Lucifers,
I created a Clownfish branch 'include_dir_fixes' to address some problems with
the way Clownfish include dirs are handled:
https://git-wip-us.apache.org/repos/asf?p=lucy-clownfish.git;a=log;h=refs/heads/include_dir_fixes
The basic idea is that if a parcel is found for the first time in an include
dir, we only process files and classes of the parcel from this include dir. If
the parcel turns up in include dirs later in the search path, anything related
to the parcel will be ignored. This makes sure that we never mix files from
different include dirs for a single parcel.
With this change, we can safely enable the option of searching all of Perl's
@INC, the directories in CLOWNFISH_INCLUDE, and user-provided locations for
Clownfish headers. This should make CLOWNFISH_INCLUDE mostly unnecessary for
Perl projects since Clownfish headers will be typically found in @INC.
Nick
