On 2014-02-10 18:43, Andrew Hughes wrote:
You're already using it:
PKG_CHECK_MODULES([LIBFFI], [libffi])
Why that's in LIB_SETUP_STATIC_LINK_LIBSTDCPP, I have no idea.
Because libraries.m4 is in need of a long overdue cleanup. :-(
* Second, I believe the original intention was to allow for a third
option, --with-libX=<dir>, which would point to an location in which the
library is installed, similar to how e.g. --with-alsa works.
Yes, but you don't allow that:
AC_MSG_ERROR([Invalid value for --with-zlib: ${with_zlib}, use 'system' or
'bundled'])
so using --with-x is confusing if someone does specify a directory.
Yes. As I said, that was the original intention -- not the current
implementation. Once again, the code in libraries.m4 is in dire need of
some TLC. Getting it in better shape *is* on my agenda, but it tends to
be pushed down all the time.
So from my point of view, Omair's patch is good. It provides additional
value. It does not solve all problems in libraries.m4, nor is it the
complete answer on how --with-libpng will behave in the future. But it
is a good step on the way. I'm willing to sponsor the patch. But I won't
do that if you object to accepting it. (After all, from Oracle's point
of view there's no real need for this patch.) I'm also not very much
interested in working out this specific patch to perfection, when the
whole of libraries.m4 needs so much work. So, it's a bit of "take it or
leave it".
To be extremely clear: Andrew, do you object to bringing Omairs patch,
as it is, into OpenJDK?
/Magnus