| |
| |
| |
|
Page: 1 2
Comments:
<0> ok, i've asked on ##linux and on ##c, so i guess asking here is next:
<0> how can i make -lfoo link against libfoo.so instead of libfoo.so.1?
<0> hello
<0> anyone?
<1> is there a way to make gcc produce debugging information for multiline macros
<1> so it is possible to step through code with macros and step in the macro
<2> -E
<1> file not recognized: File format not recognized
<1> I get that when trying to link
<1> is -E a cflag or ldflag?
<2> c
<2> also, it will suppress compilation
<1> well I can't link with that option
<1> so it obviously isn't working for me
<1> hmm
<1> wait
<1> what you are suggesting is to run the preprocessor
<1> then run the compiler
<1> then link
<2> right
<1> that would work I suppose
<1> it's slightly more confusing to debug
<1> but that is a good point
<1> I thought there is an option for gcc to include preprocessor macros with debugging information
<2> maybe there is, but usually, if i simply see the preprocessed output, i find what's wrong right away
<2> 25 ppl in the channel and nobody's there?
<3> im using inline asm, 'ljmp 0x08, $somesymbol' and right after it 'somesymbol:' and i get "Error: symbol `somesymbol' is already defined"
<3> is gcc creating its own instance of somesymbol before it reads the next line?
<2> could be the $
<3> no, just make it complaining about the improper symbol now
<3> lol i just realized what i wrote, was on the phone
<4> i have some template code that's acting up. on gcc3.3.x it doesn't compile. on gcc3.4.6 it compiles but the code crashes in some nonsensical way, with minor changes to the code it even "crashes" the compiler itself, on gcc4.0 and gcc4.1 the code works just fine. i've tried it on various processors on linux and osx. i'm suspecting it's a compiler bug, but how can i be sure? and how do i deal with it?
<4> on 3.4 the code crashes in some STL map code - with or without stlport installed. it crashes in a function that probably shouldn't be called on the circumstances ( - in the key compare function for the map, which probably shouldn't be called when inserting the first element)
<3> what flags do i use to tell gcc i dont want to use their definition of new/delete?
<5> Hi! i am quite newbie in c/c++ environment. i like to get man pages for glibc, does someone knows where to get those?
<6> in general most distributions ship a manpages-dev package
<6> it all depends on your distribution
<5> mine doesn't do that..
<6> it isn't installed yet
<5> no.. there is no manual pages for my distro..
<5> i mean glibc manuals..
<6> you just want posix stuff
<7> which distro is that?
<5> lunar linux
<5> well at least i havent found the module which could install those
<7> not this one http://modules.lunar-linux.org/index.php?option=module&module=man-pages ?
<5> ups..
<5> =)
<5> i tought that it was installed.. =)
<5> but no, i will try it.
<7> xeer: ?
<5> i'll be damned.. =) it did it. Thanks zordan, i just spend 6 hours to search manual pages.. =)
<7> you should try without the blindfold next time :P
<3> gcc is telling me my new function is required to have size_t
<7> yes of course
<7> how else would you know how much memory to allocate/free
<5> do i still miss some man pages, i cannot get man page for sizeof?
<7> I don't think there is one
<5> hmh..
<7> because sizeof is an operator
<5> ok
<7> you could get a draft copy of the C standard
<5> draft copy?
<7> yeah I think those are free
<7> as opposed to the final standard
<6> not all calls have a manpage
<5> so.. if you like to program with, C, you just must know these stuff?
<7> that's what the standard is for
<7> and there are plenty of books too
<6> or where you have a book for, to search basic stuff up
<7> http://www.open-std.org/JTC1/SC22/WG14/www/docs/n1124.pdf
<6> when you know 'enough' you use the manpages as a reference
<7> that should be good enough to look stuff up
<7> sizeof is in 6.5.3.4
<5> ok, thanks, i will check it..
