On Fri, 25 Feb 2005 07:25, Troy Rollo wrote:
On Friday 25 February 2005 03:26, Mike McCormack wrote:
Thanks for finding this problem. I think the root cause of the problem is that three structures that need to be the same require 3 seperate tweaks.
Can you confirm that this patch also fixes the problem for you?
I probably could, but given the nature of your patch there's a much more efficient test that should give much greater confidence in it - compiling the affected source files without debugging information using your patch and mine should give identical object files (except for time stamps in the object file headers).
I have done this, and except for 8 bytes that turned out to be line numbers of assert statements, the two patches compile to identical object code.