Imagine there are Test versions ....

[Akoellh]

Originally Posted by Axeia

Can't blame people for wanting NVidia drivers on their beta version can you?

First of all, you can blame them for two things.

a) Not knowing, that they will not be able to report _any_ kernel related bugs after that without reproducing the problem _after_ uninstalling that driver, because it taints the kernel and therefor bug reports with this tainted kernel can not be accepted

b)Obviously not knowing how to build the driver "the hard way" no matter what distro, they would have the same problems on a stable release. I can not use the NVidia drivers on my test installations, because they are (or better "were") VMs. What I could do (and did) was to build the drivers from my "own" (modified from jengelh's specs) src.rpms, and _none_ of the drivers made any problems building. The src.rpm contains only one patch, which is needed to adjust the Makefile when building chrooted, so it is of no relevance for building via the run-installer. This leads to the conclusion, that all "NVidia driver not building with 11.2-*whatever-pre-version*"-problems are neither related to the drivers nor to the distro but to the users themselves not being able to install the driver "by hand" no matter which distro we are talking about.

What sense does it make to start testing unstable releases if you are not even able to perform such a basic and most importantly _very well_ documented step?

You will have lots of additional problems not related to bugs in the unstable release, you will have the same problems as all other users with real bugs (which would be great if you are the one actually finding and report them, but most people don't report because they don't know how to do it), you would have a lot more work if you would start do report bugs (and consequently, most "testers" avoid that) and you will gain practically nothing.

this is a good discussion to have....i've been thinking about it for
months....the 'catch 22' of it all:

1. several (me include) routinely say "when 11.x comes out you should
wait a few months for the bugs to shake out"

2. if everyone waits there will be NO bug shakeout...and eventually no
release..

and i have to ask myself, why would i say that? and, the answer is
because when 9.1, and .3 came out (i skipped 9.2) they didn't feel
like a beta (to me, as 10.1, 2, and 3 and 11.1 did)...

and, if 11.2 plops on the street as broken as .1 it will just get worse..

but, the only way to have a robust 11.2 is to have a BUNCH of bug
smashers busy in the runup to release....so, what do 'we' (community)
do, we have an "OPEN beta", open to anyone with a computer..

and, what do we get? [another 'catch 22'] a few experienced folks and
a mass of Redmond Ship Jumpers that think anytime a milestone/RC
doesn't do as Mista does it is broke--and they file a useless 'bug
report' sucking developer time, and patience [because even after chmod
777 it won't "do right"]..

ok, i've participated in a few betas in my day...they were closed (by
invitation and/or meeting some set criteria), had strict guidelines,
and active mail lists where tester and developer shared a usable level
of technical lingo..

and, if you didn't test/report as wanted, you didn't get invited to
participate again..

i was happy a few months back to see the call for testers here (and
wished i had time to participate)

btw, each beta i participated in had a reward at the end and somewhere
around here i have a 'free' retail copy of PKZip (for OS/2), and other
now useless stuff..

do we (the community) reward competent and productive testers with
anything? T-Shirt, mug, stickers, retail package? job title? what?

oh, if you did't catch it: i think it is wrong to think of these
*user* help fora as the way out of allowing downloads of M$ like
too-buggy-to-use releases--which just serve to (in my mind) further
tarnish the brand..

ymmv!
--
palladium
Have a lot of fun..

[growbag]

1. The forum is all about asking for help, otherwise what is the point of having one?
2. It makes sense to ask other users here first in case one is simply doing something wrong - avoids wasting dev time.
3. If nobody tests the new stuff, nothing would get done.
4. People use the beta versions for various reasons, personally I use them to get better support for newer hardware. And seeing as suse policy is to stick with one kernel version per release, it's really a by-product of that decision.
5. Official bug reporting is still far too hard and complicated.

Wouldn't it be nice to have a sexy little "report as bug" button that simply transferred you to the bug site with most of the data already filled in? No extra login names to remember, no blasted passwords etc'.

I mean, the whole point of creating this forum was to consolidate all the separate sites into one. Why not go one step further and integrate bugs as well?

"....because the devs would be bombarded with rubbish bug reports and waste time"

QED!

[Akoellh]

Originally Posted by growbag

1. The forum is all about asking for help, otherwise what is the point of having one?

Great idea, so please delete all HowTos, remove the search function, all stickies and while you are at it, also automatically all posts older than let's say 4 weeks, because the forum is then certainly not about finding answers which have already been given several times.

Originally Posted by growbag

2. It makes sense to ask other users here first in case one is simply doing something wrong - avoids wasting dev time.

And if being told to open a bug report in most cases nothing happens. That's what this thread is all about.

Originally Posted by growbag

3. If nobody tests the new stuff, nothing would get done.

Yup, but who actually tests this stuff an reports problems? Again, see last remark.

Originally Posted by growbag

4. People use the beta versions for various reasons, personally I use them to get better support for newer hardware.

OK, then why wasting any more time on stable releases and people actually doing some hard work on backporting drivers etc. to older versions?

