Merge pull request #24840 from charris/backport-24778

BUG: Fix DATA statements for f2py

numpy/f2py/crackfortran.py

@@ -1425,53 +1425,61 @@ def analyzeline(m, case, line):
             if dl.startswith(','):
                 dl = dl[1:].strip()
             ll.append([dl, il])
-        vars = {}
-        if 'vars' in groupcache[groupcounter]:
-            vars = groupcache[groupcounter]['vars']
+        vars = groupcache[groupcounter].get('vars', {})
         last_name = None
         for l in ll:
-            l = [x.strip() for x in l]
-            if l[0][0] == ',':
+            l[0], l[1] = l[0].strip(), l[1].strip()
+            if l[0].startswith(','):
                 l[0] = l[0][1:]
-            if l[0][0] == '(':
-                outmess(
-                    'analyzeline: implied-DO list "%s" is not supported. Skipping.\n' % l[0])
+            if l[0].startswith('('):
+                outmess('analyzeline: implied-DO list "%s" is not supported. Skipping.\n' % l[0])
                 continue
-            llen = len(l[1])
-            for idx, v in enumerate(rmbadname(
-                    [x.strip() for x in markoutercomma(l[0]).split('@,@')])
-                                    ):
-                if v[0] == '(':
-                    outmess(
-                        'analyzeline: implied-DO list "%s" is not supported. Skipping.\n' % v)
+            for idx, v in enumerate(rmbadname([x.strip() for x in markoutercomma(l[0]).split('@,@')])):
+                if v.startswith('('):
+                    outmess('analyzeline: implied-DO list "%s" is not supported. Skipping.\n' % v)
                     # XXX: subsequent init expressions may get wrong values.
                     # Ignoring since data statements are irrelevant for
                     # wrapping.
                     continue
-                fc = 0
+                if '!' in l[1]:
+                    # Fixes gh-24746 pyf generation
+                    # XXX: This essentially ignores the value for generating the pyf which is fine:
+                    # integer dimension(3) :: mytab
+                    # common /mycom/ mytab
+                    # Since in any case it is initialized in the Fortran code
+                    outmess('Comment line in declaration "%s" is not supported. Skipping.\n' % l[1])
+                    continue
+                vars.setdefault(v, {})
                 vtype = vars[v].get('typespec')
                 vdim = getdimension(vars[v])
-
-                if (vtype == 'complex'):
-                    cmplxpat = r"\(.*?\)"
-                    matches = re.findall(cmplxpat, l[1])
-                else:
-                    matches = l[1].split(',')
-
-                if v not in vars:
-                    vars[v] = {}
-                if '=' in vars[v] and not vars[v]['='] == matches[idx]:
-                    outmess('analyzeline: changing init expression of "%s" ("%s") to "%s"\n' % (
-                        v, vars[v]['='], matches[idx]))
-
-                if vdim is not None:
-                    # Need to assign multiple values to one variable
-                    vars[v]['='] = "(/{}/)".format(", ".join(matches))
-                else:
-                    vars[v]['='] = matches[idx]
+                matches = re.findall(r"\(.*?\)", l[1]) if vtype == 'complex' else l[1].split(',')
+                try:
+                    new_val = "(/{}/)".format(", ".join(matches)) if vdim else matches[idx]
+                except IndexError:
+                    # gh-24746
+                    # Runs only if above code fails. Fixes the line
+                    # DATA IVAR1, IVAR2, IVAR3, IVAR4, EVAR5 /4*0,0.0D0/
+                    # by expanding to ['0', '0', '0', '0', '0.0d0']
+                    if any("*" in m for m in matches):
+                        expanded_list = []
+                        for match in matches:
+                            if "*" in match:
+                                try:
+                                    multiplier, value = match.split("*")
+                                    expanded_list.extend([value.strip()] * int(multiplier))
+                                except ValueError: # if int(multiplier) fails
+                                    expanded_list.append(match.strip())
+                            else:
+                                expanded_list.append(match.strip())
+                        matches = expanded_list
+                    new_val = "(/{}/)".format(", ".join(matches)) if vdim else matches[idx]
+                current_val = vars[v].get('=')
+                if current_val and (current_val != new_val):
+                    outmess('analyzeline: changing init expression of "%s" ("%s") to "%s"\n' % (v, current_val, new_val))
+                vars[v]['='] = new_val
                 last_name = v
         groupcache[groupcounter]['vars'] = vars
-        if last_name is not None:
+        if last_name:
             previous_context = ('variable', last_name, groupcounter)
     elif case == 'common':
         line = m.group('after').strip()

numpy/f2py/tests/src/crackfortran/data_common.f

@@ -0,0 +1,8 @@
+        BLOCK DATA PARAM_INI
+        COMMON /MYCOM/ MYDATA
+            DATA MYDATA /0/
+        END
+        SUBROUTINE SUB1
+        COMMON /MYCOM/ MYDATA
+        MYDATA = MYDATA + 1
+        END