<5> your link crashed M$'s explorer.. =)
<7> probably your adobe plugin that crashed it
<5> yeah, that was it..
<6> perhaps someone here has an idea
<6> I try to build a small library which will be loaded using LD_PRELOAD
<6> its task is to capture a certain opengl call
<6> these days most opengl programs dynamicly load opengl
<6> so I want to 'dlsym' dlsym
<6> I tried to emulate it using dlvsym but that doesn't seem to work well on all systems (it crashes on 64bit systems)
<6> I also tried 'ld --wrap' but it won't work for LD_PRELOAD
<6> does someone here has an idea on how to achieve this?
<8> hi
<5> newbie guestion, why this prints garbage? http://rafb.net/parte/results/ZgzjS723.html
<5> ups, there is 'parte' in link, should be 'paste'
<7> the sizeof won't work like you expect it to
<5> =(
<7> it will return the size of the pointer
<7> so probably something like 4 or 8
<5> 4.. =)
<7> try something like int i=0; while(host->h_addr_list[i++]!=0)
<5> so, how can i get the element count in list?
<5> i mean array
<7> the list is 0 terminated
<5> ok
<9> what would piss gcc off so much that it would tell me asm volatile is not a valid operand
<10> Does any1 know, if gcc 3.4.2 has problems with finding precompiled header?
<10> if e.g. my precompiled header is in src/X.h.pch, I'm giving gcc -I"./src", and *sometimes* when I include "X.h" I'm getting error that it can't find the include file, after I change "X.h" to "./X.h" I'm going through ...
<11> who can help me with making C++ clases in shared object (dynamic library (implicit linkage))
<12> hi
<12> is it possible that in an instance of a cl*** which does not use the virtual method mechanism
<12> no space for a vtable pointer is allocated?
<11> titou: hi
<11> titou: i do not know it clearly.. it's worked once properly..but then i change code (add or remove some variables, the pointer crashed)
<12> hu?
<11> for example:
<11> cl*** A -- abstract
<11> cl*** B : virtual public A
<11> cl*** C {
<11> A* base;
<11> B* derived;
<11> };
<11> so i use such polymorph mechanism:
<11> base = new B();
<11> base->someMethod(); // realized in cl*** B
<11> it is impossible to allocate some memory for A cause it is abstract.. i guess..
<11> titou: how to allocate space for vtable?
<12> you can't
<12> it's the compiler's job
<11> some gcc's flags?
<12> i don't think so
<11> so, is it my curved hands?)
<12> duno
<11> titou: how to solve this?
<7> if there are no virtual methods no vtable will be created
<12> damned it's a mistake...
<11> Zordan_: i use virtual methods
<7> andy: I was talking to titou
<12> i used to write virtual ~mycl*** .... and i write it too for this cl***!
<12> sorry :o)
<12> you're right, there is no vtable variable
<11> i never used IRC before)
<7> there is a first time for everything :)
<12> then welcome here :)
<11> how messages flows here? i see 28 users and only we talk here, and i see you're talking about some problem i didn't saw the begining of %)
<12> there are all sleeping :)
<12> or working o anything else
<7> most of the users here idle ... which means that they are just here to listen or waiting for a more interesting conversation to join
<11> ah.. so it's a chat certainly like i've supposed )
<7> that is implied by the name ;)
<11> ))
<12> (IRC = Internet Relay Chat)
<11> yeh)
<11> about dynamic library - i used C++ with implicit linkage.. so what is the vtable variable? never know about this.. i listened about virtual tables of methods in theory but never have seen it in the code
<7> you don't see them in code
<13> hm.. is there any compilation flag used to make the unlikely() optimisation work?
<11> Zordan_: what's the solution or mistake i do ?
Return to
#gcc
or
Go to some related
logs:
IA32 wiwkipedia
#linux
double size xmms slow
dhclient dropline
Braning Irons
#kernel
#php
ubuntu install aterm
ubuntu stop the xserver
vmware freeze ubuntu
|
|