Closed GoogleCodeExporter closed 9 years ago
This isn't a bug--it's standard behavior for installing GIRs. The install
location comes from the m4 macro provided by gobject-introspection (see
http://git.gnome.org/browse/gobject-introspection/tree/m4/introspection.m4)
which gets the information from the girdir variable in the
gobject-introspection-1.0 pkg-config file.
I believe (though I could be mistaken) you can do something like `make install
INTROSPECTION_GIRDIR=wherever INTROSPECTION_TYPELIBDIR=wherever` to get around
this, or simply pass --enable-introspection=no
Relevant code:
https://gitorious.org/sqlheavy/sqlheavy/blobs/518ad091d517cbf6b271b1b87cb7d338e7
30274e/sqlheavy/Makefile.am#line98
Original comment by nem...@gmail.com
on 6 May 2012 at 11:16
FWIW, libgda solved this issue with a configure option and the following code
in Makefile.am:
----
if ENABLE_GI_SYSTEM_INSTALL
girdir = $(INTROSPECTION_GIRDIR)
typelibsdir = $(INTROSPECTION_TYPELIBDIR)
else
girdir = $(datadir)/gir-1.0
typelibsdir = $(libdir)/girepository-1.0
endif
----
Please also note it seems not a lot of libraries use these INTROSPECTPON vars,
at least not gtk+, clutter, nor unique.
Original comment by emmanuel...@gmail.com
on 6 May 2012 at 8:58
I definitely don't like libgda's solution... I think something like using
AC_ARG_VAR to check for optional paths to override the pkg-config variables
would be more appropriate.
That said, I would much rather just stick with the standard behavior, so I'm
not going to change this in SQLHeavy without a change in g-i. You should talk
to the gobject-introspection guys about integrating something upstream in
introspection.m4. They already have a bug report (with a patch, though it does
something a bit different): https://bugzilla.gnome.org/show_bug.cgi?id=639782
Original comment by nem...@gmail.com
on 6 May 2012 at 10:44
Original issue reported on code.google.com by
emmanuel...@gmail.com
on 6 May 2012 at 8:56