Re: [Announce] busybox-power for Harmattan
 

hi

check this minor fix to support obs :

https://garage.maemo.org/plugins/ggi...o-confirm-text

please test the shared repo version ...

Re: [Announce] busybox-power for Harmattan
 

The commit looks good and the OBS build works as expected. Thanks again!

--
Some background information for the rest:
RzR has been helping me to get meego-confirm-text in his shared community repository for Harmattan. This way you don't have to download and install the debian archive yourself (if you got the repository set up on your device), which makes the installation of busybox-power a lot user friendly.
Other packages can now depend on meego-confirm-text as well :).

Since I don't have an OBS account yet, RzR has been so kind to upload it for me (and even fix the package so it can be build in the community OBS in the first place). I'll make sure to upload a bumped version and further maintain the package in the community repository as soon as I have an account.

Re: [Announce] busybox-power for Harmattan
 

hmmm,cant seem to install the MeeGo-confirm-text......broken package...

Re: [Announce] busybox-power for Harmattan
 

Maybe post the terminal output?

I have it up and running here without any issues...

Re: [Announce] busybox-power for Harmattan
 

The following packages have unmet dependencies.
busybox-power: PreDepends: meego-confirm-text but it is not installable
E: Broken packages
~ #

Re: [Announce] busybox-power for Harmattan
 

Yup from first post:

So install aegisctl and meego-confirm-text first :)

Re: [Announce] busybox-power for Harmattan
 

ahhh,i didnt know i must incept it,thanks :)

Re: [Announce] busybox-power for Harmattan
 

i understand the desire to caution users away from installing packages that could very easily brick their phone if something unforeseen goes wrong. thats what warnings are for.

that doesnt mean its ok to forcibly require user intervention during a deb install, ever. ESPECIALLY when you require an x server and dbus session...

it makes it impossible to install as part of an automated restore script, without seriously going out of your way {e.g.: temporarily replacing the callable binary with a script that just cats the file passed in and exits with 0}.

i personally just remove the preinst because i know im on an n9 and i know i want to install the package, but it would be nice if you wouldnt put installation guis in deb files in the first place.

Re: [Announce] busybox-power for Harmattan
 

Thank you for your feedback, appreciated. However, I'm not in favor of removing the warning, because:

• It's only shown during first time installation (upgrades or re-installations won't trigger it).
• More importantly: I think it is a very sane warning. Take kernel-power for Maemo for instance: most users will not doubt that it is rock solid. However, kernel-power can't guarantee 100% compatibility with the stock installation, just like I can't with busybox-power. I have all faith in busybox-power for Harmattan -the download statistics for its older brother are very reassuring as well-, but I think it's only fair to warn users the way it's currently done (disregarding the fact that there isn't really an alternative way to do it, unfortunately)

Nonetheless, you're making a very valid point: the current implementation does not allow unattended installations of busybox-power.
Therefore I would like to suggest a compromise: what if I would let the preinst script check for the existence of a file (e.g. /tmp/.bbp-unattended), and if found, it'll exit 0 straight away. Since you're already automating installation, you would only need to put an extra "touch $file" in your script prior to installing the deb. Let me know if this works for you and I'll implement it and document it in the OP :).

Re: [Announce] busybox-power for Harmattan
 

cool, that sounds like a decent compromise. you should also check DISPLAY, and if theres no xsession, simply print the warning to stdout and continue {or maybe prompt on the command line} instead of failing with no error printed.

one more thing; can you leave out the signing and validating of the refhashlist in the noaegis version? i dont think it makes sense to do aegis-y things in it, if we're assuming that aegis is neutered.

i found that using the deb as-is gave me perm errors on accli. this is presumably because the kernel that im using {standard openmode pr1.3} doesnt automatically give you credentials like tcb-sign. i had to remove the calls to accli and validator-init while i was removing meego-confirm-text in the deb.

p.s. heres the deb i made, if anyone else wants an aegis-free and gui-free install. note that it requires openmode and may ruin your phone and give you cancer.
https://github.com/teleshoes/n9-conf...l.deb?raw=true