| |
| |
| |
|
Comments:
<0> bye
<1> is there sleep function on unix system?
<1> void Sleep(int ms); ?
<2> antifreeze shell dependent
<2> in sh sleep 1;
<3> good morning...
<4> antifreeze: several. alarm, nanosleep, select, poll, sleep, ualarm and usleep all come to mind.
<4> possibly even pause.
<4> gah, and setitimer, of course.
<4> that seems to round them out nicely.
<5> moin
<5> mhm, only windows guys blinks
<5> gayhomo***ual!
<6> No, but thanks for the offer.
<5> :)
<7> i got kicked or something ? anyho my problem is here: http://www.copypot.com/404
<4> mirex2: you want #c++
<7> yes I do. I see that I'm messed up just as my irc connection
<7> thanks twkm
<4> your quit was ping timeout.
<8> moin
<9> mirex2: I guess your post number - 404 - was a tad unlucky :0
<9> bah
<9> how long have I been away?
<4> 3 hours.
<4> probably more, but that long since mirex2 mentioned that link.
<9> I'm like that guy from volvo ad
<10> dimaqq: phased out in front of the kbd?
<11> http://groups.google.com/group/comp.lang.c/browse_thread/thread/7d093e68e41df2da
<11> hehe
<10> haha, why?
<11> becoz the macro is preprocessed py CPP
<11> (the answer if you scrol down)
<11> s/py/by
<9> eleven: now, I phased out in front of another irc server
<12> Morgen
<13> moin
<14> It is possible to hard lock files in UNIX ( BSD ) systems ? For example when I hard lock some file, any other application cannot open this file before i unlock.
<3> if this program knows about this lock
<15> you can open it for writing maybe ? maybe then it won't be openable ?
<3> wrong
<3> it works in windows only
<3> unix has another way
<15> you can open file for reading while someone writes to it ?
<3> of course
<3> remember tail
<14> i don't have problem
<14> i am interesting about concrete unix syscall
<14> but in POSIX and C99 I cannot find some hard link
<11> wieczyk: it's not possible
<11> at last not in POSIX
<11> it's only possible to lock a file in such way, that some other process can't aquire a lock
<3> by the way, what's about O_EXLOCK?
<14> azi: i know
<14> but i want to check if it is any other possibilities
<14> thanks
<11> well sure there is ;)
<16> perhaps lockf()?
<11> guys
<11> it's not possible to do that under posix ;)
<11> wieczyk: why do you need that? to protect some file from other processes to access it?
<14> azi: do you know if some UNIX give a syscall to hard locking ?
<14> azi: no, i am doing some 'research'.
<14> ;]
<3> create a file and remove it at the same time - this is the best lock :)
<17> yup
<11> it's not possible to do that with posix, and i'm afraid any other unix specific syscal is nto portable
<17> don't forget to open it :)
<11> denk: i was just about to suggest that
<14> i don't havy any problem, i am only interesting
<14> becouse i use BSD and bsd don't give any private syscall to hard link
<14> linux has ?
<11> wieczyk: the solution also depends on who's gonna access the file
<3> azi: sorry, I am too slow today...
<11> denk: you? :)
<3> yep :)
<3> after installing of solaris 10 on x86...
<11> heh, why was that good for?
<3> I'd like to test oracle under it
<11> i c
<3> and try to write for it in C
<11> i remember the last time i checked solaris code..
<11> i browsed the login source
<11> which was like "void main()" heh
<18> haxx0r!
<19> uhm
<20> hi
<18> yes
<20> I'm trying to write sth resampling linear pcm from 8bit to 16bit and in reverse.
<18> yes
<20> I understand that I should square every byte to transform 8bit to 16bit - and square root every 16bit if I want reverse, if that's unsigned PCM.
<20> but what to do if this is signed?
<18> rather a "read the pcm specs" question I guess
<18> if it's simply for the value range, double the signed stuff before you do the unsigned operation on it?
<20> PCM isn't something needing a spec - it's simply one-byte or two-byte signal level stored in integer value, as far as i understand.
<20> (if we are talking mono)
<18> then I don't see your problem
<18> *shrug*
<20> hm
<20> as always, by describing my problem I've understood it a little better
<20> If I have an one-byte sample with range from -128 to 127 and want to convert it to value from -32768 to 32767, I need no squaring, but simply to multiply it by 256
<20> regardless of whether it's signed or not
<18> I would just shift it
<18> but then again, just multiplying it seems undynamique
<18> but then again again.. it seems the right thing to do
<20> mhm
<20> If that's, as I think, a frequency scale, it must cover the same area, whether it's 8bit or 16bit. To maintain scale it must be multiplied. Or shifted, as you said.
<18> I agree there.
<18> I just stumbled over "1 -> 256" .. much further away from 0
<20> So... let's fight with sox and see who wins >:>
<18> it looked weird to me
<21> sox is the evil brother of tux?
<20> very evil
<20> it sings when uninvited
<20> badly
<0> your daily contribution? :)
<18> op me, I want to ban something, too
<19> prevention :)
<19> /whois Ravi
<19> that`s not me, but he joins everywhere where i am ;)
<19> whata curiosity :)
<19> *yawn*
<21> well, we can always recognize you from your backticks.
<19> hurra.
<18> we could`s copy them though
<5> hmm
<5> who gave him @?
<5> now we are lost!
<19> yeah!
<19> you are ;)
<5> I knew it!
<21> it was that humpty dumpty guy with a stick.
<5> mhm reverse chanops
<5> Total of 79 nicks [73 ops, 0 halfops, 0 voices, 6 normal]
<5> :)
<21> weird fantasies you have.
<19> kay
<19> you won
<19> now you are SAFE :)
<21> like in the calvin&hobbes strip?
<19> dunno
<19> mmkay
Return to
#c
or
Go to some related
logs:
#worldchat
#gentoo
#worldchat
#worldchat
#worldchat
#sicilia
#worldchat
#worldchat
#worldchat
#worldchat
|
|