In glibc 2.10 (32 bit) fallocate() exists but fallocate64() doesn't.
When _FILE_OFFSET_BITS==64, fallocate() is a redirect to fallocate64()
and the program can't be linked (fails to find symbol fallocate64).
See http://bugzilla.redhat.com/500487
Attached patch detects fallocate() more robustly to
On Sun, 24 Oct 2010 15:10:08 +0100
Timo Sirainen wrote:
> On 22.10.2010, at 19.22, Paul Howarth wrote:
>
> > In glibc 2.10 (32 bit) fallocate() exists but fallocate64() doesn't.
> > When _FILE_OFFSET_BITS==64, fallocate() is a redirect to
> > fallocate64() and the p
On Sun, 24 Oct 2010 20:26:51 +0100
Timo Sirainen wrote:
> On 24.10.2010, at 20.16, Paul Howarth wrote:
>
> >> A lot of code just to work around a bug that apparently only exists
> >> in Ubuntu 9.10 and Fedora 11. Is there a reason for anyone to be
> >> act
A tarball created from current pigeonhole hg using "make dist" doesn't
include the files doc/man/{reporting-bugs.inc,sed.sh} and fails to build
as a result. Attached patch works for me.
Paul
--- dovecot-2.0.1/dovecot-2-0-pigeonhole-d51650c8af85/doc/man/Makefile.am 2010-08-23 18:26:12.0
Hi,
The LDAP Plugin fails to build (it doesn't attempt to build it) in
1.2.rc7 because $have_ldap_plugin is checked before it's assigned in
configure.in. Patch attached. This was not a problem in 1.2.rc6 and
earlier versions.
Cheers, Paul.
--- dovecot-1.2.rc7/configure.in 2009-06-28 04:52:16
The update of ssl_proxy_get_security_string() in dovecot 1.2.6 has
broken compatibility with openssl < 0.9.8 (e.g. on RHEL4) because the
functions SSL_COMP_get_name and SSL_get_current_compression don't exist
until openssl 0.9.8 (build fails at link time due to these symbols
being unresolved). I've
On Wed, 7 Oct 2009 10:52:14 +0200
Thomas Leuxner wrote:
> Am 07.10.2009 um 10:44 schrieb Paul Howarth:
>
> > I've worked around this by reverting the function to
> > the 1.2.5 version for now but I think a better solution would be to
> > have a configure test
On Wed, 7 Oct 2009 10:52:14 +0200
Thomas Leuxner wrote:
> Am 07.10.2009 um 10:44 schrieb Paul Howarth:
>
> > I've worked around this by reverting the function to
> > the 1.2.5 version for now but I think a better solution would be to
> > have a configure test
