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

Re: Bug#455769: same problem on wheezy + Thinkpad X220T

On Thu, Mar 28, 2013 at 06:48:09PM +0100, Daniel Pocock wrote:
> If a package fails to do what it is supposed to do, isn't that a valid
> argument that it is RC?
> gnome-power-manager is meant to suspend a user's laptop when they shut
> the lid and put it in a carry bag.  That's what the manual says.  If it
> can't do that, is the package suitable for release?

This looks to me like you're implying that gnome-power-manager does not
work at all, or at least not on Thinkpads. Is that so?

I use GNOME on two Thinkpads (an X200s, and an X220i), and suspends work
better than on any previous version of any Linux system I have ever used.
Not perfectly: about once a month a suspend fails. Given the frequency
with which I suspend and resume, that's a less than 1% failure rate.
Suspending is very tricky, in the kernel, and I'm amazed it ever works.

I had my first suspend problems with a laptop in 1999. By amazing
co-incidence, that was also the first time I had a laptop. With every
laptop since, and every kernel version I've used since, and every version
of Debian I have used since, there has been glitches in suspend. Sometimes
they've been frequent: the first laptop I had with ACPI instead of APM,
a few years ago, was very unstable for suspend, to the extend that I would
rather power off instead of suspend. Suspend problems have happened, but
things have been improving over the years. It's unfortunate that the
problems have not been completely fixed, but are we doing better with
wheezy than squeeze? In my experience, yes. (Anyone who wants a better
answer than that should collect actual statistics.)

> No, that is not my reasoning here.  My feeling is that these are issues
> which can and probably should be fixed before a release, to take off the
> rough edges and/or ensure users have a relatively smooth experience. 
> Among other things, many people take pride in referring to the
> legendary[1] stability of stable Debian releases, and it is often
> emphasized that these releases only come when ready... some of these
> issues make me feel that something is not quite ready, but very close to
> it, and it would be a shame to compromise.

Some people's laptops sometimes fail to suspend. Does this warrant
making the release even later than it currently is? I don't think so.

Perfection is unattainable. Every Debian stable release is buggy as
hell, and that's unavoidable, if we want to make any releases at all.

If we can get wheezy out, we can start updating packages in Debian,
and maybe, just maybe, some of the bugs are already fixed upstream,
and we can make backports of the packages, or the fixes, and then
the first stable point release will get the fixes.

It's a good thing I'm not in the release team, or we'd have a buggy
release every week. Possibly after every package upload.

http://www.cafepress.com/trunktees -- geeky funny T-shirts
http://gtdfh.branchable.com/ -- GTD for hackers

Reply to: