| Message ID | 54CA78F1.1020204@ubuntu.com |
|---|---|
| State | New |
| Headers | show |
On Thu, Jan 29, 2015 at 07:16:17PM +0100, Matthias Klose wrote: > On 01/29/2015 07:12 PM, H.J. Lu wrote: > > On Thu, Jan 29, 2015 at 10:05 AM, Matthias Klose <doko@ubuntu.com> wrote: > >> that fixes the build failure. ok to commit? > >> > >> 2015-01-29 Matthias Klose <doko@ubuntu.com> > >> > >> * acinclude.m4 (GLIBCXX_ENABLE_VTABLE_VERIFY): Define VTV_CYGMIN > >> unconditionally. > >> * configure: Regenerate. > >> > > > > This is wrong. You are checking vtv_cygmin before it is defined. > > my bad, somebody should obviously build both variants ... now testing this variant. Even with this patch vtv_cygmin is uninitialized. It will be likely empty string, unless somebody has vtv_cygmin in the environment, but it is better to initialize it, see the patch I've posted. > --- acinclude.m4 (revision 220257) > +++ acinclude.m4 (working copy) > @@ -2320,8 +2320,6 @@ > AC_MSG_CHECKING([for vtable verify support]) > AC_MSG_RESULT([$enable_vtable_verify]) > > - AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) > - > if test $enable_vtable_verify = yes; then > case ${target_os} in > cygwin*|mingw32*) > @@ -2341,6 +2339,8 @@ > VTV_CXXLINKFLAGS= > fi > > + AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) > + > AC_SUBST(VTV_CXXFLAGS) > AC_SUBST(VTV_PCH_CXXFLAGS) > AC_SUBST(VTV_CXXLINKFLAGS) > Jakub
On Thu, Jan 29, 2015 at 10:26 AM, Jakub Jelinek <jakub@redhat.com> wrote: > On Thu, Jan 29, 2015 at 07:16:17PM +0100, Matthias Klose wrote: >> On 01/29/2015 07:12 PM, H.J. Lu wrote: >> > On Thu, Jan 29, 2015 at 10:05 AM, Matthias Klose <doko@ubuntu.com> wrote: >> >> that fixes the build failure. ok to commit? >> >> >> >> 2015-01-29 Matthias Klose <doko@ubuntu.com> >> >> >> >> * acinclude.m4 (GLIBCXX_ENABLE_VTABLE_VERIFY): Define VTV_CYGMIN >> >> unconditionally. >> >> * configure: Regenerate. >> >> >> > >> > This is wrong. You are checking vtv_cygmin before it is defined. >> >> my bad, somebody should obviously build both variants ... now testing this variant. > > Even with this patch vtv_cygmin is uninitialized. It will be likely empty > string, unless somebody has vtv_cygmin in the environment, but it is better > to initialize it, see the patch I've posted. > >> --- acinclude.m4 (revision 220257) >> +++ acinclude.m4 (working copy) >> @@ -2320,8 +2320,6 @@ >> AC_MSG_CHECKING([for vtable verify support]) >> AC_MSG_RESULT([$enable_vtable_verify]) >> >> - AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) >> - >> if test $enable_vtable_verify = yes; then >> case ${target_os} in >> cygwin*|mingw32*) >> @@ -2341,6 +2339,8 @@ >> VTV_CXXLINKFLAGS= >> fi >> >> + AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) >> + > Or we can just do AM_CONDITIONAL(VTV_CYGMIN, test x$vtv_cygmin = xyes)
On Thu, Jan 29, 2015 at 10:28:17AM -0800, H.J. Lu wrote: > >> --- acinclude.m4 (revision 220257) > >> +++ acinclude.m4 (working copy) > >> @@ -2320,8 +2320,6 @@ > >> AC_MSG_CHECKING([for vtable verify support]) > >> AC_MSG_RESULT([$enable_vtable_verify]) > >> > >> - AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) > >> - > >> if test $enable_vtable_verify = yes; then > >> case ${target_os} in > >> cygwin*|mingw32*) > >> @@ -2341,6 +2339,8 @@ > >> VTV_CXXLINKFLAGS= > >> fi > >> > >> + AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) > >> + > > > > Or we can just do > > AM_CONDITIONAL(VTV_CYGMIN, test x$vtv_cygmin = xyes) Or that, but that is not what you've checked in. Some shells will be upset about test $vtv_cygmin = yes when $vtv_cygmin expands to empty string. Jakub
On Thu, Jan 29, 2015 at 10:30 AM, Jakub Jelinek <jakub@redhat.com> wrote: > On Thu, Jan 29, 2015 at 10:28:17AM -0800, H.J. Lu wrote: >> >> --- acinclude.m4 (revision 220257) >> >> +++ acinclude.m4 (working copy) >> >> @@ -2320,8 +2320,6 @@ >> >> AC_MSG_CHECKING([for vtable verify support]) >> >> AC_MSG_RESULT([$enable_vtable_verify]) >> >> >> >> - AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) >> >> - >> >> if test $enable_vtable_verify = yes; then >> >> case ${target_os} in >> >> cygwin*|mingw32*) >> >> @@ -2341,6 +2339,8 @@ >> >> VTV_CXXLINKFLAGS= >> >> fi >> >> >> >> + AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) >> >> + >> > >> >> Or we can just do >> >> AM_CONDITIONAL(VTV_CYGMIN, test x$vtv_cygmin = xyes) > > Or that, but that is not what you've checked in. > Some shells will be upset about test $vtv_cygmin = yes when > $vtv_cygmin expands to empty string. > Yes, I got checking for vtable verify support... no /export/gnu/import/git/sources/gcc/libstdc++-v3/configure: line 17481: test: =: unary operator expected At least it doesn't stop bootstrap. Either fix is fine with me.
Index: acinclude.m4 =================================================================== --- acinclude.m4 (revision 220257) +++ acinclude.m4 (working copy) @@ -2320,8 +2320,6 @@ AC_MSG_CHECKING([for vtable verify support]) AC_MSG_RESULT([$enable_vtable_verify]) - AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) - if test $enable_vtable_verify = yes; then case ${target_os} in cygwin*|mingw32*) @@ -2341,6 +2339,8 @@ VTV_CXXLINKFLAGS= fi + AM_CONDITIONAL(VTV_CYGMIN, test $vtv_cygmin = yes) + AC_SUBST(VTV_CXXFLAGS) AC_SUBST(VTV_PCH_CXXFLAGS) AC_SUBST(VTV_CXXLINKFLAGS)