testsuite - typos and Cygwin problems
Tobias Schlüter
tobias.schlueter@physik.uni-muenchen.de
Thu Jan 20 22:49:00 GMT 2005
Paul Thomas wrote:
> In going more systematically thorugh the test suite than I have in the
> past, I have noticed some "funnies":
>
> (1) There are some typos in /gfortran.dg/data_char_1.f90
> if ((b(1) .ne. 'abcdj') .or. (b(2) .ne. 'hi k')) call abort
> !
> ! if ((b(1) .ne. 'adcdl') .or. (b(2) .ne. 'hi l')) call abort
> !
> end program
I agree that this is the correct fix and appplied it.
>
> (2) complex_write.f90 and fmt_read.f90 both fail with Cygwin but run
> fine with Tru64/RH9
Looking forward to any progress on this.
>
> (3) a good number of the testsuite programmes have no { dg-xxxx }
> directives.
I've fixed those. It is interesting to observe that entry_1.f90 contained one
typo beyond the missing dg-do statement, therefore this has never been tested,
not even outside the tree.
Are there any I'm missing this way?
[tobi@marktplatz gfortran.dg]$ grep -L "dg-" *.f90 *.f
data_char_1.f90
direct_io_1.f90
entry_1.f90
getenv_1.f90
pr16597.f90
pr17143.f90
pr17164.f90
[tobi@marktplatz gfortran.dg]$
Diff below, committed after verifying that all these tests pass on i686-pc-linux.
- Tobi
Index: ChangeLog
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/ChangeLog,v
retrieving revision 1.4915
diff -u -p -r1.4915 ChangeLog
--- ChangeLog 20 Jan 2005 19:20:26 -0000 1.4915
+++ ChangeLog 20 Jan 2005 22:47:22 -0000
@@ -1,3 +1,12 @@
+2005-01-20 Tobias Schl"uter <tobias.schlueter@physik.uni-muenchen.de>
+
+ * gfortran.dg/data_char_1.f90: Fix typo, add dg-do directive.
+ * gfortran.dg/direct_io_1.f90: Add dg-do directive.
+ * gfortran.dg/entry_1.f90: Fix typo, add dg-do directive.
+ * gfortran.dg/getenv_1.f90: Add dg-do directive.
+ * gfortran.dg/pr16597.f90, gfortran.dg/pr17143.f90,
+ gfortran.dg/pr17164.f90: Add missing dg-do directives.
+
2005-01-20 Kazu Hirata <kazu@cs.umass.edu>
PR tree-optimization/15349
Index: gfortran.dg/data_char_1.f90
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/data_char_1.f90,v
retrieving revision 1.1
diff -u -p -r1.1 data_char_1.f90
--- gfortran.dg/data_char_1.f90 19 Aug 2004 16:45:21 -0000 1.1
+++ gfortran.dg/data_char_1.f90 20 Jan 2005 22:47:25 -0000
@@ -1,3 +1,4 @@
+! { dg-do run }
! Test character variables in data statements
! Also substrings of cahracter variables.
! PR14976 PR16228
@@ -8,5 +9,5 @@ program data_char_1
data b(:)(1:4), b(1)(5:5), b(2)(5:5) /'abcdefg', 'hi', 'j', 'k'/
if ((a(1) .ne. 'Hello') .or. (a(2) .ne. 'orld ')) call abort
- if ((b(1) .ne. 'adcdl') .or. (b(2) .ne. 'hi l')) call abort
+ if ((b(1) .ne. 'abcdj') .or. (b(2) .ne. 'hi k')) call abort
end program
Index: gfortran.dg/direct_io_1.f90
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/direct_io_1.f90,v
retrieving revision 1.1
diff -u -p -r1.1 direct_io_1.f90
--- gfortran.dg/direct_io_1.f90 21 Aug 2004 11:20:27 -0000 1.1
+++ gfortran.dg/direct_io_1.f90 20 Jan 2005 22:47:25 -0000
@@ -1,3 +1,4 @@
+! { dg-do run }
! PR 16908
! Segfaulted on second set of writes. We weren't handling partial records
! properly when calculating the file position.
Index: gfortran.dg/entry_1.f90
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/entry_1.f90,v
retrieving revision 1.1
diff -u -p -r1.1 entry_1.f90
--- gfortran.dg/entry_1.f90 17 Aug 2004 15:34:12 -0000 1.1
+++ gfortran.dg/entry_1.f90 20 Jan 2005 22:47:25 -0000
@@ -1,3 +1,4 @@
+! { dg-do run }
! Test alternate entry points in a module procedure
! Also check that references to sibling entry points are resolved correctly.
module m
@@ -11,7 +12,7 @@ end subroutine
subroutine test1
implicit none
- call indidecta (foo)
+ call indirecta (foo)
call indirectb (bar)
end subroutine
Index: gfortran.dg/getenv_1.f90
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/getenv_1.f90,v
retrieving revision 1.1
diff -u -p -r1.1 getenv_1.f90
--- gfortran.dg/getenv_1.f90 6 Aug 2004 21:47:01 -0000 1.1
+++ gfortran.dg/getenv_1.f90 20 Jan 2005 22:47:25 -0000
@@ -1,4 +1,4 @@
-! { dg do-run }
+! { dg-do run }
! Test the getenv and get_environment_variable intrinsics.
! Ignore the return value because it's not supported/meaningful on all targets
program getenv_1
Index: gfortran.dg/pr16597.f90
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/pr16597.f90,v
retrieving revision 1.1
diff -u -p -r1.1 pr16597.f90
--- gfortran.dg/pr16597.f90 27 Aug 2004 07:59:29 -0000 1.1
+++ gfortran.dg/pr16597.f90 20 Jan 2005 22:47:25 -0000
@@ -1,3 +1,4 @@
+! { dg-do run }
! pr 16597
! libgfortran
! reading a direct access record after it was written did
Index: gfortran.dg/pr17143.f90
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/pr17143.f90,v
retrieving revision 1.1
diff -u -p -r1.1 pr17143.f90
--- gfortran.dg/pr17143.f90 25 Aug 2004 00:31:30 -0000 1.1
+++ gfortran.dg/pr17143.f90 20 Jan 2005 22:47:25 -0000
@@ -1,3 +1,4 @@
+! { dg-do run }
! pr17143
! does not print 2*63 correctly
character*25 l
Index: gfortran.dg/pr17164.f90
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/pr17164.f90,v
retrieving revision 1.1
diff -u -p -r1.1 pr17164.f90
--- gfortran.dg/pr17164.f90 25 Aug 2004 00:14:06 -0000 1.1
+++ gfortran.dg/pr17164.f90 20 Jan 2005 22:47:25 -0000
@@ -1,3 +1,4 @@
+! { dg-do run }
! pr17164
! index aborts when substring is longer than string
implicit none
More information about the Fortran
mailing list