> Il Martedì 29 Ottobre 2013 10:23, Robert Lemmen <rober...@semistable.com> ha > scritto:
> > On Tue, Oct 29, 2013 at 08:50:05AM +0000, Gianfranco Costamagna wrote: >> >because of this, libcheck is shipped as a *static only* library, which >> >means you can still link against it and don't have to include > libcheck >> >*code* in your project. the static library is called liobcheck.a, and a >> >typical linker invocation involves "-static" to tell the > linker about >> >the fact that you want to link statically. >> >> Thanks, I tried, but I'm still having linking problems. >> (I don't know why travis-ci succeedes, while I have problems on my > computer > > hmm, can you tell me in detail what you are doing (i.e. linker > invocation) and what error message you are getting? > Ok I found and fixed the problem. The problem was that (seems to be a fault in debian/check build), check wasn't linked against pthread, math and time. this commit fixed the problem https://github.com/LocutusOfBorg/ettercap/commit/a8d67aa64ecea865971105fff1256de7ecf749cc I had some of this kind of problem with the switch from eglibc 2.15 to eglibc 2.17, I had to add manually some pthread or math somewhere in the makefiles on various debian programs, because the libraries weren't automatically linked (I never looked deeply at this kind of issues). can this be fixed on check side? >> Sorry I didn't undestand this part, the pourpose of check is to include > tests on a particular software right? >> >> what we do is: >> - build ettercap, >> - build tests (linked against check) >> - run tests. >> >> so check is not linked against ettercap, but only against test code, and it > isn't shipped with any installer, is just for running testsuites. >> >> Do you have any better way for running tests on ettercap project? > > no better suggestions, what you do sounds absolutely right. I just said > that because a typical reason for people wanting a .so instead of a > static lib si that they want to ship the test binary and are concerned > about the size of it. and that's just not what unit tests are for... > > Well, thanks for the explanation! > regards robert > > -- > Robert Lemmen http://www.semistable.com > -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
