[Bug fortran/43899] New: Wrong unused-variable warning with NAMELISTs

[Bug fortran/43899] New: Wrong unused-variable warning with NAMELISTs

[burnus at gcc dot gnu dot org]

Found at
http://groups.google.com/group/comp.lang.fortran/browse_thread/thread/4e34a377097524f2

gfortran gives a -Wall warning for a variable which is only used via namelists:
 character(4)  :: nml_string
                            1
Warning: Unused variable 'nml_string' declared at (1)

For:

  character(4)  :: nml_string
  namelist /nmlist/ nml_string
  read(7,nml=nmlist)
  write(*,nml=nmlist)
  end


-- 
           Summary: Wrong unused-variable warning with NAMELISTs
           Product: gcc
           Version: 4.5.0
            Status: UNCONFIRMED
          Keywords: diagnostic
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: burnus at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu dot org]

------- Comment #1 from jvdelisle at gcc dot gnu dot org  2010-04-26 19:05 -------
Well in a sense it is unused.  Regardless, adding a warning for the truncated
string, the real issue, is straightforward and I will do so.


-- 

jvdelisle at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |jvdelisle at gcc dot gnu dot
                   |dot org                     |org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2010-04-26 19:05:23
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[burnus at gcc dot gnu dot org]

------- Comment #2 from burnus at gcc dot gnu dot org  2010-04-26 20:16 -------
(In reply to comment #1)
> Well in a sense it is unused.

Well, but only in a sense - in the example a value is assigned to and then
printed ...

> Regardless, adding a warning for the truncated
> string, the real issue, is straightforward and I will do so.

It might be straight forward, but it needs to be guarded by a special option as
it is 100% valid code. I think printing a run-time warning by default is wrong.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu dot org]

------- Comment #3 from jvdelisle at gcc dot gnu dot org  2010-05-04 01:56 -------
What I am thinking of is a warning if a quoted string is terminated by an
end-of-line and there is no closing quote.

I would like to put this behind -Wcharacter-truncation which will be picked up
with -Wall. Let's see what I can come up with.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu dot org]

------- Comment #4 from jvdelisle at gcc dot gnu dot org  2010-05-04 04:11 -------
What about this?

$ gfc -fbounds-check pr43899.f90 
$ ./a.out 
 &NMLIST NML_STRING='123456789' /   
At line 9 of file pr43899.f90 (unit = 7, file = 'example.nml')
Fortran runtime error: Namelist object 'nml_string' exceeds length of character
variable


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[burnus at gcc dot gnu dot org]

------- Comment #5 from burnus at gcc dot gnu dot org  2010-05-04 05:53 -------
(In reply to comment #4)
> What about this?
> 
> $ gfc -fbounds-check pr43899.f90 
> $ ./a.out 
>  &NMLIST NML_STRING='123456789' /   
> At line 9 of file pr43899.f90 (unit = 7, file = 'example.nml')
> Fortran runtime error: Namelist object 'nml_string' exceeds length of
> character variable

I like the message, but I do not like a run-time error with -fbounds-check;
while it is good to have an easy-to-reach diagnostic like that, so far
-fbounds-check only aborted if the program was invalid according to the Fortran
standard. Here, the code probably has a problem, but the usage is valid
according to the Fortran standard.

Thus, I would like to see either only a warning or a different compile-time
flag.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

--- Comment #6 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-03 04:29:19 UTC ---
OK, I got back to this.  I will submit a patch shortly.

I now do this:

At line 9 of file pr43899.f90 (unit = 7, file = 'example.nml')
Fortran runtime warning: Namelist object 'nml_string' truncated on read.

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

--- Comment #7 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-03 15:22:29 UTC ---
Author: jvdelisle
Date: Wed Nov  3 15:22:25 2010
New Revision: 166252

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166252
Log:
2010-11-03  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR libgfortran/43899
    * runtime/error.c (generate_warning): New function to generate a run
    time warning message. Fix some whitespace.
    * libgfortran.h: Add prototype for new function.
    * io/list_read.c (nml_read_obj): Use new function to warn when a
    character namelist object is truncated.  Only warn if compiled
    with -fbounds-check.

Modified:
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/io/list_read.c
    trunk/libgfortran/libgfortran.h
    trunk/libgfortran/runtime/error.c

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

--- Comment #8 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-03 15:27:53 UTC ---
Author: jvdelisle
Date: Wed Nov  3 15:27:48 2010
New Revision: 166253

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166253
Log:
2010-11-03  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR libgfortran/43899
    * gfortran.dg/namelist_67.f90: New test.

Added:
    trunk/gcc/testsuite/gfortran.dg/namelist_67.f90
Modified:
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

--- Comment #9 from Tobias Burnus <burnus at gcc dot gnu.org> 2010-11-03 15:44:35 UTC ---
The patch in comment 7 and 8 adds a run-time warning (with -fcheck=bounds) if a
string in a namelist gets truncated.

