difference between g95 and gfortran 
Author Message
 difference between g95 and gfortran

 Waiting for the free gnu f95 compiler, i am a bit confused:
Are g95 and gfortran different projects? If they are, which one
of them is going to be integrated into gcc? Which one is closer to
a final release? I'm sorry for asking this here, but i was unable to find
the answer in any of the projets' sites.


Sun, 17 Dec 2006 20:38:12 GMT  
 difference between g95 and gfortran

|  Waiting for the free gnu f95 compiler, i am a bit confused:
| Are g95 and gfortran different projects?

different projects, same origin.

| If they are, which one of them is going to be integrated into gcc?

gfortran is in gcc CVS now. g95 will as I understand be offered as a
standalone compiler (it is only a 3.5MB download and will probably
not grow significantly)

| Which one is closer to a final release?

featurewise g95 is very close to be a full Fortran 95 compiler, but
lacks optimization. gfortran has optimization but lacks many Fortran
95 features (Fortran 77 is more or less covered)


Sun, 17 Dec 2006 21:06:54 GMT  
 difference between g95 and gfortran


>> Waiting for the free gnu f95 compiler, i am a bit confused:
>>Are g95 and gfortran different projects?

> Yes. Coincidentally, your question was asked on the gfortran mailing
> list on Monday.

> To quote the short explanation by Steve Kargl:

> """

> The projects at one time were both known as g95.  Andy Vaught decided
> to close the g95 tree and gfortran was forked from g95.  gfortran
> is the compiler that is now included in GCC.  Andy continues to work
> on g95 and uses a fairly old GCC development branch for the back end.

> """

This is a very brief description of what happened.

Basically (I think all of what I'm saying can be verified in mailing
list archives, cvs histories etc.), Andy wanted to retain complete
control of g95, cut everybody out from development, alienating everybody
who had worked on g95 at that time (which caused the fork), alienated
gcc's developers, started to only publish obfuscated source code, and
only at irregular intervals [1], not providing current, unobfuscated
code, even when asked to, thereby violating the GPL (the copyright on
gcc, large parts of which have found their way into g95), and developed
a fairly complete Fortran 95 frontend. Unfortunately, the original goal,
as once stated on the webpage, "the g95 project is developing a Fortran
95 frontend for inclusion in gcc" seems to have been removed someplace
along the way.

gfortran has suffered from the loss of this capable programmer, who also
has a lot of time to put into it. But still, gfortran has developed,
even though at a slower pace. It is now part of the development tree of
gcc which will become gcc 3.5, and since it has attracted a few
developers again lately, we hope to be able to bring gfortran into a
much better shape till the release. In my perception, a realistic (if
pessimistic) goal is to make gfortran a fully functional g77 replacement
by the time gcc 3.5 is released, hopefully covering most of Fortran 95.
Gcc could already have a fully functional Fortran 95 frontend if someone
who has as much time at his hands as Andy, had worked on it, together
with other people, but unfortunately, that's not the case.

If g95 decides to play fair, gfortran will move along faster, as code
transfer is legally possible. Unfortunately, during the last few days
Andy seems to have decided to accept code from people who have not
assigned their copyright to the FSF, making those portions unacceptable
for gcc (and gfortran), and without a public development history it will
be difficult to isolate them.

I sincerely hope all these troubles will be overcome.

Best regards,
- Tobias Schlter

[1] see the CVS history (which is no longer linked to from g95's
webpage) of some random file, e.g.
or look at the diffs between successive revisions:
to get an idea, of much Andy despises other people looking at the
development of g95.

Mon, 25 Dec 2006 22:03:47 GMT  
 [ 4 post ] 

 Relevant Pages 

1. automatic arrays with negative size in g95 and gfortran

2. Common statement does not work with gfortran, but g95 gives the correct answer

3. gfortran or g95

4. gfortran vs. g95

5. g95 AND gfortran problem ONLY on PPC OSX Tiger

6. Compiler flags for compiling FEM2DLiB with g95 and/or gfortran

7. problem with very small numbers - g95 versus gfortran

8. gfortran, g95, and dual-core

9. g95 versus gfortran

10. Gfortran 2 years behind G95 and still not ready for prime time

11. Vector instruction support in G95 or Gfortran?

12. recl on g95 and gfortran


Powered by phpBB® Forum Software