Hi,
building the master branch (a267bc2) on Debian systems fails when
linking the unit tests to the "check" library with this error:
CCLD unit/test-eir
/usr/bin/ld: /usr/lib/gcc/x86_64-linux-gnu/4.6/../../../../lib/libcheck.a
(check.o): relocation R_X86_64_32 against `.rodata.str1.1' can not be
used when making a shared object; recompile with
-fPIC /usr/lib/gcc/x86_64-linux-gnu/4.6/../../../../lib/libcheck.a:
could not read symbols: Bad value collect2: ld returned 1 exit status
make[1]: *** [unit/test-eir] Error 1 make: *** [all] Error 2
AFAIU this is a Debian quirk, the facts of the issue are more or less
these (please correct me if I am wrong):
- libcheck.a, as per upstream default, is generally built with -fPIC
- Debian policy tells that by default .a are not expected to be
built with -fPIC, so they provide libcheck_pic.a as alternative
instead.
- BlueZ build system tries to dynamically link the unit tests to
libcheck.a, but on Debian systems this is a non-PIC library and on
X86_64 linking to non PIC as shared library is not allowed.
After a little bit of research I found out that gstreamer people also
had run into that issue and they have a m4 macro to pickup
libcheck_pic.a if present:
http://cgit.freedesktop.org/gstreamer/common/tree/m4/check.m4
I have no idea if you BlueZ dev too want to handle this in-tree or
leave the burden to distributors.
The workaround I use for now is:
- run "make" and let the build system fail
- show the failing command with "make V=1"
- change -lcheck to -lcheck_pic and rerun the command
- "make" will now proceed with the build
Regards,
Antonio
--
Antonio Ospite
http://ao2.it
PGP public key ID: 0x4553B001
A: Because it messes up the order in which people normally read text.
See http://en.wikipedia.org/wiki/Posting_style
Q: Why is top-posting such a bad thing?
Attachment:
pgpphf0zzN5EF.pgp
Description: PGP signature
