[geeks] Compilation warnings...

[Michael C. Vergallen]

Jonathan C. Patschke wrote:
>
>
> If restaurants cooked food the way gcc validates code, a "well done"
> steak would've passed through a kitchen while the grille was turned up,
> but you wouldn't be assured of much more than that.
>
>   
Nice analogy ... however not completely true ... mostly it would be the 
cook that was at fault because
he  forgot to put it trough the grill....or the  cook  doesn't know how 
to operate the grill ...a third possibility
is that the cook is trying to hide the fact that he is a bad cook...
 
> GCC isn't just poor at making sure code adheres to even the loosest
> interpretations of the language standards, but also extends the language
> in ways that create entirely new[0] language constructs that aren't
> valid on other compilers.  I suspect the problem is merely that the CUPS
> maintainers hadn't bothered to run their code through someone else's
> compiler, and GCC didn't have the presence of mind to issue that
> particular warning[1].
>   
It was the GNU TLS package & sub-packages that was at fault ...the 
programmers admit that they only work
on Linux with GCC.  I half expected it to be ... the problem with 
programmers  who are only exposed to one
Unix flavor is that they don't seem to understand that what works for 
them doesn't mean it is correct coding
practice that will deliver portable code...CUPS uses those libraries so 
I have to compile them... It seems that
most G.P.L packages need 100 sub-packages to compile....is this because 
instead of using a routine form an
other library in their code they just link the library even if the 
routine could be implemented without the use of
someone else's code...I start to wonder ...because even gnupg  now needs 
other sub-packages to compile...all
by the same authors so they could have  bundled  their work  so you 
don't have stop & fetch some more code
every 5 minutes...just to find out that by the time you have built one 
package you need and you start on the next
package  you require  you can start all  over again because  the 
sub-packages  you  build  are outdated and the
next package requires just one feature that was not in the previous 
version ...makes me wonder  if the  whole
Object Oriented  approche to coding is really that beneficial...
> The Good And Proper thing to do would be to fix the code, either by
> bandaging it with a typecast or by changing the variable type.  However,
> as warnings go, that one probably -is- innocuous, unless the most
> significant bit of that integer is likely to be flipped.
>   
I'll have a look at those three packets of code .... that make up the 
GNU TLS library...and try and figure out what was done
badly...
> [0] Like the way it handles partial initializers.
> [1] Or, as seems too often the case, the maintainers released the code
>      because it compiled with warnings instead of errors.