Re: elog/ereport noreturn decoration

On 29 June 2012 22:35, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> IOW I am not on board with reducing
> the number of warnings in clang by increasing the number everywhere
> else.

I successfully lobbied the Clang developers to remove some warnings
that came from certain sites where we use a single element array at
the end of a struct as pseudo flexible arrays (Clang tried to do
compile-time bounds checking, with predictable results). Now, I had to
make a nuisance of myself in order to get that result, but you might
consider trying that.

I have been using the Clang static analyser some. I've seen all those
false positives. The greater problem is that the analyser apparently
won't work across translation units.

--
Peter Geoghegan       http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training and Services