[patch, fortran] more about namelists

Daniel Franke franke.daniel@gmail.com
Tue Jul 31 09:45:00 GMT 2007


A recent patch of mine [1] introduced attribute bits for pointer and private 
components of derived types respectively. It was only shortly after I 
committed this patch, that I found the functions derived_pointer() and 
derived_inaccessable() in resolve.c, obviously meant for the same purpose as 
the attributes just introduced ...

Further, while checking the usage of the functions mentioned to see whether 
they could be savely removed, I learned that the changes I made to fix the 
known problems with namelists introduced new ones (see the changed testcases 
for examples).

As a sidenote: with this patch, gfortran will be the only compiler I tested 
(intel, sun, lahey) that detects the nested private components in 
namelist_33.f90 correctly - for a given value of correct. I would be grateful 
if someone could please double-check this, thanks :)


:ADDPATCH fortran:

2007-07-29  Daniel Franke  <franke.daniel@gmail.com>

	* resolve.c (derived_pointer): Removed, replaced callers by access 
	to appropiate attribute bit.
	(derived_inaccessable): Shortcut recursion depth.
	(resolve_fl_namelist): Fixed checks for private components in namelists.


Bootstrapped and regression tested on i686-pc-linux-gnu.
Ok for trunk?

Regards
	Daniel

[1] http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02016.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: namelist-2.patch
Type: text/x-diff
Size: 6109 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20070731/4d7257dd/attachment.bin>


More information about the Fortran mailing list