| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8
Comments:
<0> openbsd is the one that hates the GPL, right?
<0> well, maybe not hate
<0> but dislike due to its restrictiveness
<1> http://lwn.net/Articles/176478/
<2> Humanoid: all the large users of the BSD license have a different philosophy than large users of the GPL
<2> restrictiveness is a subjective term in their arguments
<0> If I remember correctly though, they don't even advocate the BSD licence that much
<2> that's fine; not everyone feels the need to advocate things like that
<0> I was more asking about which they actually use in a roundabout way :D
<2> i think they do advocate sharing though
<2> especially with their letters to hardware mfgs. about opening docs
<0> ah, okay I found it
<0> "To this end, the Internet Systems Consortium (ISC) licence, a simplified version of the BSD licence"
<2> *shrug*, the BSD license is already damn simple
<0> Yeah, I wonder how they cut it down more
<2> go read it and find out :)
<0> I'm already on it :)
<0> "The ISC copyright is functionally equivalent to a two-term BSD copyright with language removed that is made unnecessary by the Berne convention. This is the preferred license for new code incorporated into OpenBSD"
<0> So I guess it does the same thing, just in different language.
<3> hi
<4> swish
<5> sw1sh
<6> good evening
<6> anyone ever had this error?: *** glibc detected *** free(): invalid next size (fast): 0x0806cab8 ***
<7> no, I don't corrupt my memory :-)
<6> heh, I don't want to corrupt it either
<6> i just can't manage to understand why that's happening.
<7> how big is your program?
<6> 206 lines.
<7> http://rafb.net/paste
<6> http://rafb.net/paste/results/LGWjgX46.html
<6> it's in spanish, but if there's something you don't understand, I can make it clear for you.
<8> what would be faster shared data access than pre-generating a hash or btree and storing it in an mmapped file?
<6> the error is bewteen the line 139 and 150
<8> (using a semaphore to lock the file when reading and updating)
<7> cursed: http://rafb.net/paste/results/VSF8Am22.html
<8> is it possible to ever have an SQL lookup be as fast as mmap()'ing a file?
<6> hrm.
<6> you need the files.
<6> wait, give me a second.
<7> cursed: I fixed the warnings: http://rafb.net/paste/results/jGsQOw48.html
<7> (it still crashes, though)
<7> cursed: oh, it crashes because you're fclose()ing a null pointer
<7> cursed: don't cast malloc
<6> mauke, http://black.slackware.org/datos_e32k6_bin.zip
<6> er.
<6> mauke, http://black.slackware.org/~curse/datos_e32k6_bin.zip
<6> there.
<6> well, casting malloc have always worked for me tho.
<7> int i = (int)1 - (int)1; also works but you should still write int i = 0;
<9> cursed: people tell me i shouldn't speed when i'm driving because i'll get a ticket. but i still speed and nothing happens. are they wrong?
<7> (cleanup): pointer=0x81c6540, size=4, from irc.c:140:leerJuegosBin: memory corruption detected (buffer overflow?)
<7> line 140: temporal->portsID = (int *)malloc(sizeof(int));
<7> line 141: for(j=0;j<ports;j++){ temporal->portsID[j] = portsID[j];}
<7> do you see anything wrong with that?
<6> nope.
<6> let me explain you why i set that.
<6> ports is an integer that holds the lenght of how many [j]'s portsID has.
<7> cursed: wrong
<6> i read ports from "juegos.bin"
<7> portsID has a single element
<7> you're only allocating space for one (1) int
<6> portsID may have more than 1 int-
<7> cursed: wrong
<7> portsID = (int *)malloc(sizeof(int)); <- there. a single int.
<6> hrm, it should be malloc(sizeof(int)*ports); then?
<9> portsID = malloc(sizeof(int)); // i can't believe you casted malloc, mauke :-P
<7> Alipha: not my code :-)
<7> cursed: portsID = malloc(ports * sizeof *portsID);
<7> nice, now I get SIGFPE
<6> heh, works for me.
<6> it was the sizeof.
