This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH] Fix PR35469, revert jboolean semantics changes
On Wed, 12 Mar 2008, Andrew Haley wrote:
> Richard Guenther wrote:
> > On Tue, 11 Mar 2008, Andrew Haley wrote:
> >> Richard Guenther wrote:
> >>> I will bootstrap & test the following for the 4.3 branch and the trunk.
> >>> Andrew, is this ok with you? It reverts jboolean back to be a single
> >>> bit integer, and g++ will actually enforce that single-bitness since
> >>> 4.3. I don't know if that makes the most sense
> >> Yes, thanks. My identical patch overlapped with yours.
> >>> or if just making
> >>> it an integer of the same precision as its mode (that is, 8 bits on
> >>> any target but darwin(?)).
> >> Arrays of jboolean must always have elements one byte in size.
> >> We can just make assignment of any value other than 0 or 1 undefined.
> > Will you go ahead with your patch or shall I do it? (I can approve
> > reversion of patches of myself myself)
> Better for you to do it. I don't have Darwin, so I can't test,
> but I assume that after your patch of jboolean will have elements one byte
> in size.
I also cannot test darwin but have tested x86_64 on both trunk and
the branch. Yes, jboolean will have one byte size everywhere.