Just my two cents.
Thanks Larry, cents appreciated. I must admit still I'm not convinced of
the evilness of the approach I hoped could be followed, regardless of what
you've kindly illustrated. What explained is right, if the new installed
OS works somehow differently from the old one; so, if I can interpret your
words, we users don't know how different the OS after upgrade is, so we
cannot be sure that everything will work as before after an "easy restore"
(let's call "easy restore" the alternative method proposed in my question,
and "normal restore" the plain install-all-again way).
But: what if the system is reflashed with the very same OS? Would an "easy
restore" work? I do not ask this because I have interest in reflashing a
working system with the same firmware, but because the answer can help me
better understand the issue.
If the answer is "no", than I'll understand there are problems beyond what
I'm considering with my poor knowledge of Linux, and I'll sadly force me
to do a "normal restore", if I'll decide to upgrade.
If the answer is "yes", than I'll understand that differences in the new
firmware from the old one *could* prevent the "easy restore" method from
working as hoped. In this case, the firmware developers could try
answering to the question: ''is the (current) new firmware different from
the previous one enough to prevent the "easy restore" method from working
as hoped?''; and they should provide this information for each eventual
new firmware release, so to let users know if the "easy restore" method is
feasible for that particular upgrade.
I would advocate a high-level approach to
the problem, tracking the installs (and repository additions) done by
the user with the install tool. After an upgrade the installer tool
could have an option to try and reinstall the applications they had
formerly installed. Successful installations could be "checked-off"
the list of applications to install, failures marked as such.
This is a good idea for me too, and at this regard, again, I'm still
waiting for a reply to my previous answer about the opportunity of
recommending to developers of apps for the 770 the use of the Zero Install
system, and a good application manager managing stuff like proposed
(merging mine and your proposals):
http://maemo.org/pipermail/maemo-developers/2006-November/006158.html
PS: Please note when I say "I'm still waiting for a reply" does not imply
"I want a reply and I want it now!"... I just want to make sure some
questions I care lay not hidden under tons of stuff, and let them emerge a
bit - like a personal "Komulainen-like-Unresolved-issues" post :)
--
Antonio
_______________________________________________
maemo-developers mailing list
[email protected]
https://maemo.org/mailman/listinfo/maemo-developers
