On Friday 14 May 2010 15:54:45 Timo Teräs wrote: > On 05/14/2010 09:34 PM, Denys Vlasenko wrote: > > 2010/5/14 Timo Teräs <[email protected]>: > >> On 05/14/2010 09:22 AM, Rob Landley wrote: > >>> On Sunday 09 May 2010 09:19:58 Denys Vlasenko wrote: > >>>> On Sunday 09 May 2010 07:31, Rob Landley wrote: > >> > >> While at it, we could get rid of include/applets.h too. One very nice > >> way would be to use linker section. You can use gcc __attribute__ > >> section("sectionname") [1] to put the applet definitions to separate > >> memory area. It is after that possible to iterate over the structures > >> in that special section using __start_sectionname and __stop_sectionname > >> symbols which are autocreated [2]. This is also roughly how Linux kernel > >> initcalls work [3]. > > > > This will break ld --gc-sections. > > I think ld handles --gc-sections properly if __start_sectionname symbol > is referenced. If not, using __attribute__((used)) should work. > > - Timo
Having a script create a human readable source file is something you could explain to a middle school student with two weeks experience with C. The tools to do so are standardized by posix, and the versions busybox itself produces support all necessary functionality. Having a magic linker function use vendor-specific extensions in a manner only supported on certain toolchain versions is not just non-obvious, but you yourself aren't sure off the top of your head how you'd have to go about implementing it, and the explanation you just tried to give is full of non- obvious conditional requirements. Given that the two techniqes produce equivalent results, this doesn't look like a hard decision... Rob -- Latency is more important than throughput. It's that simple. - Linus Torvalds _______________________________________________ busybox mailing list [email protected] http://lists.busybox.net/mailman/listinfo/busybox
