Undetected error in array declarations...

[Mike Simons]

Thanks for the speedy reply...

Alexandre Oliva <oliva@dcc.unicamp.br>, 03 Dec 1998 10:29:03 -0200:
>On Dec  2, 1998, Mike Simons <msimons@fsimons01.erols.com> wrote:
>>   If a function call is given as a array size, 
>
>>   I think the compiler should flag this as an error... or if for some reason
>> that may be valid syntax, at the very least issue a warning when -Wall is on.
>
>Try -pedantic.  This is a gcc extension.

  Sorry for the false bug report...

  I "knew" that variable sized arrays were illegal.  I "knew" that 
array sizes had to be know at compile time.  I was surprised when 
I saw some code using this and that it worked, and shocked that it
wasn't even a warning with powerful "-Wall -W".  ;)
  This isn't a invalid, in the GNU dialect of C. 


  I read around the egcs/gcc info pages and now understand that the
'-Wall -W' combo doesn't actually issue warnings for _ANSI_ C problems...
or even portability issues, I thought it did.  I also thought that 
variable sized arrays are a "danger", and they are, only if used 
badly (but so is everything else) or in code that needs to compiler
independent.  I may start using them in the future.
  To get as close as possible to ANSI in addition to most useful 
warnings, the correct flags would be "-Wall -W -pedantic -ansi". 


  However, I think there is a bug in this attitude:

(from the info pages, Node: Warning Options)
:     Some users try to use `-pedantic' to check programs for strict ANSI
:     C conformance.  They soon find that it does not do quite what they
:     want: it finds some non-ANSI practices, but not all--only those
:     for which ANSI C *requires* a diagnostic.
:
:     A feature to report any failure to conform to ANSI C might be
:     useful in some instances, but would require considerable
:     additional work and would be quite different from `-pedantic'.  We
:     recommend, rather, that users take advantage of the extensions of
:     GNU C and disregard the limitations of other compilers.  Aside
:     from certain supercomputers and obsolete small machines, there is
:     less and less reason ever to use any other C compiler other than
:     for bootstrapping GNU CC.

  To a certain extent I think the attitude "take advantage of our 
extensions and disregard other products" is bad.  The second paragraph 
sounds very close to "embrace, extend, extinguish" but from a very 
different source (using public standards (ANSI C), adding extensions, 
encouraging users to rely on the extensions ;).  Standards are good, 
extensions are dangerous but only when they can not be disabled.
  While variable sized arrays can be flagged as a warning or error,
that excerpt says to me, that there are extensions have no warnings,
that can not be disabled, and the combination of these is not considered
a problem.  I think that extensions that silently creep in are evil.
  I think there should be an easy option to force ANSI compliance 
and disable all GNU extensions, or at the very least issue a warning
about any extensions in use.  So that people who choose to (for whatever
reason) have the option to detect problems.  But I don't have the 
knowledge to add this option, and I doubt many people who do see any 
value in it.

    Just my opinion,
      Mike Simons