Il giorno 26/apr/06, alle ore 12:35, Eduard Bloch ha scritto:
How would that interact with running squid installations? Please
tell us
what the actuall constraints are instead of repeating "needs to stay
here, needs to stay there". I assume you want to:
- make a package name more clear to show what the purpose of that
package is. Eg. "squid-experimental"
- clearly state that this is an experimental package, that it does
not
share configuration or cache data with default squid installation
- upload the package either to unstable or to experimental, depending
on your opinion of how much damage the package may create.
Actually,
I would upload it to experimental only.
- keep config directories and cache data separate
- when squid-3.x is released, upgrade squid package with the new
version PLUS the changes from squid-experimental PLUS transition
code
if needed. THEN upload the package as "squid" to experimental, THEN
ask people to install it from experimental to test the upgrade
path.
And then move it to unstable for more wide testing.
Yes, something like this. I fail to see why packaging squid3 now
seems to be a 'damage' for debian users, but still this is my plan.
Since Rob Collins asked I'm actually planning for a package that can
be installed alongside with the actual squid 2.5 package.
Unstable is where such testing can happen, think zsh-beta or
different versions of postgres.
Ehm... zsh-beta is a different class of application, not something
that
can fsck up your system RSN by offering remote accessible
vulnerability.
Just like any network-service package, by the way. Bugs happen. And
fixes too. Why should squid 3.0 be worst than anything else
(sendmail, php, etc.)?
In addition, unstable is not 'supported' by a security service and
security fixes are just maintainer's duty, and I think I can cope
with it just like I did with all the security related bugs of squid
in the last three years.
Regards,
--
Luigi Gangitano -- <[EMAIL PROTECTED]> -- <[EMAIL PROTECTED]>
GPG: 1024D/924C0C26: 12F8 9C03 89D3 DB4A 9972 C24A F19B A618 924C 0C26
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
