[Pm-utils] pm-utils 1.3.0-rc3 available for testing

[Michael Biebl]

Hi Victor!

2010/1/15 Michael Biebl <mbiebl at gmail.com>:
> 2009/12/14 Victor Lowther <victor.lowther at gmail.com>:
>> On Thu, 2009-12-10 at 23:16 -0600, Victor Lowther wrote:
>>> On Sat, 2009-12-05 at 22:24 -0600, Victor Lowther wrote:
>>> > PM-utils 1.3.0-rc1 is not available for testing.  Notable changes from
>>> > the pm-utils 1.2 series:
>>>
>>> Fresh after -rc1, -rc2 is out.  This includes several bugfixes forward
>>> ported from the soon-to-be-released pm-utils-1.2.7, along with a new
>>> --quirk-test parameter that bypasses the sanity checking imposed by the
>>> quirk database handler -- now you too can see what happens when you post
>>> your video card in userspace while running a KMS driver!
>>
>> Now at 1.3.0-rc3!  What should have been a tiny bugfix in is_set ended
>> up breaking everything.
>
> Here is some feedback/suggestions
>
> 1.) suspend/hibernate broke for me. I got this in pm-suspend.log
> /usr/lib/pm-utils/sleep.d/98-video-quirk-db-handler suspend
> suspend:/usr/lib/pm-utils/sleep.d/98-video-quirk-db-handler: line 225:
> ((: 0x10de ==  : syntax error: operand expected (error token is "==
> ")
> success.
>
> The result was, that no quirks were applied. I'm having bash 4.1 installed.

Thanks to some tireless debugging of Martin Pitt, we were able to
track this down and push some important fixes to make it actually work
for systems using quirks

One issue we noticed is, that the pm-utils quirksdb uses ";" as
separator in its match targets.
We were wondering, if we shouldn't just use "|" directly.
import-fdi-quirkdb makes one suggest that is the intention, but
apparently what ends up in the *.quirkdb files is ";"
Victor, what's your take on this? See also
http://bugs.freedesktop.org/show_bug.cgi?id=26296


> 2.) Could we please rename 98-video-quirk-db-handler to 98video-quirk-db-handler
> for consistencies sake

I just went ahead and did that.

>
> 3.) It would be great, if 98-video-quirk-db-handler was a /bin/sh script.

With the advanced features used in there, that is probably not easily
possible, so let's skip that for now.

> 4.) I think we should do a one-time conversion of hal-infos' suspend
> quirks and ship that within pm-utils.

Aside from that, having an updated pm-quirks tarball based on the
latest hal-info git master would be nice.

Victor, do you have time to roll a new, up-to-date of pm-quirks (maybe
including that in pm-utils directly)
and also a 1.3.0-rc4 with the fixes that have been committed.

Thanks,
Michael



-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?