https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82708
--- Comment #7 from joseph at codesourcery dot com <joseph at codesourcery dot com> --- On Wed, 25 Oct 2017, keno at juliacomputing dot com wrote: > First, the build process looking for the headers in <path>/sys-include rather > than <path>/include where glibc installs them. Leads to the same symptoms as > reported in this issue. You should be using a sysroot; directories such as $target/include and $target/sys-include should only ever have libc headers in bare-metal cases where you're using e.g. newlib and the concept of a native directory structure is meaningless. The sysroot must *not* be the same as the $prefix where the compiler is installed, or as $prefix/$target; it needs to be another directory, e.g. $prefix/$target/sysroot. glibc must be configured with --prefix=/usr and installed under the sysroot. > Then, I tried using --with-sysroot which I wasn't using before, but then I got > confusion between <sysroot>/usr/include and <sysroot>/include (most parts of $sysroot/include should not exist. Again, sysroot must not be $prefix/$target. (However, sysroot/lib and sysroot/usr/lib must exist even for configurations using directories such as lib64, though they can be empty directories in that case. Again, see how build-many-glibcs.py does things.)
