diff options
| author | Ken Raeburn <raeburn@mit.edu> | 2004-05-30 20:06:15 +0000 |
|---|---|---|
| committer | Ken Raeburn <raeburn@mit.edu> | 2004-05-30 20:06:15 +0000 |
| commit | 524da05167bbcaf31dde0ff60488191ade422f04 (patch) | |
| tree | c99b53603355c6d9b5f3c29db55dab2ef33cddcd /src | |
| parent | 78026363dabef1a5f216a165e7b0049d3ada125c (diff) | |
| download | krb5-524da05167bbcaf31dde0ff60488191ade422f04.tar.gz krb5-524da05167bbcaf31dde0ff60488191ade422f04.tar.xz krb5-524da05167bbcaf31dde0ff60488191ade422f04.zip | |
* aclocal.m4 (KRB5_AC_GCC_ATTRS): On AIX 4, mark the destructor attribute as
always not working, until we can construct a good test for the order of
destructors in the multiple shared library case.
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@16378 dc483132-0cff-0310-8789-dd5450dbe970
Diffstat (limited to 'src')
| -rw-r--r-- | src/ChangeLog | 7 | ||||
| -rw-r--r-- | src/aclocal.m4 | 13 |
2 files changed, 20 insertions, 0 deletions
diff --git a/src/ChangeLog b/src/ChangeLog index 6cf8877060..632676843d 100644 --- a/src/ChangeLog +++ b/src/ChangeLog @@ -1,3 +1,10 @@ +2004-05-30 Ken Raeburn <raeburn@mit.edu> + + * aclocal.m4 (KRB5_AC_GCC_ATTRS): On AIX 4, mark the destructor + attribute as always not working, until we can construct a good + test for the order of destructors in the multiple shared library + case. + 2004-05-04 Ken Raeburn <raeburn@mit.edu> * configure.in: Generate a makefile in util/support. Include diff --git a/src/aclocal.m4 b/src/aclocal.m4 index 61cb2ef9f9..51664026fa 100644 --- a/src/aclocal.m4 +++ b/src/aclocal.m4 @@ -1602,6 +1602,19 @@ void foo2() { unlink("conftest.2"); } int main () { return 0; }], [test -r conftest.1 || a=yes test -r conftest.2 || b=yes], , AC_MSG_ERROR(Cannot test for constructor/destructor support when cross compiling)) +case $krb5_cv_host in +*-*-aix4.*) + # Under AIX 4.3.3, at least, shared library destructor functions + # appear to get executed in reverse link order (right to left), + # so that a library's destructor function may run after that of + # libraries it depends on, and may still have to access in the + # destructor. + # + # That counts as "not working", for me, but it's a much more + # complicated test case to set up. + b=no + ;; +esac krb5_cv_attr_constructor_destructor="$a,$b" ]) # Okay, krb5_cv_... should be set now. |