Eliminate write-only variables

[Joseph S. Myers]

On Sun, 18 May 2014, Sandra Loosemore wrote:

> Hmmmm, I'm guessing this was some concern about invalid code motion around a
> setjmp.  Our original analysis document lists "F does not call setjmp" as a
> requirement for the optimization, so this was probably a case where we were
> being excessively conservative.

Reading my internal analysis from 24 Nov 2005, it was in terms of 
converting a function-local static variable to automatic.  The concern may 
well have been about "the values of objects of automatic storage duration 
that are local to the function containing the invocation of the 
corresponding setjmp macro that do not have volatile-qualified type and 
have been changed between the setjmp invocation and longjmp call are 
indeterminate" meaning that program that's valid when the variable is 
static can become invalid on converting it to automatic.  Of course that 
depends on the particular uses of the variable (whether it's possible for 
it to be changed between the two calls); I think we can presume the 
existing optimizations (that deal with everything except the final store) 
already deal with keeping the transformations valid in the presence of 
setjmp / longjmp.

-- 
Joseph S. Myers
joseph@codesourcery.com