On 07/21/2011 08:13 AM, Steve Kargl wrote:
There is generate_warning(). If one can disable this extension vai -std=f95, then one should be able to emit a warning
But one does not have to. While for the compiler, it might be okay and often is useful to emit warnings, having warnings printed from the library is much less useful. Either the output gets drowned in all the other output such that no one sees it - or it breaks scripts which do not expect this extra output. Especially, the programmer might not use input which causes the library to emit warnings - and the user may not have the source code to fix it.
I think the extention, which was removed over a year ago, should not be restored
And I disagree. The extension itself was not removed - there are even testsuite testcases which test for it.* However, as it turned out, there exists namelists which are valid according to the documented GNU extension but are rejected. - That's not new given how many constructs namelists allow.
Additionally, it is not a year ago but not even 9 months - and less than 4 month since 4.6.0 which was the first GCC release with the bug. You cannot expect that every user uses unstable versions or immediately updates to a released version. Having someone using 4.6.0 after less than 4 months is already very timely.
Additionally, as mentioned before, except for NAG all 7 of the compilers I tested plus gfortran 4.1 to 4.6 work (with 4.4 to 4.6 failing after 2010-10-26/2010-11-03). Furthermore, it is documented in gfortran's manual and it is not a very invasive patch.
Tobias * For instance namelist_24.f90
