https://bugzilla.redhat.com/show_bug.cgi?id=1016809
--- Comment #23 from Michael Schwendt <[email protected]> --- > I_PREFIX=`$PKG_CONFIG --variable prefix rodent 2>/dev/null` Yep, there won't be any rodent.pc file in clean buildroots, such as used by Fedora. But primarily, the rodent.pc file is "weird", because of what it does with its "includedir" and "Libs" lines: > includedir=/usr/include/rfm/modules Not even librfm installs anything into that path. In case you want module/plugin packages to use pkgconfig to retrieve an installation path at configure-time, consider something like libdir=/usr/lib64 moduledir=${libdir}/rfm/modules $ pkg-config rodent --variable=moduledir /usr/lib64/rfm/modules and a similar thing for plugindir. Such a query can be embedded into configure scripts for external module/plugin packages. > Requires: librfm In a .pc file, this creates an inter-dependency with librfm.pc and pulls in the --cflags and --libs for librfm. Considering that librfm has been split off and there is no own API for rodent, that is odd. > Libs: -L${libdir}/rfm/modules -lbcrypt -lcompletion -lcombobox > -lproperties -lmime -lmimemagic -lmimezip -licons And these are rodent module libs appearing on the linker options line for unknown reasons. Hope this makes clear why I've questioned the rodent-devel rpm for just this file. As a last resort one could have included the file in the base rodent rpm, even if nothing uses it at runtime. :-) -- You are receiving this mail because: You are on the CC list for the bug. You are always notified about changes to this product and component _______________________________________________ package-review mailing list [email protected] https://admin.fedoraproject.org/mailman/listinfo/package-review
