Bug 435093

In case I was unclear, the current release of hal is preventing F8 hal-info from
working properly, not the upstream fix ;)

This is fixed in the current F9/rawhide hal package; an update is still required
for F8 to clear bug #280251 (via bug #434960) - or a clear statement that there
are no plans to update hal in F8 so that workarounds/kludges can be added to the
pilot-link package.

Do you want an update for F8? Theres a big difference between versions. Can't we
just build an errata package with this fix?

And it goes a little something like this... (hit it).

pilot-link needs a working PolicyKit and related hal configs; it's been broken
in F8 since release (well, since ~F5 in one way or another).

I pushed the relevant config upstream; the device info would come back down via
hal-info, the policy in hal.

So, we could solve this by either:
1) a hal-0.5.11 + hal-info >= 20080508 update (you seem unkeen)
2) an errata package solving this bug + hal-info >= 20080226 + policy rules
carried in pilot-link
3) carry all config in pilot-link, with added verbosity to work around this bug

Since we'd waited 3 months, last week I made a patch implementing 3) in
pilot-link - it's at the end of bug #280251, and there's a scratch build in koji.

I guess that leaves pilot-link vulnerable to duplication from hal-info updates
(e.g. the NetworkManager ones), and doesn't fix the audio player problems that
seem to have occurred (see link to fedora-list in comment #0).

To be honest, I'm at head-banging-against-the-wall stage. I don't care which
solution you plump for (and it's not my call). Communication between package
maintainers hasn't happened and it clearly needs to, given the interdependencies.

Bug #280251 has been fixed for ages now. It's a shame we seem incapable of
shipping the fix to users.

Do you have a link to a patch to hal-info or hal that I could easily apply?

Are you suggesting an errata to hal/hal-info just for in_outof, or to also carry
the fdi and policy needed for pilot-link? (i.e. make pending pilot-link patch
redundant? No problem with that if it's the best way)

I'm afraid I neglected to flag up bug #431377 (just tested: still present) which
means any update of fdi for pilot-link will take out hald when the PDA is
connected; so safest to reboot after install (or a naughty haldaemon restart).

Patch for in_outof fix (as per comment #0):
http://lists.freedesktop.org/archives/hal-commit/2007-November/003894.html

The following were against hal master at the time; I'd hope they'll apply to F8 hal?

Patch for Palm fdi:
http://gitweb.freedesktop.org/?p=hal-info.git;a=blobdiff;h=9bf7b6583bf38505bd660e768a7b3d157ba7d26d;hp=b81389604a97425e784df2f305f50824d8136a5a;hb=a196ae310a30b70f6b9de86b02cefc6ff4d06d9f;f=fdi/information/10freedesktop/10-usb-pda.fdi

and if in_outof is fixed, also apply (functionally identical, but more succinct):
http://gitweb.freedesktop.org/?p=hal-info.git;a=blobdiff;h=156c1517913c7dda8a967c0eea1c4f1b554c5f49;hp=9bf7b6583bf38505bd660e768a7b3d157ba7d26d;hb=5d8b2c479871d27f5a17023e20848b611c13e615;f=fdi/information/10freedesktop/10-usb-pda.fdi

Patches to policy:
http://gitweb.freedesktop.org/?p=hal.git;a=blobdiff;h=5620cae826dc60e31f2551df6ad2b19c0a7edebd;hp=50294e37f1307ac89c1d9badcf98cef5554b16a4;hb=af4487e991bbba50c7be29b9709da37e45fc20a8;f=fdi/policy/10osvendor/20-acl-management.fdi
http://gitweb.freedesktop.org/?p=hal.git;a=blobdiff;h=a7adf9f9ce2ad08cad81c4e93f27200285e7ad28;hp=12027b9c08efb0a998c4b825cae2ffc5013bc4d4;hb=af4487e991bbba50c7be29b9709da37e45fc20a8;f=policy/org.freedesktop.hal.device-access.policy

I think org.freedesktop.hal.device-access.policy may have been
hal-device-file.policy back then.

Apologies for any gitweb wonkiness on my part, I think this is what you need?

(In reply to comment #3)
> Do you want an update for F8? Theres a big difference between versions. Can't we
> just build an errata package with this fix?

I don't understand why going from hal-0.5.10 -> hal-0.5.11 in F-8 is such a big
deal, it would solve bug #280251 (which has been open for about 3 years now)
once and for all.  

It's a subminor version change, if it was a more major change surely it would be
a 0.6.x release, or does upstream HAL not honour those kinds of sensible rules
for version numbers?

Either way, please let us know whether you plan to push an errata fix for
pilot-link, and if so, what the time frame would be, so we can decide whether or
not to workaround it in pilot-link.

Works for me with hal-0.5.10-3.fc8 (and hal-info-20080607-2.fc8) in
updates-testing. Thanks.