myDSL Extensions (deprecated) :: removing an extension



DSL itself and the DSL Team do not try to limit the ways that you can use the system.

If you wish to do a full hd install and use the system like a stripped down version of Debian KNOPPIX, go right ahead.

As others have said, it is easier to manage your installations via the APT Debian Package management system than it is to install *.tar.gz , *.dsl  or *.uci extensions.

APT will automatically take care of installs and uninstalls and it will handle dependencies as best it can, too.

So, in summary:

DSL hd install works fine.
Debian APT is the preferred way to install new software to a DSL hd install.
*.dsl, *.tar.gz, and *.uci installs are also possible but in the case of *.tar.gz and *.dsl, you are SWIMMING AT YOUR OWN RISK because you cannot easily uninstall or undo any unexpected complications that may result from the install.

It seems clear to me that the developers like the current functionality and have other things to work on than a DSL-specific unintaller/dependency handler.

Since DSL also can be used with APT, I agree with the developers.

Use the right tool for the job.

Quote
It seems clear to me that the developers like the current functionality and have other things to work on than a DSL-specific unintaller/dependency handler. Since DSL also can be used with APT, I agree with the developers.

Use the right tool for the job.

I understand that now, in a roundabout fashion, as ke4nt said that they "are working on it", and SaidinUnleashed said its not only backburner, its "not even on the radar".
Again, im not trying to convince anyone else of anything!
My curiosity arose from what i felt to be stifling and resistance towards _someone_else_ working on such a thing.  I see DSL as powerful in a way i hadnt before, for faster machines that can handle loading the majority of it into RAM.  But i also see the way that it is powerful for my usage, which is getting a lot of otherwise bloated/slow applications into a small space that work much more efficiently at a loss of minimal functionality.  Its again not my priority to try and decipher what the developers should work on, nor am i trying...
i just wanted to understand if, or if even, anyone would be against the idea of others pursuing those ends.  and if so, why??  Every response to this seems to be 'that won't work', or 'what a waste of time' or such lot.  why?  is it soo far out there that installing/uninstalling of stripped down apps that can turn an old hotpad into an incredibly functional system?? why is that seen as hopelessly useless?

Hi, undertow.

I think maybe all the messages just aren't connecting so I'll give it a shot.

The mydsl system was created specifically for LiveCD and Frugal users. Actually, Frugal was created for use on cf cards but has grown wildly into an option for the hard drive. It has always been stated that we STRONGLY DISCOURAGE people using the mydsl system on a typical hd install because it's not meant to be a complete package management system. Apt-get is the best method of managing a hd install since it is a true package management system. There was no reason for the DSL developers to re-create the wheel.

If you want to create your own package management system for hd installs, that's fine but don't expect the DSL developers to embrace it, even if it works. That is just not a direction that is of interest to us. I'm not sure why you really want to pursue this unless you just want to see if you can do it.

It seems to me it would be time much more well spent to learn how to remaster the base and add in the mydsl apps if you want to have an OS full of apps that will run on old machines. As long as you don't add alot of extra files to /home/dsl or /opt you can run a remastered frugal install on a p1 with 32mb ram that includes the gimp and any other large applications without using much ram. I have even created a gtk2 remaster using the gtk2.dsl that uses little more ram than the base DSL iso.

Please understand.....If you roll out this process in these forums, people start to think this is a DSL supported process and then get frustrated if people don't answer their questions. If you start this going and then go on vacation for two weeks and can't answer questions then people get mad at the DSL community and think we're just a bunch of unhelpful snobs.

If you want to start this project and put up a project page somewhere to manage it, I think that's fine as long as it's stated that the project is not supported by the DSL community as far as technical support. The reason I stated above is also the reason  why Saidinunleashed doesn't want unofficial .dsl files linked in the Wiki. The Wiki is supported by the DSL team. If people are getting .dsl files from the Wiki, they are assuming they can come here to get support questions for that app. That is not the way to do it.

Ke4nt does ONE HELL of a job managing extensions. It is a HUGE job. As DSL has changed he has had to rebuild the entire extension database at times to keep up with the changes. He examines every extension and, if he thinks there are issues, he might send it out to the testing team for input, then post to the mydsl testing area before putting it in the main directory. This is all done to protect the community and ensure a good extension. HD installers, more than anyone, should appreciate the thoroughness of his work. They are almost guarenteed a working app that won't break their install. See, if we make major changes to the base for the next release and someone installs a .dsl package from the DSL Wiki that breaks their precious HD install, they are going to be upset at us for making that package available.

We are not trying to squash anyone's creativity here. The DSL community trusts us to give them a good OS with minimal problems and we strive to do that. It means that sometimes you have to draw lines that not everyone is going to embrace.

I am very proud that we released 1.5 with major changes and hardly a peep in the forums about issues. To me, that tells me we are doing something right.

I hope this came out the right way. I think what you are wanting to do is interesting but don't be affended because the development team does not see the value of supporting it. Everyone has their ideas of how things should be done. That's why there are hundreds of distros out there to choose from.

Good luck

Chris

my own opinion....
I see no reason not to develop ideas that you feel would help.  As others have said, several user-developed features have already been incorporated into the base.  I guess it depends on how much the core team feels the idea would improve DSL, and as far as i can see you are not getting much positive feedback because the team are probably not interested in doing any work on this particular subject themselves.  There is already a full package system that works in DSL, so maybe the idea of replicating some features in myDSL isn't worth their time.  If, however, you were to put together a working uninstall system you'll probably get better feedback that you get from just presenting an idea.

"that won't work" - meaning that the existing extension install method where original files can be overwritten is incompatible with a simple uninstall routine.  With computers, anything is possible provided that you spend enough time and creativity to develop a solution.  But the as-is system will not work and cannot be easily fixed without causing other complications like ramdisk bloat.

I think that your biggest misunderstanding is that you keep confusing the myDSL extension system with a traditional linux package management system like RPM or DEB/APT.

myDSL is not a true package management system AND NEVER WAS INTENDED TO BE ONE. In fact, there are NO management controls that are designed into the extensions.  On the plus side, this means that extension building is a simple process that is much easier than building a "real" package like a *.deb or an *.rpm  On the minus side, no management = no ability to safely uninstall or verify dependency requirements.

The idea is that these would be EXTENSIONS of the original livecd and not traditional software packages.  For use with a livecd or frugal, this is no big deal, but for a hard drive installation this can be a problem.

So, like I already said before:

There already exists a true package management system for DSL.  It is called Debian APT, aka apt-get.

It is very easy to activate and use dpkg/apt-get and it supports the full installation/uninstallation/dependency handling that is required of a traditional hard disk installed linux system.

Because the system is pointed towards the "oldstable" repository, it is also an ideal solution for older computers.  There are many GTK1.2 applications that are lightweight and ideal for use with an older computer system.

The DSL guys even provided a "Syntaptic" choice in the DSL main menu that provides a graphical install/uninstall interface.  I suggest that you give it a try.

It is also possible to install/uninstall non-Debian packages using a program called "alien", so it is also possible to use programs that are available in RPM format or Slackware packages.

Debian package management is an excellent and sophisticated way to manage software for your hard drive installation.  It is the ideal solution.

It also took man-YEARS of design reviews and codewriting to produce this excellent system and in my opinion it would be a "waste of time" to duplicate this effort by trying to turn the myDSL extension system to suit a similar purpose.

Hopefully, this makes some sense to you because I am running out of ways to explain it any better.

Next Page...
original here.