05:43 <tankf33der> in case of pil21 libedit cant replace readline. in general it can.

05:43 <Regenaxer> ok

05:44 <Regenaxer> Why not in pil21?

05:44 <Regenaxer> Different API?

05:44 <tankf33der> you use rl_basic_quote_characters

05:44 <tankf33der> which is not in libedit

05:44 <Regenaxer> ok

05:44 <tankf33der> libedit even provides readline.h for compatability, libedit is bsd reimplementation.

05:45 <Regenaxer> What is the advantage of libedit?

05:45 <Regenaxer> Most important is portability

05:45 <tankf33der> lets keep readline.

05:45 <Regenaxer> ok

05:50 <Regenaxer> My biggest problem with readline is that there seems to be no key-hook

05:50 <Regenaxer> So *Run is stopped in pil21 after editing started

05:57 <Regenaxer> I use rl_pre_input_hook to handle *Run

05:57 <Regenaxer> i.e. *before* input starts

05:58 <Regenaxer> Perhaps this is also an advantage, *Run does not mess with the line being edited ;)

06:02 <tankf33der> http://ix.io/2BzZ

06:02 <tankf33der> what i see here

06:03 <Regenaxer> yeah

06:03 <Regenaxer> no key hook

06:04 <Regenaxer> I used only these docs: http://tiswww.case.edu/php/chet/readline/readline.html

06:05 <Regenaxer> and for readline history http://tiswww.case.edu/php/chet/readline/history.html

06:06 <Regenaxer> I call the new 'history' function in my .pil/rc

06:15 <Regenaxer> Still can't locate the stress bug

06:15 <Regenaxer> What I know is that under current test *someone* modifies a global

06:15 <Regenaxer> it is $Avail

06:16 <Regenaxer> but when I try to find it, it vanishes (heisen)

06:17 <Regenaxer> Definitely some part *may* overwrite $Avail with a nonsense value

06:17 <Regenaxer> I don't know where and when

06:17 <Regenaxer> It goes undetected for a long time

06:17 <Regenaxer> possibly never detected, as it is overwritten later with a correct value

06:18 <Regenaxer> Grr, I'm struggling with this for days now :(

06:37 <Regenaxer> *we* are

08:49 <Regenaxer> Now running dedicated checks. Traverses whole Avail list to detect corruption. Extremely slow

08:50 <tankf33der> is patch available?

08:51 <Regenaxer> I've put this http://ix.io/2BAH into main.l (after 'dbg'

08:52 <Regenaxer> then call (chkAvail 1) or (chkAvail 2) at interesting places

08:52 <Regenaxer> Dont put into often-called places like 'cons', it will run forever

08:53 <Regenaxer> Even startup takes minutes then

08:54 <Regenaxer> Explanation: From core I believe that the Avail list is overwritten somewhere

08:54 <Regenaxer> So this check should detect that immediately

08:54 <Regenaxer> If reproducible, do a binary search with (chkAvail ..) until the bad place is found

08:55 <Regenaxer> As it happens so seldom, I believe it is sogewhere where plio is read and/or external symbols are handled

08:56 <Regenaxer> thus not in normal (startup) code

09:08 <beneroth> Regenaxer, BigBlueButton (the online teaching thingy I considered to look into for pilcon) just got on my blacklist, their security awareness / care seems to be seriously non-existent: https://blog.fefe.de/?ts=a16eb3a2

09:09 <Regenaxer> Oha

09:10 <Regenaxer> Indeed serious

09:10 <beneroth> the first thing that came to light with mic muting being implemented on server-side only was bad enough, but this one kills it

09:10 <Regenaxer> T

09:11 <beneroth> it's the equivalent of giving untrusted unchecked client data into (eval)

09:12 <beneroth> someone who is doing is, and gets it reported and does nothing about it for half a year, should be banned to write anything interfacing with networks...

09:13 <beneroth> (yeah we cannot to this, much big companies would stop working xD)

13:41 <Regenaxer> Avail tests did not work out anything

13:46 <beneroth> :/

13:46 <Regenaxer> :(

13:46 <Regenaxer> Heisen

13:46 <Regenaxer> Measuring changes the result

14:08 <tankf33der> Regenaxer: oh.

14:19 <tankf33der> my code still running under gc+

14:20 <Regenaxer> ok

14:20 <Regenaxer> I think you can stop it, it will not help

14:20 <Regenaxer> Even if it crashes, it gives no hint about the reason

20:38 <beneroth> Regenaxer, I just got a notification that Ubuntu fixed our complaint about the outdated picolisp version (the one where debian was many versions ahead of debian) xD

20:40 <beneroth> 2016-07-25 https://bugs.launchpad.net/bugs/1606142

20:43 <beneroth> over 4 years reaction time

20:43 <beneroth> ah you should also got an notification