orivej has quit [Ping timeout: 256 seconds]
m_mans has joined #picolisp
<
tankf33der>
m_mans: ку
<
Regenaxer>
Hi m_mans, tankf33der!
<
m_mans>
Hi Regenaxer! How are you?
<
Regenaxer>
Good :) and you?
<
m_mans>
Me too, thanks
<
m_mans>
What about pil21 state? Do you plan to switch to it in your production environments?
<
Regenaxer>
yes, though currently only the wiki runs on pil21
<
Regenaxer>
I also asked the Debian maintainer to switch to pil21
<
m_mans>
Sadly I not visit chat too often )
<
Regenaxer>
Too busy?
<
m_mans>
I think I'm not organized :)
<
Regenaxer>
hehe :)
<
m_mans>
I'm trying to start new project and also want to try pil21
<
Regenaxer>
I think it is quite stable, but I'm glad for any bug report
<
tankf33der>
m_mans: btw, you should try setup your old code to run under pil21.
<
m_mans>
tankf33der: yeah, maybe. But currently I have no active projects
<
m_mans>
Regenaxer: I found strange behavior in REPL when I press Ctrl+C
<
Regenaxer>
it is now readline(), so it changed a little
<
Regenaxer>
though Ctrl+C should be basically the same
<
Regenaxer>
ie. interrupt
<
m_mans>
./pil + , then I press Ctrl-C, I see '!' - it's OK. Then I press "Enter" and see empty string
<
m_mans>
It seems that it hangs, no reaction to futher "Enter"s
<
Regenaxer>
that is strange
<
Regenaxer>
You need to press ^D to leave the breakpoint
<
Regenaxer>
no longer ^X
<
Regenaxer>
but you are right, just Enter behaves strangely
<
Regenaxer>
It is hard to get readline to interoperate properly
<
Regenaxer>
I will check what happens with Enter
<
m_mans>
another bug: 1) start pil, 2) evaluate (+ 1 1), 3) press Ctrl-C -> free(): double free detected in tcache 2, Aborted (core dumped)
<
m_mans>
Fresh Linux Mint (x64)
<
Regenaxer>
pil21 ?
<
Regenaxer>
Seems I cannot reproduce. Is (+ 1 1) relevant?
<
m_mans>
pil21, yes. I think you can enter any expression and press Enter. Then press Ctrl-C
<
Regenaxer>
Without evaluating the expression?
<
Regenaxer>
i.e. before <enter>?
<
m_mans>
No, you need to press Enter first
<
m_mans>
any expression, even 1 <Enter>
<
Regenaxer>
Like that?
<
Regenaxer>
./pil +
<
Regenaxer>
: (+ 1 1)
<
Regenaxer>
I pressed ^C
<
m_mans>
yeah, here it goes to core dump
<
Regenaxer>
before printing NIL and : ?
<
m_mans>
free(): double free detected in tcache 2
<
m_mans>
Aborted (core dumped)
<
Regenaxer>
Perhaps it is related to the other problem
<
Regenaxer>
hanging on <enter> after !
<
Regenaxer>
I must rething this whole part
<
Regenaxer>
what to do in readline on interrupt
<
m_mans>
No urgency, just for info
<
tankf33der>
double free crash in my list, known.
<
m_mans>
It can look strange for newbies
<
Regenaxer>
^Z seems to behave well
<
Regenaxer>
But I also noticed that sometimes after exiting pil21 the tty is not set back to cooked
<
Regenaxer>
needs an stty sane then
<
Regenaxer>
This whole issue is a bit tricky
<
Regenaxer>
I think all those problems are related to rlSigBeg() and rlSigEnd() in @src/lib.c
orivej has joined #picolisp
<
aw->
i also get the double free error
<
Regenaxer>
Not sure what it depends on
<
Regenaxer>
I think I saw it too, but cannot reproduce
<
Regenaxer>
I first check why it hangs, this is what I can reproduce
<
Regenaxer>
In any case, hitting ^D (instead of <enter>) always works, right?
<
Regenaxer>
tracing ...
<
m_mans>
hitting ^D - seems yes
<
m_mans>
afk for a while
<
Regenaxer>
Thanks! I keep debugging :)
TeddyDD has quit [Ping timeout: 246 seconds]
<
Regenaxer>
Seems it has to do with re-entrance into 'waitFd'
TeddyDD has joined #picolisp
orivej has quit [Ping timeout: 264 seconds]
DKordic has quit [Remote host closed the connection]
rick42_ has joined #picolisp
orivej has joined #picolisp
<
tankf33der>
this is new sources, i have *.ll files and Makefile anyway tries call ../bin/pil to generate *.ll files, so bootstrapping fails.
<
Regenaxer>
This must be a time stamp error then
<
Regenaxer>
The sources are older than *.ll
<
Regenaxer>
is that the case?
<
Regenaxer>
Anyway, I probably released in the wrong moment
<
Regenaxer>
Let me release again
<
Regenaxer>
I released
<
Regenaxer>
Can you try again?
jibanes has quit [Ping timeout: 265 seconds]
jibanes has joined #picolisp
<
tankf33der>
doing.
<
tankf33der>
the same. since this is only for pijul control version lets skip it for now.
<
Regenaxer>
Strange
<
Regenaxer>
*is* it the case that the *.ll files are older?
<
Regenaxer>
It is the basic principle of make
<
Regenaxer>
did you directly unpack the TGZ? It contains the timestamp
<
Regenaxer>
Git destroys the timestamps
<
Regenaxer>
that's why I hate it
<
tankf33der>
this is pijul, like git or darcs or anything else.
<
Regenaxer>
So it destroys the file meta infos?
<
tankf33der>
maybe. checking.
<
Regenaxer>
must ae
<
Regenaxer>
'make' fails then
<
Regenaxer>
If the time stamps are destroyed, you need a 'touch *.ll'
m_mans has quit [Ping timeout: 240 seconds]
<
Regenaxer>
I could not solve the problem of re-entrance into 'waitFd'. Needs a different internal strategy
<
Regenaxer>
Will think with a fresh brain tomorrow morning.
viaken has quit [Quit: reboot]
viaken has joined #picolisp