Backports to 4.2

[Steve Kargl]

On Sat, Jun 16, 2007 at 03:37:58PM -0700, Jerry DeLisle wrote:
> Daniel Franke wrote:
> >
> >As Mark plans to release 4.2.1 in mid-July, I would like to see whether 
> >there are any of my patches to trunk that need backporting to the 
> >4.2-branch. Some of the associated PRs are still open an could be closed 
> >either way. These are:
> >
> >	PR 22359 - FSEEK intrinsic (open)
> >	PR 31760 - missing elemental applicability (closed)
> >	PR 31919 - min/max do not check array conformance (open)
> >	PR 31929 - atan2 accepts non-conforming array shapes (open)
> >	PR 31930 - bes[jy]n intrinsics not elemental (closed)
> >	PR 32001 - MAX intrinsic with large number of arguments (closed)
> >	PR 32002 - insufficient conformance check (open)
> >
> >Of these, only PR22359 and PR32002 represent regressions (g77 and 4.2 
> >respectively), the others are bug-fixes. Nonetheless, as FSEEK was sort of 
> >intrusive, I'd like to omit this in backporting (demand is low anyway) but 
> >would ike to backport all others as they are generally quite simple and 
> >correct some oversights that may be annoying for the users.
> >
> >Ok to proceed?
> >
> Only the regressions should be backported.  The rest can be closed.  You 
> are approved on the two regressions (provided testing completed on the 
> branch of course).
> 

I agree with Jerry, here.  Even though this may be simple to backport,
I think we (the gfortran hackers) should concentrate on trunk (and
the F-E branch).

You do have the option of doing the backport and doing the necessary
regression testing and then ask markm for approval to commit the
non-regression patches.  markm is expecting gfortran to follow the
normal release branch rules.

-- 
Steve