no?
<Regenaxer> yes, sure, everything interprets everything ;)
<razzy> you are just teasing me
<Regenaxer> No, the point is that it consists of different layers
<Regenaxer> src/lib/llvm.l redefines PicoLisp functions
<Regenaxer> so that they generate low-level code instead of maniplating cell structures
<Regenaxer> The low-level code is the interpreter
<Regenaxer> that interpreter interprets cell structures
<razzy> now i feel problems, redefining gives me head ache
<Regenaxer> It redefines by building new symbols in the 'llvm' namespace
<Regenaxer> (symbols 'llvm 'pico)
<Regenaxer> For example, 'de' is redefined:
<Regenaxer> (local) de
<Regenaxer> (pico~de de Prg ...
<Regenaxer> i.e. the pico 'de' is used to define 'de' in llvm namespace
<Regenaxer> So in the followin, executing (de _foo will compile the code for _foi
<Regenaxer> *_foo
<razzy> I think i cannot have my-new-pil with my-new-code interpretted in old-pil. it would be convenient to just compile my-new-pil into machine code if i like it
<Regenaxer> Why is it not possible?
<razzy> I am looking at Makefile and i do not know how :D. maybe PilCon?
<Regenaxer> yes, let's talk about it
<razzy> Thank you.
<Regenaxer> :)
<Regenaxer> I'm looking forward to next PilCon
<razzy> me too :]
<razzy> for record i also think that speed argument is not obvious.
<Regenaxer> speed for an interpreter - compiled code is of course faster
<razzy> speed of the interpretter would depend on programmer, self hosted interpretters tends to be bigger and slower, because it is easier to write them. in my unexperienced opinion :]
<Regenaxer> I would agree with these statements
mario-goulart has quit [Remote host closed the connection]
mario-goulart has joined #picolisp
beneroth has quit [Quit: Leaving]
aw- has joined #picolisp