PR 18501. Before I mark it WONTFIX.

[Nathan Sidwell]

Diego Novillo wrote:

> It will boil down to what we want to prioritize: false positives or 
> false negatives.  I'm almost tempted to prioritize false positives (ie, 
> give a warning)

this is my inclination too.

>, but that would break uninit-5.c which seems to be a 
> common case:

the reported case is more obviously an uninitialized use than the uninit-5.c
case.  I.e. every path into the loop body results in an uninitialized
use, whereas the uninit-5 case not every path into the second if is an
uninitialized use.  Furthermore, gcc used not to be able to detect the
uninit-5 case, but could detect the reported case -- so in 'enhancing'
uninit-5 we get a regression.

nathan

-- 
Nathan Sidwell    ::   http://www.codesourcery.com   ::     CodeSourcery LLC
nathan@codesourcery.com    ::     http://www.planetfall.pwp.blueyonder.co.uk