Op vrijdag 26 augustus 2011 09:56:43 schreef Florian Hubold: > Am 26.08.2011 09:50, schrieb Sander Lepik: > > 26.08.2011 10:27, Guillaume Rousse kirjutas: > >> I don't see where it is stated than firefox can't get updated until all > >> extensions in the distributions work with it. > > > > I would call it a regression that shouldn't pass QA. One day your ads > > ands scripts are blocked and the other day they are not. Great user > > experience.. > > > >> The large advantage of packaged extensions is that they are managed by > >> sysadmins, instead of users, which makes a difference when they are > >> different people. > > > > And even larger disadvantage comes in when they are not updated. Contain > > memory leaks and so on. The biggest problem here is that the packager who > > imported those addons is not keeping them up-to-date. > > > > Sysadmin can copy those extensions into place anyway. Then it's up to > > him/her to keep them up-to-date. Currently we have to take care of that > > and we don't (or if anyone wants to claim that we do then we do it > > really poorly as there are already newer versions of addons that got > > pushed with latest update). > > Well, as there are mostly no bigger changes like in %files, can't an update > of those addons > be scripted or automated halfway? So someone just needs to push this prior > to a Firefox update? > > Not saying i have an opinion pro or against packaged addons, but it is > really a nuisance > as those who do the Firefox update also need to update all those addons, > which is unfair, > seems to me if it's not the importer/maintainer of those addons who does > the update.
thinking about what i said in previous email... i myself maintain a package which contains a firefox addon. maybe maintainers of dependant packages could be notified if the requirement is pushed to updates_testing? so they can test if this is working or not? if anyone maintains a list: beid-middleware is the name of the package, and i haven't yet tested if it does work or not.
