28 replies to this topic

[samukets]

Is there any multiboot that works well with PLI 3.0 and other images oe2.0?

[WanWizard]

No.

[samukets]

This is an intentional limitation by PLI team?

[littlesat]

No....a limitation by multiboot.... OpenPLi 3.0 != OE 2.0

Edited by littlesat, 21 November 2012 - 21:43.

[WanWizard]

No.

This is due to the fact that OpenPLi isn't a copy of the original DM image with a fancy skin and some plugins bolted on, but our own development, both for the OpenEmbedded environment as for the Enigma environment.

The multiboot solutions available all expect a (copy of a ) DM environment, based on DM's OE.

[pieterg]

real multiboot won't be a problem.
But the last time I checked, all socalled multiboot solutions were sharing the kernel and some initscripts between all boot choices. That won't work of course.

[MostlyHarmless]

gutemine is pretty sure that it works :-)

[WanWizard]

Well, as long as he provides the support (not only on multiboot but also on the images running on it), and nobody comes here to complain about it, by all means, go ahead...

Our experience is different, booting an OpenPLi image on a not-native kernel leads to instability of the image, and the weidest errors. We don't wish to spend any time investigating those.

[MostlyHarmless]

you are not understanding how gutemine works

The users try out if the images boot and report if they are stable usable, gutemine only provides the technical infrastructure and doesn't support any images.

And with Wally West or Dumbo you can nicely boot your image with your own kernel too.

[Happysat]

you are not understanding how gutemine works

Last time gutemine said smth like this:

"Well, but the PL* Team has multiply communicated that they don't need or want any Multibooting of their images, hence I decided to stop the 'official' support :-)
And the gap to the normal OE 2.0 images will become bigger and bigger (partly intentionally from their side with things like playing hide and seek with modules) where I don't have time or fun for ever in closing the gap.
And inofficial support means simply if it doesn't work (anymore) nobody can complain at my place ...

The unsupported feature currently is only existing because it was fun to proove them wrong that it doesn't work anymore - so you are just benefitting from my sense of humor at the moment.
The same approach applies to dFlash - if they produce images without offering a working backup tool - it is not my fault, but I don't want to be annoyed with comments like 'then gutemine should fix it'
This is time and effort that I have to spend and it is not for granted that I do so."

At least he has fun @ his work

Edited by Happysat, 23 November 2012 - 21:01.

[MostlyHarmless]

Yes, but I don't care if an unsupported feature simply works (as it currently does) - and I think there are some users who can live with that status too.

Hence the correct answer to the original question would have been ... find out yourself ...

And isn't everything in life about having fun?

Edited by MostlyHarmless, 23 November 2012 - 22:26.

[WanWizard]

I know perfectly well how he works. All I'm saying that we will not repond to issues reported on images running in a multiboot environment.

[samukets]

If we do not have multiboot, have some version of dflash, PLI functional in 3.0?

[MostlyHarmless]

gutemine is pretty sure that it works too :-)

[WanWizard]

I'm very happy for him. In case it doesn't, or you encounter weird errors or problems, I'm sure he's happy to sort them out for you.
We won't, we've seen too many of those to be willing to chase ghosts. Our position on supporting OpenPLi on multiboot stands.

[pieterg]

What I've seen from gutemine, he's not usually going for 'hacked' solutions. So I would be surprised that if he would make a multiboot solution, that the multiboot targets would share the same kernel (and perhaps other parts of the image as well).
But reading his comments, quoted by Happysat, it seems that he does have problems with different kernel configurations, indicating that the kernel might be shared after all. That would surprise me, gutemine did seem to me like a 'rather no multiboot than crappy multiboot' kind of guy. But I could be wrong.
(just assuming that the quote is really from gutemine, I might be wrong there as well, because it doesn't appear to be very respectful to the work we do here, again not something I would expect from someone like gutemine).

Anyway, that's not the point here.
As Wanwizard says, just use your multiboot installation if you want, but if you experience problems, please do mention you're using multiboot, so we do not have to spend a lot of time trying to find a problem which might turn out to be caused by incompatibility with a kernel (or initscript) from some other image.

[MostlyHarmless]

The quote is correct, but your reply (That won't work of course) is not correct.

[MostlyHarmless]

PS: The driver hide and seek is related to the stv driver ... which is the only real difference in the kernel causing problems when multibooting as the different jffs2 compression of the kernels is not relevant when booting from ext4.

Edited by MostlyHarmless, 24 November 2012 - 17:05.

[WanWizard]

The quote is correct, but your reply (That won't work of course) is not correct.

I don't think anyone said that it wouldn't work. Al we said is that "if" it doesn't work, we won't spend time researching it unless it can be reproduced without multiboot.

But I'm repeating myself...

[Blogger]

Reading am post I get impression that MostlyHarmless is Gutemine !!??