Given it's hard to get .hgignore rules to be absolutely correct in a repository where nightly has run and hg status -mard won't display files intentionally created but not added to the repository I'm thinking that this model of development would be nice:
+-------------+ |gate-hg-clone| +-----+-------+ | zfs snapshot +-----+----------------+ zfs clone +-------------+ |gate-hg-clone-snapshot|-.-------------|working_repos| +----------------------+ `-. +-------------+ `-._ | rsync `-. +---v---+ zfs clone `-.| build | +-------+ Where: gate-hg-clone is a local copy of the onng.sfbay gate-hg gate-hg-clone-snapshot is a zfs snapshot of get-hg-clone working_repos is a zfs clone of gate-hg-clone-snapshot. build is zfs clone of gate-hg-clone-snapshot. I will make all intentional file changes in working_repos that way hg status will be meaningful when ? status files are shown. I will rsync the usr dir in working_repos to the build directory and that is where I will do my builds. Much cleaner I think and very efficient. Note, actually I would have two build dirs (sparc, x86) which would be rsynced from my working_repos. -- Will Fiveash Sun Microsystems Office x64079/512-401-1079 Austin, TX, 78727 (TZ=CST6CDT), USA Internal Solaris Kerberos/GSS/SASL website: http://kerberos.sfbay Info about krb-diag: http://kerberos.sfbay/krb-tool-info.html http://opensolaris.org/os/project/kerberos/