On Mon, Jul 25, 2011 at 01:36:23PM -0600, Eric Blake wrote: > On 07/25/2011 09:39 AM, Daniel P. Berrange wrote: > >>(struct testTLSCertReq): Alter time members. > >>(testTLSGenerateCert): Reflect the change. > >>(mymain): Reduce stack usage. > >>--- > >> > >> > >>- /* if zero, then the current time will be used */ > >>- time_t start; > >>- time_t expire; > >>+ /* zero for current time, or non-zero for hours from now */ > >>+ int start_offset; > >>+ /* zero for 24 hours from now, or non-zero for hours from now */ > >>+ int expire_offset; > >> }; > >> > > > > >This is actually a change in semantics introduced here. The > >start/expire values were treated as absolute values, eg in > >places we pass '1' for expiry time to indicate a time way > >in the past. This has now become 1 hour into the future. > >This is why the 3 expiry tests were broken. > > > >I pushed the following change to make sure the expiry tests > >have times in the past again > > > >+++ b/tests/virnettlscontexttest.c > >@@ -1112,7 +1112,7 @@ mymain(void) > > true, true, true, > > true, true, GNUTLS_KEY_KEY_CERT_SIGN, > > false, false, NULL, NULL, > >- 0, 1, > >+ 0, -1, > > Ah - while I fixed the future stamps, I forgot to fix the past > stamps. ACK to this change; however, I still find it odd that the > start time is now but the expire time is in the past. Instead, > shouldn't we be using something like start time of -2 and expire > time of -1, so the start time still comes before the expire time? Well we're not really testing the start time in this scenario, so it doesn't really matter either way. Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list