Goodbye

[FX]

Dear all,

Both short-term and long-term modifications to my plans mean that I  
have had very little to no time to devote to gfortran in the past, and  
that I don't expect to have more time in the one year or two to come.  
Although I'm glad to have had such an amazing experience as gfortran  
user, contributer, and then maintainer, I feel it's time to say  
goodbye and to move to other. I'm sorry I have to do it in this  
exciting period where a lot of new features are making it into  
gfortran. I'm also sorry it comes at a time where time to design plans  
and review patches seems sorely needed (but I'm sure you already  
noticed that I haven't had time to participate much in recent weeks).

I've checked that there are no unposted, exciting patches in my local  
trees (sorry again!). Here is a short list of things I said I intended  
to work on, so that other maintainers can pick them up if they want  
to, or list them somewhere not to loose the ideas (but most of them  
have a PR anyway):

   -- Fix multiple decls per function: I've been defeated by this one,  
taking the approach of fetching decls after they're created. The best  
way to fix this is to use the other approach: file-scope resolution.

   -- Handle GNU attributes (including STDCALL): the more I think  
about it, maybe it's not such a priority. It's not hard to do at all  
(a good junior job, I think), but as it's a new language extension  
that has to be designed, it requires some thinking beforehand. 	My  
first patch was poor, but it can be found at http://gcc.gnu.org/ml/fortran/2008-03/msg00006.html 
  . My opinion on this is that we should not try to support Intel's  
extensions, but go with our own "!GNU$ ATTRIBUTES attrib1, attrib2,  
attrib3 :: symbol" syntax to allow use of GCC attributes, just like C  
does.

   -- Allow for reading of UTF-8 (and other encodings) source files.  
This one should have been made easier by the integration of libcpp  
done by Daniel (Franke), and I have not done any actual work on it. It  
just feels a shame that we are able to deal so well with charsets and  
UTF-8, and yet can only read \u-escaped source files and not UTF-8 ones.

   -- There is some not-very-nice-stuff in my wide character patch to  
allow assignment of character variables of different kinds (the  
infamous __convert_s1_s4), but it appears that my reading of the  
standard was incorrect and this isn't allowed after all. That is a  
rather simple job, I think.


About the packages I am building and hosting: the i686-linux are  
automatically built and require not intervention, so I'll leave it as  
is; as I use the MacOS ones for myself, I'll continue to provide them;  
I will not contribute any Windows builds any more, though, as I don't  
have access to a Windows machine any more. I'll be happy to send my  
Windows build scripts and setup and tips to anyone who asks; in recent  
years, it's been rather smooth to use. As for hosting on  
quatramaran.ens.fr, no change is expected.

I have unsubscribed from the fortran mailing-list and the Fortran  
google group. I have also opted out of all mail from the bugzilla. If  
you need help or advice on something I built or broke with one of my  
patches, please feel free to send me a mail (though I don't promise  
fast response time).

Well, this has turned into quite a long mail, so I'll end it here.  
Thanks again to anyone in the community and the developer team, it's  
been from the beginning to the end wonderful working with you.

Kind regards, and all the best!
FX