| |
| |
| |
|
Page: 1 2
Comments:
<0> toying with stty, there is 'stty insane'
<1> ... ?
<0> 'stty sane' resets to reasonable defaults, 'stty insane' does the opposite, I read :)
<2> that would be ... weird.
<3> no insane here :(
<4> stty sane makes ^H appear when i press backspace :-P (also no insane here)
<1> insane in the m...
<0> netbsd
<0> I was doing difficult stuff in C to open/configure serial port, but that works fine using stty I see now
<0> I thought that was much more limited
<0> it can do most if not all things from man termios
<1> aye
<1> I fondled it back when I did my bluetooth ****, too
<0> ic
<0> 'fondled'
<1> touched.. if you prefer that ;)
<0> same issues
<2> ''stty parenb parodd ignpar'' should make for a mostly insane terminal. perhaps toss in a cs5.
<0> I'll p***
<1> michai: one day I will just set your baudrate to -1
<1> don't complain then, I warned you!
<0> grrr!
<0> take that biaach
<1> ok
<1> mum called
<1> must go to bed
<0> well... laters then
<0> laters
<5> udv
<5> hi
<5> for some reason my program exits with exit code 0, however it is a fastcgi app that shouldnt exit. but dont know why and where... is there way to find out where my program exited? with gdb or something?
<3> with gdb it should be possible
<3> or something
<3> just compile it with -g or -ggdb and then step over it
<6> yea, compile ya program with -g flag, and fire it up in gdb, set some breakpoints and then step thru teh code
<6> ^^
<3> or attach to the pid of running process
<7> maybe bp for exit will suffice
<7> or even truss/strace/ltrace/whateveryoucallit.
<6> strace requires kernel recompilation
<7> does it?
<3> _sparky: on what platform?
<6> all linux, well afaik
<3> hmm
<7> none linux, well afaik
<6> thought taht was the major + point of truss on solaris
<3> never had to recompile the kernel on linux for strace to work
<6> ah ok, i must have out of date docs then =/
<7> you just looked up your docs? ;)
<7> whatever.
<6> "Unfortunately the stock Linux kernel does not allow strace to do everything it wants. You can get some kernel patches here to fix that"
<3> weird.
<3> anything more precise than "everything it wants"?
<3> oh nevermind gotta go
<5> ok thnx
<8> _sparky: those patches are for 2.2.x kernel...
<6> as i say, thats teh docs i read re. strace..out of date :~
<6> regarding original question, think compile with -g and set breakpoint on exit
<9> moin
<10> moi
<11> a bit offtopic.. if the env variable is se to /sbin, will the shell look for binaries in the relative subdirs too?
<12> hi all
<13> azi: the env variable?
<11> lindi-: yes
<11> i'm wondering, if adding /sbin to it, matches /sbin/foo/fooo.sh as well
<14> azi: you mean $PATH ?
<11> cehteh: yes
<14> and no, it wont match subdirs
<10> $PATH is not the env variable, it's a env variable
<14> you can try /sbin/*/ ... but i really dont know if the shell expands that
<10> s/ a / an /
<14> you better dont, it is likely unsafe
<14> if you want to store your scripts in a dedicated directory then put them to /usr/local/sbin
<14> ensure that that is in PATH
<11> i see thanks
<12> is there any free c ide that i can use instade of visual studio ?
<14> many
<11> spe{trum: dec-c++ afaik
<10> emacs
<15> cehteh you saw the code which I gaved you ?
<14> I_v0: eh which one?
<15> the one we were talkin about at night yestarday or 2 days ago
<15> the binary tree
<14> ah that
<15> you seemed interested, it's on gpl you're free to use it ;-)
<14> lol ... no i dont need it :P
<15> ahh :P
<14> well its quite to specific to your problem and hardly useable for anything else .. and not complex anyways :)
<15> I understanded other thing after this question about free soft
<14> if you want to make code which is useable by others, then you have to make it either really complete for some task which is common and well document or very generic
<15> yeah, it's only a tree, you can insert there whatever you wan't still the only difference is that it can have more than two subnodes to each node
<12> many thx
<14> well i could offer you my linked-list
<14> I_v0: btw try 'man tsearch'
<15> cehteh will it manage more than two subnodes ?
<15> (I'm just reading it now)
<14> I_v0: no if you need some specific tree you have to code it by yourself .. but i see no point in your problem to use more than 2 subnodes
<15> What I thought was that every page will be one node, and each page can have more than 2 links pointing to somwhere
<14> but i dont want to demotivate you, go on, youll learn
<15> ;>
<15> thanks :)
<12> this dev-c++ is really a great alternative of m$ vc++
<12> even for compiling
<15> spe{trum free compiler for windows?
<15> gcc
<14> spe{trum: i use emacs ;)
<15> ahhh the ide
<12> lol, even noptepad can go
<15> I use vim
<15> spe{trum no, notepad isn't what you want :P
<12> :) lol, i like cool ide :P
<12> do gcc for windows or mingw produce fast code as m$vc ?
<15> I would say that gcc is the the best compiler I ever seen
<12> yes, i've seen for linux, it's really invincible from my asm
<15> spe{trum just give it -O3 and it should produce the fastest possible code
<12> very good. so i have to download gcc separated ?
<14> uhm
<14> if speed is your problem ... then first fix it by profiling and optimizing your program and choose the proper algrithms before blameing the compiler
<15> well, cehteh is right if you'll **** up in the code, then even god won't help you, the compiler will do the best but not miracles
<12> yes, i was just asking, since i've seen that copilers as c builder was 40% slower that my bad asm replaced code
<14> 40% is nothing if you are in the wrong complexity domain
<12> ok, btw i've done some asm tests on linux, i was 40% slower than the gcc produced code
<12> i don't need a very fast compiler, just like to test the performance sometime, but with gcc, inserting may asm code can just slow down the performance :)
<14> performance of what?
<12> i mean my asm just slow down the speed on esecution
<12> better i let gcc do his job :)
<16> spe{trum: if you keep your asm routines VERY small, they can be faster than what GCC creates, however such cases are rare indeed...
<11> + the produced code is most likely to be unportable :-)
<16> that, too :)
<16> depends on what you're doing
<16> for a, say, brute force p***word cracker you need just for yourself, it's all fine and well
<16> and for, say, a KDE contribution almost totally out of the question :)
<11> + lde guys don't care about optimisations ;-p
<12> fine, :) i just like to have fast executables, even if i don't need it much, except maybe for the rsa alghoritm we use
<11> most people tend to optimize "bad" algorithms...
<16> and the wrong parts of the code
<16> I've done that
<16> one thing you just have to learn the hard way is not to optimize your surface culling routine when your texture mapper is the bottleneck...
<11> yeah, and, what my co-worker is used to say :" optimize ****ty code, and you'll end with fast crap"
<12> ive seen that dev c++ use his own gnu static libraries, for example as libwsock32.a
<12> how do they have recompiled all that libraries ? are the sources of apies of microsoft ?
<16> I had severe problems parsing that
<12> sry, i've difficult in explain in english
<12> i mean
<16> Microsoft only owns their implementation of the W32 api... the standard C library, as provided by GNU, is open source
<16> and freely usable by anyone
Return to
#c
or
Go to some related
logs:
#gentoo
#worldchat
stone freedom palermo
#london
#worldchat
#london
#worldchat
#worldchat
#worldchat
#worldchat
|
|