Re: Sparc 6.0.3: multiple Nautilus file manager process after few minutes from fresh install
Jurij Smakov wrote:
On Sat, Dec 17, 2011 at 04:53:51PM +0000, Mark Morgan Lloyd wrote:
Obviously, assuming that those are the only 4 RC bugs affecting
sparc would be very naive. But if people do not bother reporting
them, we'll never take any action, and release will proceed, again
leading to complaints along the lines of "I can't believe this
buggy stuff was deemed a stable release". And if the port does not
have enough users reporting bugs even against the most common
desktop environment to the point that it turns out to be
hopelessly broken, maybe it's time to consider retiring it.
My apologies if I sounded unduly negative, but as you say, it's frustrating.
I wonder if I could make a couple of suggestions:
i) If somebody experienced with the platform and Debian procedures
sees discussion of a reproducible problem, with or without a fix, he
should suggest that the originator raise a bug and should specify
the URL etc. to do so. Where the OP isn't able to do that but
multiple people agree there's a problem, he should consider raising
it on their behalf.
This is a perfectly reasonable suggestion, in fact that's how it is
intended for the system to work. See some recommendations on bug
reporting I've just posted in another message in this thread.
I'm tied up for a few days (trial compilations on Solaris which might
run into routine server maintenance over the "holiday"), but as soon as
I'm able I'll have another session with Squeeze+KDE. If it's as flaky on
a U60 as it was on a U80 I'll kvetch here first to make sure it's not a
known or an obvious problem, and possibly to ask for suggestions- for
example it's not obvious how to get a useful backtrace out of konsole
since it's stripped.
Once I'm confident that I'm looking at a real issue and have wrung as
much info as possible out of it, I'm happy enough to raise a bug report.
However I think that for this sort of thing community involvement is
valuable before bugging the developers, simply to make sure that silly
mistakes don't cause unnecessary noise in the system. Or at least that's
how I work elsewhere, and don't in general get any complaints.
I'm not sure to what extent I'm prepared to invest time in "testing" or
[gasp] SID unless other people here make a similar commitment, since I
do have other things on my plate- both open source projects and the not
insignificant requirement to make some sort of living.
ii) My involvement with a different project has identified a number
of issues caused by operand alignment errors which manifest
themselves on both SPARC, (at least some) ARM, and possibly in some
cases on x64; one of the reasons I run multiple platforms here is to
allow me to distinguish between endianness and alignment errors. It
could turn out to be useful if developers for the platforms had some
way of coordinating effort.
Traditionally, responsibility for figuring out the arch-specific
problems lies with the package maintainer, who will contact porters
as necessary. On top of that, I'm regularly monitoring the
release-critical bug list and will try to at least debug any
sparc-specific issues which end up there.
But as ARM gradually pushes into SMP, the developers might find they can
use the SPARC community's experience.
Mark Morgan Lloyd
markMLl .AT. telemetry.co .DOT. uk
[Opinions above are the author's, not those of his employers or colleagues]