TODO: Comment 0, namely: Don't warn that a variable is unused, if the namelist
to which they belong is used.

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

--- Comment #10 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-03 17:47:51 UTC ---
This patchlet gets rid of the bogus warning.  All that remains is to check that
namelists are used or not.  Also note that we do not now check to see if
commons are used either.


Index: trans-decl.c
===================================================================
--- trans-decl.c    (revision 166182)
+++ trans-decl.c    (working copy)
@@ -4015,9 +4015,10 @@ generate_local_decl (gfc_symbol * sym)
     }

       /* Warn for unused variables, but not if they're inside a common
-     block or are use-associated.  */
+     block, a namelist, or are use-associated.  */
       else if (warn_unused_variable
-           && !(sym->attr.in_common || sym->attr.use_assoc || sym->mark))
+           && !(sym->attr.in_common || sym->attr.use_assoc || sym->mark
+            || sym->attr.in_namelist))
     gfc_warning ("Unused variable '%s' declared at %L", sym->name,
              &sym->declared_at);

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

--- Comment #11 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-09 00:08:24 UTC ---
Author: jvdelisle
Date: Tue Nov  9 00:08:20 2010
New Revision: 166461

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166461
Log:
2010-11-08  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR fortran/43899
    * trans-decl.c (generate_local_decl): Do not generate unused warning
    for variables in namelists.

Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/trans-decl.c

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

Jerry DeLisle <jvdelisle at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|jvdelisle at gcc dot        |unassigned at gcc dot
                   |gnu.org                     |gnu.org

--- Comment #12 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-09 00:24:02 UTC ---
This latest patch gets rid of the bogus unused warning.  Remaining, but not
real important is to warn on an unused namelist.  I think this will involve
marking namelists as used when resolving or translating I/O statements and then
at some point, scan for any namelists not marked.  I almost think its not worth
it, but will leave this opened until others weigh in on it.

Even if a namelist is used, a particular variable within it may not be used. 
For example, if the users nml file does not contain an assignment for a
particular variable in the list, it will remain unused.  No value in the read
operation.

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

Jerry DeLisle <jvdelisle at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |WAITING
                 CC|                            |jvdelisle at gcc dot
                   |                            |gnu.org

--- Comment #13 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-09 00:25:11 UTC ---
Waiting

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|---                         |WONTFIX

--- Comment #14 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> This latest patch gets rid of the bogus unused warning.  
> Remaining, but not real important is to warn on an unused namelist.  
> I think this will involve marking namelists as used when resolving 
> or translating I/O statements and then at some point, scan for any 
> namelists not marked.  I almost think its not worth it, but will 
> leave this opened until others weigh in on it.
>
> Even if a namelist is used, a particular variable within it may 
> not be used.  For example, if the users nml file does not contain 
> an assignment for a particular variable in the list, it will remain 
> unused.  No value in the read operation.

No activity since over two years. Closing as WONTFIX. Please reopen if I am
wrong.

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|WONTFIX                     |---

--- Comment #15 from Tobias Burnus <burnus at gcc dot gnu.org> ---
(In reply to Jerry DeLisle from comment #12)
> Remaining [...] is to warn on an unused namelist.  I think this will involve
> marking namelists as used when resolving or translating I/O statements and
> then at some point, scan for any namelists not marked.

I think the later scanning can be done when generating the DWARF debug symbols
(cf. PR37132) in trans-decl.c. While not very important, I find such a warning
useful. (One needs to be careful about those declared in a module or accessed
via host association. Fortunately, namelists may not be declared in a BLOCK.)

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

--- Comment #16 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
(In reply to Tobias Burnus from comment #15)
> (In reply to Jerry DeLisle from comment #12)
> > Remaining [...] is to warn on an unused namelist.  I think this will involve
> > marking namelists as used when resolving or translating I/O statements and
> > then at some point, scan for any namelists not marked.
>
> I think the later scanning can be done when generating the DWARF debug 
> symbols (cf. PR37132) in trans-decl.c. While not very important, I find 
> such a warning useful. (One needs to be careful about those declared 
> in a module or accessed via host association. Fortunately, namelists 
> may not be declared in a BLOCK.)

If there is a real need for a warning for unused namelist or namelist items,
then a new PR should be open and this one closed. IMO it is just a waste of
time: unused stuff having a strong tendency to produce false positive.

[Bug fortran/43899] Wrong unused-variable warning with NAMELISTs

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=43899

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|---                         |WONTFIX

--- Comment #17 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> If there is a real need for a warning for unused namelist or namelist items,
> then a new PR should be open and this one closed. IMO it is just a waste of time:
> unused stuff having a strong tendency to produce false positive.

No feedback for more than two years. Closing as WONTFIX. Please open new PR(s)
for remaining issue(s).