Build-Breaking MingW 3.3 Stuff (was Published Sources & Scripts for MingW GCC 3.3 Build)

Andrew Haley aph@redhat.com
Sun Apr 13 17:58:00 GMT 2003


Mohan Embar writes:
 > Hi Andrew,
 > 
 > >Well, we seem to be stuck.  If the MingW build is to work, someone
 > >with 
 > >
 > >a.  a 3.3 tree 
 > >b.  a Windows box 
 > >c.  commit access
 > 
 > I'm new to this, so can you fill me in on why one would need a Windows
 > box for this? I was the author of this patch and tested it on MingW and
 > Linux before submitting it. Michael found it to be correct and committed
 > it to the trunk. (The code in question is placeholder, do-nothing code.)

Okay.  That sounds like it's fairly safe.

At this stage in the 3.3 release process we need to be careful about
what we do.

A maintainer should apply a patch, do a full regression test, and then
commit the change.  It's not possible to do a full regression test on
a Windows specific patch without Windows.

The only other way to do it is to commit the patch untested and then
have someone else do the regression test.  This is not generally good
practice, especially on a stabilization branch.

Andrew.



More information about the Java mailing list