This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Fwd: Re: Fortran Parser -- last call]


--- Begin Message --- Hi Tobias,


No, the output format is not specified anywhere. These ideas are new, they are mine ( well I give them to you if you want them :-). Now the point is that XML is very easy to use, a lot of libraries are available to make some search through it, or modify it.
Of course, I know that the parser is something which is evolving all the time, so if you add a new feature and I have written its counterpart in the XML output, something will go wrong if somebody tries to dump XML on some code using this feature. Otherwise regular compile will be OK.


Regards,

Philippe

Tobias Schlüter a écrit :

Hi,


Daniel Franke wrote:
On Wednesday 03 December 2008 17:47:32 Philippe Marguinaud wrote:
I would like to have an answer about whether it is OK or not to work on
the gfortran parser. If the answer is yes, please tell how I should
proceed.

Philippe,


I've followed your messages up until now and am still not 100% sure what you propose.
Nobody will "disallow" you to work on gfortran's parser. I'm sure, everybody would appreciate it if it would get some additional love. Thus, if you wish to bring in your experience in parser-writing and improve the existing parser, get the paperwork with the FSF started and dig in. Your patch(es) will be reviewed as any others and happily be accepted (or maybe not). Bugzilla currently has 99 diagnostic-related PRs [1], those would probably be a good starting point.


Also, as Paul suggested in earlier mail [2], improving the parse-tree-dump, for example an XML-output option, would be highly welcome as well. However, replacing the internal tree of gfortran by an XML-based representation is, most likely, not going to happen.

One thing one should keep in mind is that the XML output code will inevitably bitrot, as new features will get added to the compiler and not be added to the XML dumper. I don't know if that is a bad thing, and if that's a reason not to pursue it. The other thing that I'm worried about is the output format. Is this specified somewhere? Is it following some standard used in other software (I mean beyond the level of saying "hey it's XML")


Regards,
- Tobi




--- End Message ---

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]