"ext Steve Greenland" <steveg at moregruel.net> writes: > According to Marius Vollmer <marius.vollmer at nokia.com>: >> "ext Marius Gedminas" <marius at pov.lt> writes: >> >> > My personal pet peeve with the app manager is all those confirmation >> > dialogs. >> >> Yeah, and there will be more in the future... There should be at >> least one before starting the operation, > > Please Ghod no. > > I confirmed the operation when I selected "install" or "remove". We could maybe get rid of this confirmation dialog for operations that are activated from the toolbar button and the menu, but not when they are activated from the list view... hmm. I put that on my list. > The obvious exception, of course, is attempting to remove some required > package, or removing something that's going to break dependencies. The AM wont let you do that, not even in red-pill mode. >> and we can't get rid of the legal "Notice" dialog for non-certified >> software. > > How about letting me mark repos as "okay by me"... You can hack the configuration of the AM that tells it which catalogues are considered 'certified'. There is no UI for this, obviously, but look around in /etc/osso-application-manager. This only works for signed repositories, tho. >> When using the AM to install system software updates, there will be >> lots more: "please take a backup", "all applications will be >> closed, continue?", "device will reboot, continue?", "don't touch >> me while I do scary things to the kernel", maybe even more. We >> should try to combine these messages. > > This needs some thought -- in particular, what defines a "system > software update"? A package with "Maemo-Flags: system-update" counts as a system update. System updates right now are based on meta-packages. There are more flags: "reboot" if you want a reboot after installing your package, "close-apps" if you want all apps to be closed before installing it, etc. The meta package for a system update will typically use something like "Maemo-Flags: system-update, reboot, suggest-backup". We might need to extend this a bit to cover flashing of kernel and initfs in a nice way.
