I am facing below errors when trying to statically link libDuma, Can you tell me how to ask g++ to use malloc from libDuma?
sunny@sunny-laptop:~/CodeTest开发者_如何学Python$ g++ ./testDuma.cpp -g -o testDuma -static -lduma -pthread
/usr/lib/gcc/i686-linux-gnu/4.4.5/../../../../lib/libc.a(malloc.o): In function `free':
(.text+0x4b00): multiple definition of `free'
/usr/lib/gcc/i686-linux-gnu/4.4.5/../../../../lib/libduma.a(duma.o):(.text+0x25f0): first defined here
/usr/lib/gcc/i686-linux-gnu/4.4.5/../../../../lib/libc.a(malloc.o): In function `malloc':
(.text+0x4bc0): multiple definition of `malloc'
/usr/lib/gcc/i686-linux-gnu/4.4.5/../../../../lib/libduma.a(duma.o):(.text+0x2730): first defined here
/usr/lib/gcc/i686-linux-gnu/4.4.5/../../../../lib/libc.a(malloc.o): In function `realloc':
(.text+0x5950): multiple definition of `realloc'
/usr/lib/gcc/i686-linux-gnu/4.4.5/../../../../lib/libduma.a(duma.o):(.text+0x23d0): first defined here
collect2: ld returned 1 exit status
Don't force a completely static link (don't use the -static
flag) -- doing so on any modern UNIX system is an extremely bad idea(TM).
Instead, link just the libduma statically. Either of these commands should work:
g++ ./testDuma.cpp -g -pthread -o testDuma /path/to/libduma.a
g++ ./testDuma.cpp -g -pthread -o testDuma -Wl,-Bstatic -lduma -Wl,-Bdynamic
Add -nodefaultlibs
flag to not link to libc. Or, remove -lduma
and link it dynamically after compilation with:
LD_PRELOAD=/usr/lib/libduma.so ./testDuma
精彩评论