[pca] Create a --singleuser flag?

[Eric Ham]

On Wed, Nov 26, 2008 at 12:29 AM, Martin Paul <martin at par.univie.ac.at> wrote:
> Yes, they would be installed. And they are good examples for the precision
> with how these patch properties are set. As far as I can tell, the
> singleuser property is just plain wrong for 128402-02 (installs only
> /etc/default/kbd) and 138247-01 (installs /usr/bin/cdrw). Both files are not
> held open by running processes, they aren't drivers, libraries or anything
> else which could be dangerous to replace in multi-user mode.
>
> The README for 127755-01 is unclear, but I guess it suffices to disable the
> fmd service before installation of the patch.

Ok, that helps, I was wondering if the singleuser property on some of
those was really valid.

> When I added support for the PATCH_PROPERTIES to pca, I looked at the
> patchinfo files from about 1650 patches to see how reasonable the properties
> are. The conclusion was that you can't really trust them; "noreboot" turned
> out to be the only one to be halfway useful. I found only a handful of
> patches which had singleuser but *not* noreboot, and for most of them the
> singleuser property was plain wrong, just like for 128402 and 138247. So I
> decided that it's neither worth to add a --singleuser option, nor to treat
> singleuser as an equivalent to noreboot.

I guess Sun still has a lot of work to do in creating valid patchinfo
files. Thanks for the explanation and keep up the good work as PCA is
a wonderful tool.