[fpc-devel] Data flow analysis (dfa) and "case ... of"

[Denis Kozlov]

On 05/06/2017 19:49, Jonas Maebe wrote:
> The DFA simply does not take into account the result of 
> case-statements at all right now. If you have a "case bytevar of" with 
> every possible value from 0 to 255, you will still get the warning. 

On 05/06/2017 19:43, Sven Barth wrote:
> They are legal in the sense that they result in undefined behavior and 
> in that case the compiler is free to act as it sees fit.

By covering all possible cases I meant including "else" condition in all 
case statements, which either defaults to some acceptable value or 
raises an exception to highlight an invalid value. This is just a 
standard practice for me, so my code never produces "uninitialized 
result" warnings, unless I actually made a mistake by not covering all 
possible cases, in which case I would really like to see that warning no 
matter what. I prefer to see a controlled exception instead of having an 
invalid value producing a pseudo-valid semi-random result which may be 
much more difficult to trace later.

Complier currently warns (cares) about such execution flows which are 
triggered by invalid data, and I happen to find it useful :)

However, I do understand that compiler can rightfully not care about 
invalid data, hence, choose not to issue the "uninitialized result" 
warning if all valid values have been enumerated.

Denis