@# Quotes DB     useful, funny, interesting





Google
 
Web www.quotesdb.info
Undernet  |  EFnet  |  Quakenet  |  Freenode  |  Dalnet  |  Ircnet  |  Galaxynet


Comments:

<0> ok zordan
<0> it gets worse
<0> gcc virtual methods are cdecl, msvc++ methods are thiscall
<1> heh
<1> and name mangling?
<0> no names involved
<0> in this case
<0> hah it seems regparm(1) does the trick



<2> I'm trying to write a JIT interpreter. Is there a way around NOEXEC pages or whatever is causing segfaults so that I can generate machine code dynamically, then run it ?
<2> (if I use const char v[], no problem; if I use char v[], segfault)
<2> Found mprotect.
<3> i have a problem in RHEL 4
<3> bash-3.00# rpm -ivh xerces-c-2.6.1-2.i386.rpm
<3> error: Failed dependencies:
<3> libstdc++.so.6(CXXABI_1.3.1) is needed by xerces-c-2.6.1-2.i386
<3> from where can i get this libstdc++.... so on?
<4> have you check from your distribution package manager?
<3> apastinen, yeah, i don't have this
<3> although i read somehwere that RHEL 4 already hve this
<3> actually i wanna installed xerces-c package
<3> apastinen,
<5> hi, I was writing some test code, and is working ok with -O only
<5> http://rafb.net/paste/results/znGw5597.html
<5> any ideas why?, I debugged the -O3 generated executable and no instruccion was generated for lalala[50] = '\0' ;
<5> btw Im using gcc 3.4.4
<6> perhaps gcc thinks that value can never be used..
<7> hi
<7> In file included from libffi-src/src/types.c:26:
<7> libffi-src/include/ffi.h:83:20: error: limits.h: No such file or directory
<7> this is one of the libraries it is missing for gcc
<8> have you installed kernel-headers and glibc-headers
<7> khemHome, i have at least limits.h.. i searched for it
<7> maybe it is looking in the wrong place
<7> but this is weird on a mac
<7> i have the same software installed in different places
<7> by fink, by darwinports



<7> and other local location
<7> if i knew where gcc wasn't looking for it, i could make a symlink
<5> I have a little problem with inline asm
<5> i have:
<5> "subl %1,%0\n"
<5> : "=r" (ret)
<5> : "r" (ptr)
<5> and gcc sometimes choose the same register for %0 and %1
<9> heres my problem. I'm on an amd64 build, with gcc 3.4.6. a library I need to use in my project has been built with gcc 4.x, and it is also 32bit. If I upgrade gcc to 4.x, can I still link my 64bit proj with the 32bit lib?
<7> well but gcc is from the package xcode-tools.. it shouldn't be a problem.. i compiled several apps before
<10> plaf: I'm quite sure that it is not possible to link 32 bit with 64 bit code
<9> really? i read on the library's forums the owner said it was. alright, what can I do to make this all work, is there a way to force gcc to make my project 32 bit?
<10> -m32 I think
<9> so even though i'm on amd64, using -m32 will compile my project as 32bit and allow me to use the library?
<10> that will require that all the other libraries that you require are available as 32 bit versions on your system aswell
<10> yes
<9> uhoh
<9> i dont know where to get the 32bit versions for any of them
<10> on a friendly distribution your multilib setup probably contains 32 and 64 bit versions of the standard libraries
<9> its not standard libraries though, i need openal, ogre, newton, ogrenewt
<10> you might have to compile 32 bit versions of those then ... and their dependencies aswell
<10> unless there are prepackaged binaries somewhere
<9> i don't have the sources to make the libs, and the dependancies grow exponentially high
<10> maybe you could convince the owner of the 32 bit library to create a 64 bit version?
<9> o
<9> *i've been reading up on his forums, he's adamant there will be no 64 bit version
<10> well I don't see any other feasible options
<9> so, short of recompiling my whole system, theres no way out
<10> there might be 32 bit versions of libraries provided by your distribution for some of those ... but probably not for something like ogre
<11> 32-bit arch. If you start a structure with a double, are you garunteed that it will be quad-word aligned?
<11> (4.1.1)
<8> yes that right plaf
<9> alright then, abandoning the newton library


Name:

Comments:

Please enter the result of the sum 63 + 46 (to avoid spam):






Return to #gcc
or
Go to some related logs:

qdepend gentoo
#perl
#linux
forget-chrono
page scrape python mechanize beautiful soup
#openzaurus
#debian
#debian
KERİM SAY
ubuntu edgy timetable



Home  |  disclaimer  |  contact  |  submit quotes