The good thing about @marxian work is that he can target both SSU and CSSU with the same code and maintain functionallity with very little tweaks in code. At the end, you always find a workaround againts bugs / missing features, but, this extra work is really insane.

Honestly I think you should really get involved. I dont believe it is necessary to go through 477 pages. If you just ask the team I'm sure that they would explain to you why it needed to be packaged this way.

The team is looking for someone to do documentation with the knowledge to understand the inner workings of maemo, something I do not posses, and the ability to clearly and concisely document it so that it could be understood.

I'm certainly willing to help flesh out the Wiki. A bare changelog is just not enough to sell users on the value of the CSSU; it looks like I'll be crawling through the details of the changelog and corresponding source code myself, so I'd be more than happy to provide a more concrete description of what has been fixed and why users should care. (I'm already trying to work my way through the changes to Modest, as those seem intriguing...)

Anyway, I'll help where I can, if I can. If only to sell this product to myself...

And you can always join #maemo-ssu and ask there

What is this "lock-down" you keep talking about? What do you mean by that? Do you mean that it's not something you can uninstall without a re-flash? That's true of many packages. Ever PR release has required a re-flash to uninstall it. Kernel updates require that you reflash the kernel area if you uninstall (which someone finally made a meta-package for as part of the KP install). There are things that CSSU touches (hidon-desktop, QT, and others) that would require a reflash to uninstall. That's why it's being done in a separate repository and NOT just a package thrown into extras!

Sorry, but PR1.3 was also a "lockdown", also requiring it's own extras repos. If you're so against that, why are you running PR1.3, and not staying on PR1.2? Wouldn't want to be "lockedown".

It took the cleanest path available to it within the structure that was already in place. If you "remove the meta-package" you break things like OVI store, and remove any ability for third party repositories to create or indicate which PR version they require. So you then would have to require all CSSU users to not use 3-rd party repos that rely on that, including OVI store and Nokia's base repositories.

Lots. H.E.N. requires KP to do most of the work in connecting to USB devices. TrueCrypt has a package version with and without kernel support, as having kernel-based support would be required for root-volume mounting (and speed improvements, less user/kernel mode switching). Several apps that monitor battery use require the kernel modules to gather that data for them. The repositories are chock-full of "apps" that require kernel packages.

And there are lots of apps that require the latest version of QT, as developers are using features new to that release for performance or coding ease. Nokia saw fit to tie it's QT version to the platform, so changing it requires a change that would break other things unless you do it as a bundle. CSSU has to choose to continue that, or cause CSSU to break lots of other things, which others here have cited as why they wouldn't install CSSU, because it "broke" something. Imagine if CSSU broke OVI store, or QT app installs. Who would install it?

I think the hostility comes from the fact that you're presenting yourself as being intelligent enough to ask for code samples, or direct "proof", but too lazy to bother to go find them yourself. Even after being told it's all public, and being given links to release notes, you still insist others do the work to provide you with specific answers and details.

For example, I as a user (not a developer) noted hildon-desktop in PR1.3 had a memory leak, as it's use would go up every day for me. It could have been hildon-desktop, some default widget, the way something was called in a library, a QT related leak... I don't know the specifics. I do know that when I installed CSSU, it stopped doing it. Your response was essentially "pics or it didn't happen", asking for "proof" of an exact diff of where that memory leak occurred, and what patch fixed it. Doesn't that seem a bit childish to you?

If I asked for "proof" that the World Trade Center was hit on 9/11, you'd probably tell me "Google it!" And if I said I wanted you to provide links, after you said to google it and presented a link to the CBS video archive page, would you provide more links, or scoff at me and treat me like an idiot for not using the tools you pointed me at earlier?

That's what's happening here. You've been told where the change logs are, and where the source is at. Yet you still continue to ask for exact details and "proof". Said simply: When you act like a child, people tend to treat like one.

That on the other hand is great. Being willing to actively do something to make the wiki more reflective of the work going on in the project would be much appreciated. Given the choice between spending an hour updating a wiki page to show details of things done vs doing more work, most coders would choose the later. As a result, wiki's tend to be maintained more by people that are not coders, who do just what you've now said you did: Look over change logs and check-in comments and update the wiki to match.

Doing that is not only a huge plus for the community, but it also causes developers to use the wiki more frequently, since all they need to do then in make minor corrections, vs entering all the data by hand. That frees them up to do more development, and makes them happy at the same time. ( I say this as both a developer and as someone who's done this for public projects. )

Well, of course I'm lazy! That should be obvious from what I've written already.

