[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: pbbuttonsd script interface - future?

> > AFAIK the wakebay utility isn't needed anymore with recent kernel
> > versions (I'll double check to make sure). The 3400 hacks are still
> > needed, though.
> Do you know from which kernel version on the tool wakebay isn't needed?
> I then can add a test.

Tough. I found a patch I sent to Alan Cox on the CD wakeup issue dated
August 2000. I believe BenH applied the patch to his tree around that
time, Alan applied it later, and it was eventually solved for good in the
IDE rewrite. All 2.4 kernels should be OK, and anything later than 2.2.18
from BenH's tree should be OK as well. Maybe BenH can be more specific on
kernel versions - the code in question is revalidate_disk for IDE CDs
which had a default blocksize of 512 applied on wakeup erroneously.

I totally forgot about that - the wakebay call never hurt anyone but it
should have gone from pwrctl since ages. For all practical purpose you can
drop the special PMac_G3 script.

> > On a quick glance, I don't see how the PMac_3400 or PMac_G3 scripts get
> > called from the master scripts, BTW. And I fear that '. config' might not
> > work under all circumstances (absolute path preferred).
> You need a link from event.d to scripts.d/PMac_3400 to activate the script.

Which is set up by the user, or postinst?

> The '. config' will always work because the master script change working
> directory. I prefer relative paths because the scripts will also run from
> /usr/local without change. This makes it easier for 'Make install' to install
> a working version into the $prefix that the user has defined without
> complicated scripts to patch all the paths in the scripts.

Yikes. I always considered packages that come up with wrong path when
installed using $prefix as buggy (and I've seen a lot of them). Anyway,
I'll give your scripts a test.


Reply to: