[Bug gcov-profile/41359] New: [fortran] 'else if' statement is discarded and incoherent line count

[Bug gcov-profile/41359] New: [fortran] 'else if' statement is discarded and incoherent line count

[Edouard dot Canot at irisa dot fr]

Considering the following fortran code:

=== begin code ===
program main

   implicit none

   integer :: a = 7

   if( a == 0 ) then
      print *, "a is null"
   else if( a > 0 ) then
      print *, "a is positif"
   else
      print *, "a is negatif"
   end if

end program
=== end code ===

compiled with the "-fprofile-arcs -ftest-coverage" option flags,
linked with the same flags, with gfortran-4.3.4 or gfortran-4.4.1

After a single execution, the profile file is built as usual with:
$ gcov main

The resulting profile is then:

$ cat main.f90.gcov
        -:    0:Source:main.f90
        -:    0:Graph:main.gcno
        -:    0:Data:main.gcda
        -:    0:Runs:1
        -:    0:Programs:1
        1:    1:program main
        -:    2:
        -:    3:   implicit none
        -:    4:
        -:    5:   integer :: a = 7
        -:    6:
        3:    7:   if( a == 0 ) then
    #####:    8:      print *, "a is null"
        -:    9:   else if( a > 0 ) then
        1:   10:      print *, "a is positif"
        -:   11:   else
    #####:   12:      print *, "a is negatif"
        -:   13:   end if
        -:   14:
        -:   15:end program

First remark: line 7 is marked '3' instead of '1'

Second remark: line 9 is discarded (marked '-') instead of '1'





When using the current snapshot of gfortran-4.5 (2009-09-15), the
result is slightly different:

$ cat main.f90.gcov
        -:    0:Source:main.f90
        -:    0:Graph:main.gcno
        -:    0:Data:main.gcda
        -:    0:Runs:1
        -:    0:Programs:1
        1:    1:program main
        -:    2:
        -:    3:   implicit none
        -:    4:
        -:    5:   integer :: a = 7
        -:    6:
        2:    7:   if( a == 0 ) then
    #####:    8:      print *, "a is null"
        -:    9:   else if( a > 0 ) then
        1:   10:      print *, "a is positif"
        -:   11:   else
    #####:   12:      print *, "a is negatif"
        -:   13:   end if
        -:   14:
        2:   15:end program

Note that line 7 is counted twice, as well as the last line !


-- 
           Summary: [fortran] 'else if' statement is discarded and
                    incoherent line count
           Product: gcc
           Version: 4.3.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: gcov-profile
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: Edouard dot Canot at irisa dot fr
 GCC build triplet: i686-linux-gnu
  GCC host triplet: i686-linux-gnu
GCC target triplet: i686-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41359

[Bug gcov-profile/41359] [fortran] 'else if' statement is discarded and incoherent line count

[jv244 at cam dot ac dot uk]

------- Comment #1 from jv244 at cam dot ac dot uk  2009-09-15 13:07 -------
As you can see with

gfortran -fdump-tree-original-lineno test.f90

which leads to garbage line numbers in 

gfortran -fdump-tree-gimple-lineno test.f90

I believe there is some similarity to  PR40823

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40823

I guess that this case could be a regression with respect to g77, but can't
test it.


-- 

jv244 at cam dot ac dot uk changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2009-09-15 13:07:14
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41359

[Bug fortran/41359] Wrong line numbers for debugging/profiling

[jv244 at cam dot ac dot uk]

------- Comment #2 from jv244 at cam dot ac dot uk  2010-04-10 19:04 -------
still present in 4.6. The issue seems to be missing location info for the
nested if [if (a>0) ], the missing info in the original dump appears as a
incorrect line:7 in the gimple. It is specific to the 'else if' form, on a
single line.


-- 

jv244 at cam dot ac dot uk changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|4.3.4 4.4.1 4.5.0           |4.3.4 4.4.1 4.5.0 4.6.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41359

[Bug fortran/41359] Wrong line numbers for debugging/profiling

[mikael at gcc dot gnu dot org]

------- Comment #3 from mikael at gcc dot gnu dot org  2010-04-10 21:27 -------
With this :
diff --git a/trans-expr.c b/trans-expr.c
index 7e95ce1..a6f8616 100644
--- a/trans-expr.c
+++ b/trans-expr.c
@@ -1382,6 +1382,8 @@ gfc_conv_expr_op (gfc_se * se, gfc_expr * expr)
   else
     se->expr = fold_build2 (code, type, lse.expr, rse.expr);

+  SET_EXPR_LOCATION (se->expr, expr->where.lb->location);
+
   /* Add the post blocks.  */
   gfc_add_block_to_block (&se->post, &rse.post);
   gfc_add_block_to_block (&se->post, &lse.post);
diff --git a/trans-stmt.c b/trans-stmt.c
index 0b215f2..4b8ae69 100644
--- a/trans-stmt.c
+++ b/trans-stmt.c
@@ -718,6 +718,7 @@ gfc_trans_if_1 (gfc_code * code)

   /* Build the condition expression and add it to the condition block.  */
   stmt = fold_build3 (COND_EXPR, void_type_node, if_se.expr, stmt, elsestmt);
+  SET_EXPR_LOCATION (stmt, code->loc.lb->location);

   gfc_add_expr_to_block (&if_se.pre, stmt);



I get :
        -:    0:Source:pr41359.f90
        -:    0:Graph:pr41359.gcno
        -:    0:Data:pr41359.gcda
        -:    0:Runs:1
        -:    0:Programs:1
        1:    1:program main
        -:    2:
        -:    3:   implicit none
        -:    4:
        -:    5:   integer :: a = 7
        -:    6:
        1:    7:   if( a == 0 ) then
    #####:    8:      print *, "a is null"
        1:    9:   else if( a > 0 ) then
        1:   10:      print *, "a is positif"
        -:   11:   else
    #####:   12:      print *, "a is negatif"
        -:   13:   end if
        -:   14:
        2:   15:end program
        -:   16:


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41359

[Bug fortran/41359] Wrong line numbers for debugging/profiling

[jv244 at cam dot ac dot uk]

------- Comment #4 from jv244 at cam dot ac dot uk  2010-04-11 18:02 -------
looks like we have a patch...


-- 

jv244 at cam dot ac dot uk changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41359

[Bug fortran/41359] Wrong line numbers for debugging/profiling

[mikael at gcc dot gnu dot org]

-- 

mikael at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |mikael at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2009-09-15 13:07:14         |2010-07-28 20:01:09
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41359