On Fri, 19 Oct 2018 10:47:25 -0300 =?UTF-8?Q?Lisandro_Dami=C3=A1n_Nicanor_P=C3=A9rez_Meyer?= <perezme...@gmail.com> wrote: > Hi! I have been looking at this bug and then at the packaging and > installed files. > > dpkg -L libsane tells me that the *only* public library shipped is > /usr/lib/<m-a>/libsane.so.1 > > The rest of the libraries are under /usr/lib/<m-a>/sane/ so they are > *private* libraries. > > I have also taken a look at libsane-dev and it seems to ship headers > only for the public library. > > So what you should really do is simplify symbols handling: you only > need symbols for public API, ie, for /usr/lib/<m-a>/libsane.so.1 > > You can safely drop all the others except if at some point you need to > publish private headers. >
You are right. Those symbols file shipped by sane-backends do not look correct. I also question, whether a symbols file per architecture is really needed: libsane1.symbols.alpha libsane1.symbols.amd64 libsane1.symbols.arm64 libsane1.symbols.armel libsane1.symbols.armhf libsane1.symbols.hppa libsane1.symbols.i386 libsane1.symbols.kfreebsd-amd64 libsane1.symbols.kfreebsd-i386 libsane1.symbols.m68k libsane1.symbols.mips libsane1.symbols.mips64el libsane1.symbols.mipsel libsane1.symbols.powerpc libsane1.symbols.powerpcspe libsane1.symbols.ppc64 libsane1.symbols.ppc64el libsane1.symbols.s390x libsane1.symbols.sh4 libsane1.symbols.sparc64 libsane1.symbols.x32 My guess is that after restricting the symbols file to the public library, i.e. /usr/lib/<$arch>/libsane.so.1, you'll have a single symbols file which is identical on all architectures. Lisandro, could you file a separate bug report to get the symbols file(s) cleaned up properly. Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