But this thread was started by the question "People who do not use the cssu, why?" My answer was, to be honest, the counter-question "why should I use the CSSU?" The CSSU folks appear to have done an awesome amount of work fixing bugs. But there is no good, clear, simple explanation of what it is they have done.

Yes, I do insist on specific details! The CSSU folks, for all their brilliance, are not Nokia. My N900 is my phone, my only phone, and I'm not willing to perform major surgery on it if the only description of the CSSU I can find is "it works for me" or "it fixes more than it breaks."

So yes, I am going to now do the grunt work and plow my way through a year's worth of IRC logs and 500-600 pages of TMO discussion to sieve out what I can about what has been changed, and browse the source code to boot. And since I don't intend to install CSSU until I've done that, this would be another answer to "why I haven't installed CSSU".

Here's the thing: if you don't know what was causing the leak, then you don't actually know that installing CSSU fixed the leak -- all that you actually know is that something changed around the time that you installed CSSU. It could have been CSSU (and indeed probably was), but it could also have been something unrelated to CSSU that changed around the same time, or it could even be that the leak is still there and a change in usage habits (perhaps due to features of the CSSU) have hidden the leak.

The truth is, you don't actually know. The CSSU gives you a warm fuzzy that something good occurred, but for all we know, installing or uninstalling something else would have given you that same warm fuzzy.

It's nice to pin all the warm fuzzys on the CSSU, but it's even nicer to prove that the CSSU improves your phone. The proof is there, somewhere, in the code. I'll see if I can find it...

Correct! I have been acting like a child here. I'm asking to be taught. I want instruction, schooling, anything. If you're saying that self-study is the only option available, that's fine. I'm not going to be happy about it, but I can deal.

Welcome to the club.

I take a slightly different take on this. See, just before I installed CSSU I used this really awesome little app called BackupMenu to do a complete dump of the rootfs and opt areas. It tars the whole thing up, and offers a way to wipe it clean and restore both partitions with a backup file if so desired. That way, for me to "uninstall" would be as simple as doing a restore, no flashing or reconfiguring involved. (I did this before PR1.3 as well, for the same reasons.) I only needed to restore once, and that was because of a goof on my part. But I know it works well from that experience.

So how's that for lazy?

Bravo! Much appreciated... especially if you plan on updating the wiki with that information as well. You could use it as a place to keep notes on what you've read?

I see, so you can fly then? Because I can't prove how gravity works, so clearly it doesn't affect you? Sometimes you can in fact know something without having directly witnessed it. Quantum physics wouldn't exist if what you suggest was always the case.

You see, I had all the updates applied several days before going to CSSU, and the leak was still there. I know for sure because it would eat enough that it would slow the system down from swapping after 5 or 6 days. So before the CSSU update, I rebooted to BackupMenu, did my backup, added the CSSU installer, installed CSSU, and ran with it for several days without adding/updating anything else. The only thing that changed in that time-frame on my N900 was the CSSU update. So I know for a fact that it was something from CSSU that fixed that bug (and a few others, like the widgets vanishing after hildon-desktop being killed, etc.).

Some things (like the dsp bug and the sgx issues) I can't pin specifically on CSSU, since they happened very infrequently, and I didn't have a way to trigger them explicitly. I suspect the sgx bug was fixed by KP49, and the dsp bug was fixed by a recompile of the upstream-patched drivers by freemangordon for 720p video. But I installed those a week or two after I updated the CSSU, and didn't have an sgx issue in that time frame. (I did have the dsp issue though, so I know that wasn't fixed by CSSU, though maybe by KP49?)

But again, even knowing that one item, having tested it in the "vacuum" of no other updates, I couldn't specifically point to a particular line of code from one package of the several hundred that updated as part of the CSSU install and say "that was it!" Nor should I have to.

Do you insist a doctor tell you how a drug works, at the biological level, and how it reacts to every other compound in your body before taking a drug? If you did, you'd never even take an aspirin, since modern science still doesn't quite understand how it does everything it does. Sometimes you have to take quantitative correlations into account and move on with things because it's not worth the time (or is impossible) to prove directly.

I'm saying some self-study is required. This is college, not kindergarten. People are happy to point you to the right resources, and get you started in the right direction, but you won't be spoon-fed everything. The tone before was "read me a story!", which got you a returned tone of "here's a book, kid, learn to read". If you ask questions that show you'd done some work, like "But why did Jack and Jill need water?", you'll find you tend to get a more responsive answer.

I just took the leap and installed the stable CSSU. Worked great. On the surface very little is visibly different, but I do like the portrait mode for the desktop. I'll probably enjoy the bug fixes/memory leak fixes more.