<6> hrm, our teacher told us to cast malloc
<7> oh, right. you're doing malloc(0) and my wrapper can't deal with that
<7> burn your teacher
<6> thanks for the help mauke
<7> ok, fixed
<7> wow, that's a lot of memory leaks
<4> muke
<10> mauke, a lot of mallocs round up 0 instead of returning NULL.
<7> wait, returning NULL is actually a better idea
<2> heh, i'm TA'ing a freshman CS lab and always give solutions to labs when the labs are over... they had to do one ***ignment where they checked a floating pointer number in several if-else statements to see if that raw score was a letter grade of A, B, C, D, or F.... in my solution i made it needlessly fast by rounding up to the next whole number (because 59.5 was D, 69.5 a C, etc.), casting to an int, dividing by 10 to cut off the last dig
<2> i figured the compiler could probably optimize the switch to a basic jump table, but if not, at least integer tests are faster than floating point tests ;p
<10> floats get pipelined though on a modern processor.
<7> "ABCDF"[f(n)]
<10> that's a nice way:)
<10> and converting floats to ints is rather slow
<2> i think the cost is reclaimed in the checking
<10> highly doubtful.
<10> float compares are 1 cycle.
<2> these are actually "double"
<10> might be 2 cycles then.
<10> of course it's like 40 cycles to convert them to ints
<2> well, i could always fix that by using fixed point numbers
<10> sure. i don't know why you are optimizing a crappy grading program
<10> i like mauke's method the best:)
<2> to be needlessly fast
<10> well you could cast the floats to ints. and shift and mask it and do a big wad of switches.
<11> it doesn't seem like this is something to be caring about in a freshman CS course
<10> i'd fail someone for wasting thier time this way:P
<2> i think it is; they can see what different solutions look like
<2> OrngeTide: you'd probably be fired then ;p
<10> i'd be like: you don't seem to understand what really matters in software development. start over.
<10> nah. it's impossible to fire a professor
<2> it is if they're not tenured
<10> even non-tenured professors are hard to fire
<10> as long as you don't piss off anyone important your job is way more secure than in the "real world"
<11> Vratha: that's true, there are different solutions, but i don't think they need to be worrying about optimizing their grading program at this point
<10> i think it would be a nice change for me to get fired with cause instead of my usual terminated without cause
<11> :P
<4> OrngeTide: you aint no certified nigga
<10> don't be a dumb ***
<11> too late
<7> All tests successful.
<7> Files=3, Tests=34, 1 wallclock secs ( 0.48 cusr + 0.02 csys = 0.50 CPU)
<7> ah, it feels so good
<10> you test your software?! wtf.
<10> you're making the rest of us look bad
<12> mauke: what's that?
<7> hah, I've even written half of the tests before starting coding
<13> .
<7> evilgeek: make test | tail -n2
<12> right, but what are you testing?
<7> a terminfo parser/interpreter
<12> oh, ugh.
<12> (terminfo or termcap? and why do you only have 34 tests?)
<10> mauke, that's terrible. please leave the testing for the testers. if you bother even telling people about changes you check-in. (I don't)
<7> I just have no idea how $<123> is supposed to work
<12> (there are thousands of terminals in my termcap.)
<13> wwhy ???
<10> i could really use a terminfo parser/interpreter for a project
<12> and, like, doesn't libtermcap do this already?
<7> evilgeek: this is just the basic stuff, to see if it runs at all
<12> ah.
<10> gambit could m***ively use one. i don't think he wants to write one either
<7> the first test happens to check for the existence of all methods
<14> You aint no certified nigga
<10> i'm pretty tired of this.
<10> Gambit-, you want to write one?
<7> (about 500 functions)
<10> Gambit-, you decided you still want to do it using ncurses?
<15> Well on the immediate short term
<7> OrngeTide: don't worry, I haven't written any docs for it :-)
<15> on the immediate immediate short term it'll be unformatted.
<10> mauke, well trying to pull the routines out of ncurses is worse than not having docs:P
<10> Gambit-, good man. least effort is always the best.
<7> ncurses just feels icky
<14> ;]
Return to
#c
or
Go to some related
logs:
#flash
#c
gianan from star trek
#nhl
#nhl
#freebsd
#computers
#computers
#gentoo
Mr.PenCap
|
|