Michael Meskes <meskes@postgresql.org> writes:
> On Sun, Nov 24, 2013 at 09:53:29PM -0500, Stephen Frost wrote:
>> I tend to agree with Tom here but I feel like I'm missing something-
>> will this only make things "work" which would fail previously..? We
>> certainly don't want to risk breaking things for existing users, but
>> perhaps it's not possible to have existing code which works and this is
>> only 'loosening' of what we accept.
> We're not exactly loosening but we aligning with what the C compiler accepts,
> if that makes any sense.
What I was concerned about was the fear that previously-working ecpg
programs would be broken by this change. I think what you're saying
is that any such program would have failed anyhow at the compilation
stage --- is that right? If so I withdraw the complaint.
regards, tom lane