[Bug fortran/41678] Format label as second item in io control list is rejected

["dominiq at lps dot ens dot fr" <gcc-bugzilla at gcc dot gnu dot org>]

------- Comment #9 from dominiq at lps dot ens dot fr  2009-10-13 18:03 -------
> It would be interesting to get the opinion of Richard Maine.

> > ...
> > The wording in the f2003/2008 drafts I have at hand is more ambiguous: 
> > ...
> > Does a io-unit in C917 includes the optional characters UNIT=? If yes, 
> > is it correct to understand that read(unit=2, 100) is allowed in 
> > f2003/2008? 

Richard Maine's answer:

dominiq <domi...@lps.ens.fr> wrote: 
> read(unit=2, 100) is invalid under f95 because of: 
> Constraint:... [elided] 
> The wording in the f2003/2008 drafts I have at hand is more ambiguous: 


I'd disagree. I'd say it was more concise and precise in that it uses 
bnf terms instead of repeating them as textual descriptions. 
... 
> C917 (R913) If format appears without a preceding FMT=, it shall be 
> the 
>             second item in the io-control-spec-list and the first item 
> shall 
>             be io-unit. 
... 
> R901 io-unit is file-unit-number 
>              or * 
>              or internal-file-variable 
> Does a io-unit in C917 includes the optional characters UNIT=? If yes, 
> is it correct to understand that read(unit=2, 100) is allowed in 
> f2003/2008? 


No. You have misinterpreted this as illustrated by the extra word you 
added. It is a very small extra word, but an important one. C917 refers 
to "io-unit". It does not refer to "a io-unit". Note also that io-unit 
is italicized. (This doesn't show up if you have only a plain text copy, 
but the actual standard is not plain text). 
The italicized io-unit is a bnf term. You cited the definition (R901). 
That definition does not include UNIT=. The thing that includes an 
optional UNIT= is the UNIT= specifier. "Io-unit" is not a sloppy 
abbreviation for "the UNIT= specifier", which is how it looks like you 
are interpreting it. 
When the standard uses an bnf term, it means it literally - that exact 
syntax. The fact that a UNIT= specifier can have the form io-unit does 
not imply that "io-unit" is a synonym for the UNIT= specifier. 
-- 
Richard Maine                    | Good judgment comes from experience; 
email: last name at domain . net | experience comes from bad judgment. 
domain: summertriangle           |  -- Mark Twain 

So the PR can probably be closed as INVALID.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41678