On Tuesday 02 December 2008, Jonathan S. Shapiro wrote: > Thomas: > > I understand what the patch is doing, but I am not clear what > problem it is trying to solve. Is the idea to determine SystemDirs > at run time based on the install location of bitcc?
Yes. The problem I encountered was that bitcc failed to find libbitc*a and as well as the basic *.bitc includes and thus required explicit -I and -L settings. > If so, this is not the correct way to go about it, because /proc is > not universally available. What I think you want to be doing here > is calling realpath() on the value of argv[0]. Only if agrv[0] contains at least one '/', otherwise searching $PATH for argv[0]'s existence _may_ result in the correct location. On Windows adding ".exe" might be necessary. [snip] > So first, do I understand correctly what you are attempting to do > here? Yes > And if so, should this be done dynamically, or should it be done at > configure time? Is the goal to make the bitcc package relocatable? > I think that would be good, I just want to make sure that we solve > the right problem here. At runtime would be nice due to the relocatability. However there is atleast one potential trap: autoconf ./configure --bindir=foo/bar ./configure --includedir=foo/bar ./configure --libdir=foo/bar make install exec_prefix=/foo/bar Thomas _______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
