This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [gfortran patch] PR 23661
Uttam Pawar wrote:
> On Thu, 2005-09-01 at 00:58 +0200, Tobias SchlÃter wrote:
>
>>Uttam Pawar wrote:
>>
>>>diff -urN gcc.mline/gcc/fortran/parse.c gcc.mine/gcc/fortran/parse.c
>>>--- gcc.mline/gcc/fortran/parse.c 2005-08-31 10:49:55.000000000 -0700
>>>+++ gcc.mine/gcc/fortran/parse.c 2005-08-31 11:51:35.000000000 -0700
>>>@@ -249,6 +249,7 @@
>>>
>>> case 'p':
>>> match ("print", gfc_match_print, ST_WRITE);
>>>+ match ("print fmt", gfc_match_print, ST_WRITE_FMT);
>>> match ("parameter", gfc_match_parameter, ST_PARAMETER);
>>> match ("pause", gfc_match_pause, ST_PAUSE);
>>> match ("pointer", gfc_match_pointer, ST_ATTR_DECL);
>>
>>How is this supposed to work if the variable is not named "fmt"?
>
> "fmt" or "FMT" is not a variable, but an identifier for the format
> specification. The standard says that,io-control-spec-list shall not
> contain both a format and a namelist-group-name.
Funny, I thought this line
character(len=5) :: fmt = "( a)"
from the testcase you posted in the PR declares a variable.
I've already diagnosed the bug, see in the PR.
- Tobi