Is udisks worse that HAL?

David Zeuthen zeuthen at gmail.com
Thu Jun 24 12:59:39 PDT 2010


Hey,

On Thu, Jun 24, 2010 at 1:17 PM, Maxim Levitsky <maximlevitsky at gmail.com> wrote:
> However I think that 1st priority for udisks should be removal of all
> hardcoded stuff.

Before you write any patches please do ask (in the udisks bugzilla)
about whether such configurability is wanted. I mean, you seem to be
pretty much assuming that hard-coding == bad and options == good.
Anyway, my view is that this generally isn't the case -  see
http://ometer.com/features.html for some discussion of why sometimes
"less is more". Havoc's write-up is old but still 100% relevant.

(In fact, letting users and 3rd party packages interfere with core
functionality often leads to tons of headaches and self-inflicted
wounds and usually leads to whole taxonomies of FAIL
(/etc/modprobe.conf comes to mind here). And, just FWIW, that's the
main reason why we hard-code some things today.)

Anyway, the few examples you mention (for example whether the desktop
automounter should try to mount the device) do seem like OK candidate
for something that can be controlled via udev.

    David


More information about the devkit-devel mailing list