Brian Masinick
When I had “raw” Debian installed on my system a couple of years ago, I experimented with converting it from Stable to Testing but in the end I didn’t really feel it had been worth it. Most of the packages in Testing were either the same or only slightly updated from Stable, and there was a major penalty in terms of the size of the regular updates I had to do on it compared to Stable (more than double the size each time). It may be different now.
Yeah, depending on WHERE in the development cycle of the current stable, versus the current testing, it may be a slight movement or a fairly significant movement. For instance, right after a release there might be a bit of a surge in changes to testing; the same might be true just BEFORE a new release, when the development team tries to get a bunch of packages promoted from unstable to testing, so they can go through an “official” testing cycle prior to release.
I don’t know why I couldn’t “see” any V6 kernels when I was experimenting with this earlier; it MAY have been because I didn’t get the repositories initialized (and online) in the version where I was jumping from one to another. This is precisely why distributors do such a good job. They know how to bring software in and build it from source, a package at a time, so if the question is “can it be done?” the clear answer is YES.
The better question is “how involved is it to go from one version to another or from one repo to another?” The answer again is that it CAN be done, but it does involve many distinct components and it is non trivial to do.
In the past I’ve done it. In my current state of current experience – my practical experience is WAY out of date; my general knowledge is still there, but it’s not current.
Regarding the relative efficiency in making a change, note that antiX has increased in size for every release I know of throughout its history; they have not been huge increases, but from what I can see in comparing my admittedly optimized instance of antiX 23.2 to any instances of antiX 26, best case I can squeeze within 20 MB of earlier releases; most of the normal builds are between 20-30 MB larger.
The very first instance of antiX I can remember used a VERY memory conservative 58 MB, but it was also a 32-bit image. As soon as I moved to a 64-bit image, that tacked quite a bit onto it. If I go back to my Dell Inspiron 5558 laptop I may be able to see a slightly leaner image of antiX 26 than what I see with my newest hardware, but it’s still likely to be more than the 215-220 MB I can squeeze out of my antiX 23.2 Full Runit 64-bit image.
Take a look at my “What are you Here With” thread and you’ll get an idea of the changes in sizes of the images.
EDIT: I forgot to mention the impact of systemd; I had intended to include this in the post but apparently forgot to do so; that’s a significant reason that attempting to do this kind of thing is not easy – especially if Debian keeps changing packages involving the inclusion of systemd as a dependency.
--
Brian Masinick
Alternate "Search B":
This search page