On Sun, Jun 28, 2015 at 11:48:48AM +0200, Dimitry Andric wrote: > On 28 Jun 2015, at 10:57, Baptiste Daroussin <b...@freebsd.org> wrote: > > > > On Sun, Jun 28, 2015 at 01:44:21AM -0700, NGie Cooper wrote: > >> On Sun, Jun 28, 2015 at 12:49 AM, Baptiste Daroussin <b...@freebsd.org> > >> wrote: > >> ... > >>> Not at all I have this on a both where I haven't yet r284898 iirc it is > >>> like > >>> this since the beginning I do not remember seeing those ld scripts with > >>> absolute > >>> path. > >> > >> $ cat /usr/lib/libc.so > >> /* $FreeBSD$ */ > >> GROUP ( /lib/libc.so.7 /usr/lib/libc_nonshared.a > >> /usr/lib/libssp_nonshared.a ) > >> $ what -q /boot/GENERIC.r283337+9c333ed/kernel > >> FreeBSD 11.0-CURRENT #4 r283337+9c333ed(isilon-atf): Tue May 26 > >> 21:49:09 PDT 2015 > > > > Yes you are right, I was looking at the wrong place. > > What is actually the perceived problem with having paths in those linker > scripts? If you use --sysroot, the libraries are searched relative to > that sysroot, right? > > (And yes, I know our gcc's sysroot implementation is broken. So please > fix that instead. :-) > WHat is the point in having absolute path in the linker script? having an absolute patch (or even no path at all) will make the compiler looking in its search path (and respecting sysroot). The only case where a path is needed seems to be when the lib you want to link to is not in the search path. Am I missing something?
Best regards, Bapt
pgpcfaZYzsXeB.pgp
Description: PGP signature
