https://bugzilla.redhat.com/show_bug.cgi?id=837901
has requested boost static libraries be shipped as multilib so that they can be used with "gcc -m32".
We historically haven't done this for anything besides glibc & libstdc++, but there's nothing technically preventing us from changing the policy in this manner. Opinions?
Bill
Bill Nottingham wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=837901
has requested boost static libraries be shipped as multilib so that they can be used with "gcc -m32".
We historically haven't done this for anything besides glibc & libstdc++, but there's nothing technically preventing us from changing the policy in this manner. Opinions?
+1, do it. -static probably ought to be treated the same as -devel
-- rex
On Mon, 09 Jul 2012 18:13:29 -0500, Rex Dieter wrote:
Bill Nottingham wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=837901
has requested boost static libraries be shipped as multilib so that they can be used with "gcc -m32".
We historically haven't done this for anything besides glibc & libstdc++, but there's nothing technically preventing us from changing the policy in this manner. Opinions?
+1, do it. -static probably ought to be treated the same as -devel
Static-only -devel packages with a virtual -static package name are shipped as multilib already anyway (and -static packages with a virtual -devel package name probably, too, in case mash takes a look at virtual package names).
Michael Schwendt (mschwendt@gmail.com) said:
Static-only -devel packages with a virtual -static package name are shipped as multilib already anyway (and -static packages with a virtual -devel package name probably, too, in case mash takes a look at virtual package names).
It does not (arguably it should.)
Bill
Michael Schwendt (mschwendt@gmail.com) said:
On Mon, 09 Jul 2012 18:13:29 -0500, Rex Dieter wrote:
Bill Nottingham wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=837901
has requested boost static libraries be shipped as multilib so that they can be used with "gcc -m32".
We historically haven't done this for anything besides glibc & libstdc++, but there's nothing technically preventing us from changing the policy in this manner. Opinions?
+1, do it. -static probably ought to be treated the same as -devel
Static-only -devel packages with a virtual -static package name are shipped as multilib already anyway (and -static packages with a virtual -devel package name probably, too, in case mash takes a look at virtual package names).
OK, this is done as of mash-0.5.29-1, headed to rawhide now-ish.
Bill
-
Bill Nottingham -
Michael Schwendt -
Rex Dieter