| Dan Nagle 2007-11-27, 8:09 am |
| Hello,
Janne Blomqvist wrote:
> First, gfortran didn't simply appear out of the vacuum as "just" a
> free software Fortran 95 compiler. It also fulfills the role of g77
> replacement (which in turn was designed to replace the Unix f77). So,
> in order to provide g77/f77 compatibility, it does feature some
> "Unixisms". Now, I personally think that changing the default
> behaviour towards standard conformance, if necessary by breaking
> backwards compatibility, is a good thing. But certainly there are
> others who feel differently, and they will soon start to complain how
> gfortran 4.3 breaks their old working programs that happened to rely
> on C-style escape handling (as Steve already hinted at in another
> message). This change in behaviour is(/will be?) of course documented,
> but as usual, few people bother to read the release notes.
I'd like to thank the gfortran developers for changing the default
in response to being criticized in this newsgroup. It takes some
courage to respond to public criticism by changing.
I think the change is a Good Thing, Dick has outlined the proof
of the Theorem that backslash escapes violate even the f95 standard.
(I know there's a counterargument, but I don't like it.) And f03
is completely unambiguous here.
But backwards compatibility is also important, and some will use
the inconvenience of reading release notes as a good excuse
for complaint.
So, Thanks, developers! Thanks for past efforts and the risk
of future complaints.
--
Dan Nagle
Purple Sage Computing Solutions, Inc.
|