On Tue, 24 Aug 2004, Mark Mitchell wrote:
Roger, I'd say that if the effecient, safe way to fix the 3.4 regression
is to make TYPE_PRECISION the width of the underlying type, and that
doesn't break anything else, that's an acceptable change, especially for
a release branch.
Sure. I'll bootstrap and regression test your 2003-07-25 patch,
together with some new testcases so that we'll know what to expect
in future.
Out of interest could you try compiling the example code given in
http://gcc.gnu.org/ml/gcc/2004-08/msg01129.html with EDG and let
me know whether it aborts or passes?