On Feb 25, 12:03am, John Darrow wrote:
} John Nemeth <jnemeth@victoria.tc.ca> wrote:
} >On Feb 8,  1:19am, John Darrow wrote:
} >} Ty Sarna <tsarna@netbsd.org> wrote:
} >} >
} >} >Module Name:	pkgsrc
} >} >Committed By:	tsarna
} >} >Date:		Sun Oct 22 20:04:30 UTC 2000
} >} >
} >} >Removed Files:
} >} >	pkgsrc/print/py-reportlab/files: patch-sum
} >} >	pkgsrc/print/py-reportlab/patches: patch-aa patch-ab
} >} >
} >} >Log Message:
} >} >nuke no longer needed patches
} >} 
} >} Shouldn't we keep around a placeholder patch-sum file so that people who
} >} update pkgsrc without removing it first (e.g. extract from a new tar file,
} >} or sup without the 'delete' flag) won't end up with leftover patches trying
} >} to apply themselves?
} >
} >     The leftover patches don't try to apply themselves.  You get a
} >warning message saying that is not in patch-sum was found and that it
} >isn't being applied.
} 
} But if they've supped without delete, or extracted a tar, they'd still have
} _both_ the obsolete patches _and_ the obsolete patch-sum, which lists said
} patches.  A placeholder patch-sum fixes this, and was created for this very
} reason.

     I was going to say that there is only one patch-sum file, so there
can't be an obsolete one left lying around.  However, I see that in
this case, since all the patches were deleted, the patch-sum was also
deleted.  Yes, I would agree that an empty placeholder patch-sum file
should be kept around.  I often untar a new pkgsrc tarball without
first deleting the old pkgsrc tree, so I would have been bitten by
this.  Yuck!

}-- End of excerpt from John Darrow