configure.ac use AC_RUN_IFELSE to determine if the system is Solaris, BSDi .. but AC_RUN_IFELSE always fail when cross-compiling cause code is building to TARGET and so can't be executed on BUILD host. And we can't set a default value if this is a cross compiling build. So here is a patch to make possible to cache if system is Solaris & so on and assume that the system is not if cross-compiling, but let users define ac_cv_os_solaris=yes if they want to. Patch also reported to developer >>> Compiling source in /usr/i586-geode-linux-gnu/tmp/portage/net-analyzer/tcptraceroute-1.5_beta7-r1/work/tcptraceroute-1.5beta7 ... * econf: updating tcptraceroute-1.5beta7/config.guess with /usr/share/gnuconfig/config.guess * econf: updating tcptraceroute-1.5beta7/config.sub with /usr/share/gnuconfig/config.sub ./configure --prefix=/usr --build=x86_64-pc-linux-gnu --host=i586-geode-linux-gnu --mandir=/usr/share/man --infodir=/usr/share/info --dat adir=/usr/share --sysconfdir=/etc --localstatedir=/var/lib configure: loading site script /usr/share/config.site configure: loading site script /usr/share/crossdev/include/site/linux ... checking for memset... yes checking for select... yes checking for socket... (cached) yes checking for strchr... yes checking for Solaris... configure: error: cannot run test program while cross compiling See `config.log' for more details. Reproducible: Always Steps to Reproduce: 1. i586-geode-linux-gnu-emerge -vat tcptraceroute 2. 3.
Created attachment 278895 [details, diff] files/tcptraceroute-1.5_beta7-BJA-cross.diff
Created attachment 278897 [details, diff] tcptraceroute-1.5_beta7-r1.ebuild.diff
the cross-compile fallback case should look at $host_os to guess a good value the rest of the patch looks pretty good and should go upstream
(In reply to comment #3) > the cross-compile fallback case should look at $host_os to guess a good value > > the rest of the patch looks pretty good and should go upstream You made a point but I don't known those os $host_os :/ Do you have some kind of list ? This could make to configure really more simple patch have been reported upstream to main developer as mailing list seems to be down
Created attachment 279117 [details, diff] files/tcptraceroute-1.5_beta7-BJA-cross.diff Here is a refresh patch to use $host
*tcptrace-6.6.7-r1 (09 Jun 2013) 09 Jun 2013; Jeroen Roovers <jer@gentoo.org> +tcptrace-6.6.7-r1.ebuild, +files/tcptrace-6.6.7-cross-compile.patch: Fix cross-compiling by Bertrand Jacquin (bug #374085). Set RDEPEND. Show notes only on first install. So that was a different cross-compile bug? Any progress on this one?
(In reply to Jeroen Roovers from comment #6) > *tcptrace-6.6.7-r1 (09 Jun 2013) > > 09 Jun 2013; Jeroen Roovers <jer@gentoo.org> +tcptrace-6.6.7-r1.ebuild, > +files/tcptrace-6.6.7-cross-compile.patch: > Fix cross-compiling by Bertrand Jacquin (bug #374085). Set RDEPEND. Show > notes only on first install. Thank you for this. > So that was a different cross-compile bug? Any progress on this one? I don't get it, sorry. Could you precise your thought ?
(In reply to Bertrand Jacquin from comment #7) > (In reply to Jeroen Roovers from comment #6) > > *tcptrace-6.6.7-r1 (09 Jun 2013) > > > > 09 Jun 2013; Jeroen Roovers <jer@gentoo.org> +tcptrace-6.6.7-r1.ebuild, > > +files/tcptrace-6.6.7-cross-compile.patch: > > Fix cross-compiling by Bertrand Jacquin (bug #374085). Set RDEPEND. Show > > notes only on first install. > > Thank you for this. > > > So that was a different cross-compile bug? Any progress on this one? > > I don't get it, sorry. Could you precise your thought ? I couldn't have been more precise. You filed bug #374085 after this one, but this bug is not fixed yet while that one is.
(In reply to Jeroen Roovers from comment #8) > I couldn't have been more precise. You filed bug #374085 after this one, but > this bug is not fixed yet while that one is. ah, wrong package. sry
Fixed in tcptraceroute-1.5_beta7-r2.