Why at all offering newer kernels for distributions in unofficial repositories to give users the possibility to change drivers without having to change the whole OS?

Complete waste of time.

Originally Posted by growbag

And seeing as suse policy is to stick with one kernel version per release, it's really a by-product of that decision.

See above, not necessary in > 99% of all cases.

Originally Posted by growbag

5. Official bug reporting is still far too hard and complicated.

Now that's an interesting statement.

- Getting an account in bugzilla is the same procedure than getting one here in the forum

- When opening a new bug, one is asked with easy to select options on what product and part of the OS (kernel, YaST, networking, etc.) one wants to file a bug

- After that one gets automatically some search results matching the last bugs related to your choice before, so searching is done for you without you having to do anything yourself

- On entering your description, you get an extra field with "steps to reproduce" and a field for the actual behavior and the expected one

Apart from that, entering your description of the problem is exactly the same, so I don't see how reporting bugs is more difficult.

Originally Posted by growbag

Wouldn't it be nice to have a sexy little "report as bug" button that simply transferred you to the bug site with most of the data already filled in? No extra login names to remember, no blasted passwords etc'.

I mean, the whole point of creating this forum was to consolidate all the separate sites into one. Why not go one step further and integrate bugs as well?

Yeah, why not, still this will not raise the quality of bug reports _and_ of descriptions made in a forum, it is also not the topic of this thread.

Originally Posted by growbag

"....because the devs would be bombarded with rubbish bug reports and waste time"

They are not at the time, because even if somebody here is telling $USER to file a bug report, in most cases nothing happens.

Telling people to not use beta versions if they obviously don't have the skills (and very often also not the need) to use them is mostly ignored the same way as telling them to file a bug report, that's what this is all about.

[BrownieCat]

Right... Recently I've been worried about whether the final 11.2 release will have as few bugs as possible or will be extremely buggy... Because not many people seem to want to file bug reports. Some experienced users just work around the problem themselves, not caring to report the bug, saying it's a 'waste of time' and a slow and cumbersome process.

[caf4926]

Going on my current use of 11.2 RC2
It's virtually bug free.

[Akoellh]

Well, I could name you two annoyances you will find still present in RC2 ATM.

I know these, because I filed bugreports on them, one has already been fixed but isn't upstream yet, the other one got no response yet.

A) If you are on x86_64 or on i686 with kernel-desktop, try to disable ipv6 via YaST, reboot and have a look at the output of ifconfig -a after that.

=> No, ipv6 is still enabled

B) Switch from NetworkManager to IFUP, configure your user to allowing him access via kinternet/cinternet and try to configure any ethernet interface (wired or wireless) via one of those frontends.

=> You can not connect to smpppd not even as root with kinternet/cinternet(/or qinternet), there are no ethernet interfaces visible.

[Cloddy]

caf4926 wrote:

>
> Going on my current use of 11.2 RC2
> It's virtually bug free.
>
>

Does that mean it's free of virtual bugs?

Graham

[caf4926]

I can tell you I dropped the Desktop kernel and went default, it resolved an issue I had with powersave (bug reported BTW and I posted the resolved) Maybe worth a try for your issues?

[Cloddy]

Akoellh wrote:

>> 5. Official bug reporting is still far too hard and complicated.
>
> Now that's an interesting statement.
>
> - Getting an account in bugzilla is the same procedure than getting one
> here in the forum
>
> - When opening a new bug, one is asked with easy to select options on
> what product and part of the OS (kernel, YaST, networking, etc.) one
> wants to file a bug
>
> - After that one gets automatically some search results matching the
> last bugs related to your choice before, so searching is done for you
> without you having to do anything yourself

Which is a good thing because the search page that you're supposed to use
before submitting a bug is really off-putting and comes up with mostly
worthless results. I find it really annoying when I submit a search
specifying 11.2 and find most of the results are not for 11.2. To cap it
all, a little while after submitting a report which seems to be entirely
new, you get an e-mail to say it's related to another report and has been
added to it. From that, you find another load of reports that never appeared
in the initial search.

>
> - On entering your description, you get an extra field with "steps to
> reproduce" and a field for the actual behavior and the expected one
>
> Apart from that, entering your description of the problem is exactly
> the same, so I don't see how reporting bugs is more difficult.

The main problem I see with the bug-reporting system is the layout,
particularly to the search page I've already mentioned. A phrase springs to
mind that I heard someone use years ago to describe a dodgy software system
- "about as user-friendly as a cornered rat."

Also, why are some fields available for bug-owners to change but, when you
do make use of them, you're given a good ticking-off as they're only
supposed to be set by the software team. Any half-way decent system would
check your userid and set certain levels of activity available to you and
close off others.

Bugzilla has been around for a long time and looks it. Novell's version is
even more antique in appearance than some others. It's well past time that
it was redesigned. Perhaps I should raise a bug report?


Graham