numpy/f2py/tests/src/crackfortran/data_multiplier.f

@@ -0,0 +1,5 @@
+      BLOCK DATA MYBLK
+      IMPLICIT DOUBLE PRECISION (A-H,O-Z)
+      COMMON /MYCOM/ IVAR1, IVAR2, IVAR3, IVAR4, EVAR5
+            DATA IVAR1, IVAR2, IVAR3, IVAR4, EVAR5 /2*3,2*2,0.0D0/
+      END

numpy/f2py/tests/src/crackfortran/data_stmts.f90

@@ -4,6 +4,7 @@ module cmplxdat
   integer :: i, j
   real :: x, y
   real, dimension(2) :: z
+  real(kind=8) :: pi
   complex(kind=8), target :: medium_ref_index
   complex(kind=8), target :: ref_index_one, ref_index_two
   complex(kind=8), dimension(2) :: my_array
@@ -15,4 +16,5 @@ module cmplxdat
   data medium_ref_index / (1.d0, 0.d0) /
   data ref_index_one, ref_index_two / (13.0d0, 21.0d0), (-30.0d0, 43.0d0) /
   data my_array / (1.0d0, 2.0d0), (-3.0d0, 4.0d0) /
+  data pi / 3.1415926535897932384626433832795028841971693993751058209749445923078164062d0 /
 end module cmplxdat

numpy/f2py/tests/src/crackfortran/data_with_comments.f

@@ -0,0 +1,8 @@
+      BLOCK DATA PARAM_INI
+      COMMON /MYCOM/ MYTAB
+      INTEGER  MYTAB(3)
+      DATA MYTAB/
+     *   0, ! 1 and more commenty stuff
+     *   4, ! 2
+     *   0 /
+      END

numpy/f2py/tests/test_data.py

@@ -15,6 +15,7 @@ def test_data_stmts(self):
         assert self.module.cmplxdat.j == 3
         assert self.module.cmplxdat.x == 1.5
         assert self.module.cmplxdat.y == 2.0
+        assert self.module.cmplxdat.pi == 3.1415926535897932384626433832795028841971693993751058209749445923078164062
         assert self.module.cmplxdat.medium_ref_index == np.array(1.+0.j)
         assert np.all(self.module.cmplxdat.z == np.array([3.5, 7.0]))
         assert np.all(self.module.cmplxdat.my_array == np.array([ 1.+2.j, -3.+4.j]))
@@ -26,8 +27,44 @@ def test_crackedlines(self):
         mod = crackfortran(self.sources)
         assert mod[0]['vars']['x']['='] == '1.5'
         assert mod[0]['vars']['y']['='] == '2.0'
+        assert mod[0]['vars']['pi']['='] == '3.1415926535897932384626433832795028841971693993751058209749445923078164062d0'
         assert mod[0]['vars']['my_real_array']['='] == '(/1.0d0, 2.0d0, 3.0d0/)'
         assert mod[0]['vars']['ref_index_one']['='] == '(13.0d0, 21.0d0)'
         assert mod[0]['vars']['ref_index_two']['='] == '(-30.0d0, 43.0d0)'
         assert mod[0]['vars']['my_array']['='] == '(/(1.0d0, 2.0d0), (-3.0d0, 4.0d0)/)'
         assert mod[0]['vars']['z']['='] == '(/3.5,  7.0/)'
+
+class TestDataF77(util.F2PyTest):
+    sources = [util.getpath("tests", "src", "crackfortran", "data_common.f")]
+
+    # For gh-23276
+    def test_data_stmts(self):
+        assert self.module.mycom.mydata == 0
+
+    def test_crackedlines(self):
+        mod = crackfortran(str(self.sources[0]))
+        print(mod[0]['vars'])
+        assert mod[0]['vars']['mydata']['='] == '0'
+
+
+class TestDataMultiplierF77(util.F2PyTest):
+    sources = [util.getpath("tests", "src", "crackfortran", "data_multiplier.f")]
+
+    # For gh-23276
+    def test_data_stmts(self):
+        assert self.module.mycom.ivar1 == 3
+        assert self.module.mycom.ivar2 == 3
+        assert self.module.mycom.ivar3 == 2
+        assert self.module.mycom.ivar4 == 2
+        assert self.module.mycom.evar5 == 0
+
+
+class TestDataWithCommentsF77(util.F2PyTest):
+    sources = [util.getpath("tests", "src", "crackfortran", "data_with_comments.f")]
+
+    # For gh-23276
+    def test_data_stmts(self):
+        assert len(self.module.mycom.mytab) == 3
+        assert self.module.mycom.mytab[0] == 0
+        assert self.module.mycom.mytab[1] == 4
+        assert self.module.mycom.mytab[2] == 0