I am just curious, what was the reason again for not using rsync for
port distribution?
Just an idea, why can't we make people who submit ports have an rsync
account:
---------------- rsyncd.cruxusers.conf -----------------
victord:pass1
cptn:pass2
---------------- rsyncd.cruxusers.conf -----------------
---------------- rsyncd.conf -----------------
#
# /etc/rsyncd.conf
#
uid = nobody
gid = nobody
# use chroot = no
max connections = 4
pid file = /var/run/rsyncd.pid
log file = /var/log/rsyncd.log
[victord]
path = /data/PORTS/victord
use chroot = true
secrets file = /etc/rsyncd.cruxusers.secrets
auth users = victord
read only = false
[cptn]
path = /data/PORTS/cptn
use chroot = true
secrets file = /etc/rsyncd.cruxusers.secrets
auth users = cptn
read only = false
# End of file
---------------- rsyncd.conf -----------------
(I checked, they don't seem to support %u to replace username)
Then people who want their repos to be public, can ask for the account,
and a simple script where user:pass is specified can generate these
config files. Then it's up to them to update their trees. A gui (I have
a php one that can handle this) can then list/search/etc the ports.
Wouldn't this be an easy solutoion? It centralizes all the HTTPUP repos
with minimal work.
Am I missing something? We can even make httpup take -user -pass -rsync
server, args so they can distribute them without learning rsync
commands. And httpup can just sync their repo to central location with
some default settings set for calling rsync.
Victor
