[repo-coord] smart(dag) and atrpms-package-config(at)

[Jeff Pitman]

file /usr/lib/smart/distro.py from install of smart-0.29.2-1.1.fc3.rf
conflicts with file from package atrpms-package-config-91-1.rhfc3.at

I think what needs to be done with distro.py needs to be discussed.

[Dag Wieers]

Post by Jeff Pitman
file /usr/lib/smart/distro.py from install of smart-0.29.2-1.1.fc3.rf
conflicts with file from package atrpms-package-config-91-1.rhfc3.at
I think what needs to be done with distro.py needs to be discussed.

Agreed, I've put Gustavo in this discussion too.

Putting distro.py in the repo config package may not be a viable solution
as only 1 repo config package can have it and I'm not even sure smart
works properly without it anyway.

Axel, is there something in there that makes it specific for you that
can't go into the master distro.py ?

Would it be possible to have that master distro.py inside the smart
sources and request Gustavo for necessary changes. Gustavo will be
responsible for the consistency and feature-set for distro.py.

-- dag wieers, ***@wieers.com, http://dag.wieers.com/ --
[all I want is a warm bed and a kind word and unlimited power]

[Axel Thimm]

Post by Dag Wieers

Post by Jeff Pitman
file /usr/lib/smart/distro.py from install of smart-0.29.2-1.1.fc3.rf
conflicts with file from package atrpms-package-config-91-1.rhfc3.at
I think what needs to be done with distro.py needs to be discussed.

Agreed, I've put Gustavo in this discussion too.
Putting distro.py in the repo config package may not be a viable solution
as only 1 repo config package can have it and I'm not even sure smart
works properly without it anyway.
Axel, is there something in there that makes it specific for you that
can't go into the master distro.py ?
Would it be possible to have that master distro.py inside the smart
sources and request Gustavo for necessary changes. Gustavo will be
responsible for the consistency and feature-set for distro.py.

To be honest, I don't really know why smart wants to leave this choice
of freedom to the packager. I suspect this may have something to do
with other rpm distros which require a different distro.py. Or maybe a
dpkg system with a sibling rpm database.

Currently smart has three config parts, /etc/smart/channels, distro.py
and the state stored under /var. I'd prefer to have all configurable
parts in a /etc/smart hierarchy and also see enforcing of
/etc/smart/channels over the internal state (e.g. modifying
/etc/smart/channels should not require to remove the state under
/var).

[Dag Wieers]

Post by Axel Thimm

Post by Dag Wieers

Post by Jeff Pitman
file /usr/lib/smart/distro.py from install of smart-0.29.2-1.1.fc3.rf
conflicts with file from package atrpms-package-config-91-1.rhfc3.at
I think what needs to be done with distro.py needs to be discussed.

Agreed, I've put Gustavo in this discussion too.
Putting distro.py in the repo config package may not be a viable solution
as only 1 repo config package can have it and I'm not even sure smart
works properly without it anyway.
Axel, is there something in there that makes it specific for you that
can't go into the master distro.py ?
Would it be possible to have that master distro.py inside the smart
sources and request Gustavo for necessary changes. Gustavo will be
responsible for the consistency and feature-set for distro.py.

To be honest, I don't really know why smart wants to leave this choice
of freedom to the packager. I suspect this may have something to do
with other rpm distros which require a different distro.py. Or maybe a
dpkg system with a sibling rpm database.

Different distro's have different policies for duplicate packages etc...
But a centralized Red Hat one that comes with the smart tarball
(distro-redhat.py) makes sense to me.

Post by Axel Thimm
Currently smart has three config parts, /etc/smart/channels, distro.py
and the state stored under /var. I'd prefer to have all configurable
parts in a /etc/smart hierarchy and also see enforcing of
/etc/smart/channels over the internal state (e.g. modifying
/etc/smart/channels should not require to remove the state under
/var).

Axel, I simplified my distro.py to:

pkgconf.setFlag("multi-version", "kernel")
pkgconf.setFlag("multi-version", "kernel-smp")

Did you add/improve anything ? What does your version contain additionally ?

-- dag wieers, ***@wieers.com, http://dag.wieers.com/ --
[all I want is a warm bed and a kind word and unlimited power]

[Axel Thimm]

Post by Dag Wieers

Post by Axel Thimm
Currently smart has three config parts, /etc/smart/channels, distro.py
and the state stored under /var. I'd prefer to have all configurable
parts in a /etc/smart hierarchy and also see enforcing of
/etc/smart/channels over the internal state (e.g. modifying
/etc/smart/channels should not require to remove the state under
/var).

pkgconf.setFlag("multi-version", "kernel")
pkgconf.setFlag("multi-version", "kernel-smp")
Did you add/improve anything ? What does your version contain additionally ?

Cut and Paste:

if not sysconf.getReadOnly():
if not sysconf.has("channels"):
sysconf.set(("channels", "rpm-db"),
{"alias": "rpm-db",
"type": "rpm-sys",
"name": "RPM Database"})

for type in ["", "doc", "smp" ]:
if type:
kernel = "kernel-%s" % type
else:
kernel = "kernel"
pkgconf.setFlag("multi-version", kernel)

Makes sense to unroll the loop, as it's more readable :)

[Gustavo Niemeyer]

Post by Axel Thimm
kernel = "kernel-%s" % type
kernel = "kernel"
pkgconf.setFlag("multi-version", kernel)
Makes sense to unroll the loop, as it's more readable :)

Certainly. I belive the origin of this loop was a copy & paste
from the distro.py we use in Conectiva, where it makes sense
because we have many more kernel packages than these.