<Regenaxer>
Man page of select() states differently
<Regenaxer>
They changed the man page too?
<tankf33der>
libc and musl different implementations.
<tankf33der>
unknown.
<Regenaxer>
Seems so
<Regenaxer>
Very bad. It was such a good feature!!
<Regenaxer>
Saved one system call to obtain the time
<Regenaxer>
On Debian the man page still says:
<Regenaxer>
"On Linux, select() modifies timeout to reflect the amount of time not slept;"
<Regenaxer>
Is this deleted from your man page?
mtsd has quit [Quit: Leaving]
<tankf33der>
Regenaxer:
<tankf33der>
On Linux, select() modifies timeout to reflect the amount of time not slept;
<tankf33der>
i see this is manual.
<tankf33der>
strange it behaves as mulsc
<Regenaxer>
Yeah
orivej has quit [Ping timeout: 260 seconds]
<Regenaxer>
I think this is a very bad decision. It breaks existing programs who rely on the man page!
<Regenaxer>
Fortunately pil21 does not depend on that
<Regenaxer>
it uses now poll() instead of select()
<beneroth>
ah breaking changes for job security of code monkeys?
<beneroth>
tankf33der, what minimal length would you recommend for blake2 file hashing? primarily to identify files, not for full tamper resistance
<beneroth>
hash length, i mean
<tankf33der>
beneroth: you must start from 32 bytes.
<tankf33der>
Anything below 32 is discouraged when using Blake2b as a general-purpose hash function; anything below 16 is discouraged when using Blake2b as a message authentication code.
<tankf33der>
this is copy-paste from monocyphers manual. :)
<beneroth>
yeah, but should I go with 32 bytes? default seems to be 128 with b2sum
<beneroth>
number of files/hashes is in the lower 5-digit range, so not that many
<Regenaxer>
Next day you can cut off the time before and after the conf
<Regenaxer>
Got segfault in misc/stress again
<Regenaxer>
...
<Regenaxer>
1 2 (1221 . 1194)
<Regenaxer>
{3} (999 . {672}) dat +A {1}
<Regenaxer>
{3} (999 . {734}) key +A {1}
<Regenaxer>
8
<Regenaxer>
664364 SIG-11
<tankf33der>
core file created ?
<Regenaxer>
Hmm, thinking about misc/stress.l I think I did a stupid change
<Regenaxer>
no core
<Regenaxer>
I old pil stress.l did in the parent
<Regenaxer>
(while (find '((P) (kill P 0)) Pids)
<Regenaxer>
(wait 1000) )
<Regenaxer>
I pil21 I changed that to
<Regenaxer>
(wait NIL
<Regenaxer>
(not (find '((P) (kill P 0)) Pids)) )
<Regenaxer>
which is stupid
<Regenaxer>
it calls the 'find' on *every* eveny
<Regenaxer>
event
<Regenaxer>
the parent gets very busy
<Regenaxer>
Now I simplified it
<Regenaxer>
using (kids)
<Regenaxer>
testing
<Regenaxer>
What I do now is simply (while (kids) (wait 999))
<Regenaxer>
No more list is 'Pids' built
<Regenaxer>
The load of the parent process decreased from 65% to 55%
<Regenaxer>
I've released it
<Regenaxer>
This probably does not fix the segfault
<Regenaxer>
but it is better this way
<tankf33der>
ok
<beneroth>
pilCon in ^42 min, Regenaxer ?
<beneroth>
not sure how much I can take part, but will try
<beneroth>
m_mans, yeah recording in the browser should be possible. I never tested it though
<beneroth>
Regenaxer, possible that recording on app is not (yet?) possible
<Regenaxer>
PilCon in 39 mins, right
<Regenaxer>
I think in the App it is possible toi
<Regenaxer>
too
<Regenaxer>
There is a recording button
<Regenaxer>
and a function to share to YouTube, whatever that means
<Regenaxer>
But I think it makes no sense to record one's own view
<Regenaxer>
I switch manually between various views during the conference
<Regenaxer>
Not helpful for anybody else
<Regenaxer>
The other problem is that most people don't attend by video
<Regenaxer>
or is this a problem in the app?
<beneroth>
ah, no
<beneroth>
if the button is there, I would expect it to work
<Regenaxer>
cause I switched to the browser sometimes, and it seemed it showed a lot more videos
<beneroth>
possibly
<Regenaxer>
But work how?
<Regenaxer>
Show what *I* see?
<beneroth>
I guess so..
<Regenaxer>
It is what *I* am interested at that moment
<Regenaxer>
Confusing for anybody watching this
<Regenaxer>
The app seems never to show all
<Regenaxer>
I click on icon frequently
<Regenaxer>
icons
<beneroth>
well Regenaxer you can not expect software developers think about minimizing the restrictions they put in place accidentally, they only see one purpose and commonly don't think about reusability ;-)
<beneroth>
no idea, sorry
<Regenaxer>
I don't think it is a restriction
<Regenaxer>
in the software
<beneroth>
I would say someone (or multiple people) should record (probably using the webinterface, dunno what is the difference), and then maybe pick the stream which is most relevant to the current talking...
<Regenaxer>
Yes, but hard to decide
<Regenaxer>
You need an editor ("cutter")
<Regenaxer>
Anyway PilCon's are not linear video presentations
<Regenaxer>
So I have no idea what it "means" to record them into video streams
<Regenaxer>
m_mans talked only of "audio" though
<Regenaxer>
this might be more clear
<Regenaxer>
But how to do it?
<Regenaxer>
Next stress.l crash:
<Regenaxer>
1 2 (1240 . 1207)
<Regenaxer>
{3} (999 . {661}) dat +A {1}
<Regenaxer>
{3} (999 . {734}) key +A {1}
<Regenaxer>
!? (NIL . {450})
<Regenaxer>
9
<Regenaxer>
NIL -- Undefined
<Regenaxer>
Seems it got completely confused
<Regenaxer>
Not useful to run the stress test, it gives no indication *what* exactly is wrong
<Regenaxer>
It was always one of the children which crashed
<Regenaxer>
All others continued and terminated normally
<beneroth>
isn't it connected to the select() problem?
<Regenaxer>
No, not at all
<Regenaxer>
pil21 doesnt even use select()
<beneroth>
ok
<Regenaxer>
And *if* it were, nothing works
<Regenaxer>
all task's I mean
orivej has joined #picolisp
orivej_ has joined #picolisp
orivej has quit [Ping timeout: 272 seconds]
orivej_ has quit [Ping timeout: 272 seconds]
m_mans has quit [Remote host closed the connection]