@nsigma, sorry after looking in my dowloads folder, it’s a .bin (self extracting) file. It’s been a while, and now that I think about it, I did have to link to the .so file in the Chrome plugin directory. But it wasn’t anything too tedious = probably why I was thinking now that it was a .deb then… My apologies.
ROFL. If you want to troll this thread you might want a better example. Or shall I bring up Microsoft’s track record in shipping “compatible” Java VM’s? ;D
@gbeebe - no need to apologize! I was interested to know if Oracle had a .deb link somewhere on their rabbit warren of a website. I’ve done manual installs enough in the past, but it’s not an ideal solution for the casual user who suddenly finds something that doesn’t work. And manually setting up /etc/alternatives is a bit more of a PITA.
@gouessej - what you’re suggesting is not that far off what’s happening from what I understand, just limited by the technical constraints imposed by the package manager.
These are 2 different things. Canonical pushes a ghost package and the final user still sees that Oracle Java is installed whereas it isn’t. My suggestion uninstalls Oracle Java cleanly if the user accepts the update and if he tries to check whether Oracle Java is still installed on his machine, he can clearly see that it isn’t installed, it allows him to know exactly what is wrong and to try installing it by another way instead of supposing Java itself has a problem. Canonical solution is a lie, proposing an update that explicitly removes Oracle/Sun packages instead of replacing them by empty ones would be far better.
@JL235 I agree with nsigma. Windows is not a champion of backward compatibility, how many APIs have been suddenly discontinued? the joystick API? Silverlight? There are even 4 different ways of retrieving the full path of the desktop in its API since Windows 3.1, none of them works on all Windows, you have to use them all if you want to support all Windows versions.
I understand the difference - please note I said “limited by the technical constraints imposed by the package manager.” AFAIK, what you’re suggesting is not possible using apt/dpkg. No idea whether the same is true for other packaging systems.
Lying to your users by satisfying a package dependency with a broken package is simply as poor as managing a distribution can get. If you can’t remove a package with the package manager, then perhaps the proper answer is don’t.
But I already know dist-upgrade can substitute one package with another, so if Canonical proceeds with the whole broken package trick, then it not only demonstrates they can’t be trusted, it also glaringly points to clear incompetence at managing their own distribution in the first place.
Yar boo sucks to Linux package management. Can’t stand the concept. Now at least I see there was some reason for my apparently unreasonable paranoia. Not that I give 2 poos about Oracle Java mind - it seems to not work very well compared to OpenJDK.
Cas 
Linux package management is just fine, thanks. It’s Canonical that’s screwing the pooch here.
And your experience has been Oracle JDK being of lower quality than OpenJDK? That’s unique. Should be moot now though, since Oracle JDK 1.7 is OpenJDK 1.7, so at least this clusterf*ck not only has a horizon, it’s one we’re halfway over as it is.
For no particularly well explained reason my games have never run on Oracle/Sun Java, but always worked fine with OpenJDK.
Cas
Ditto…
You think so ? I also thought that OpenJDK is better than Oracle JDK, and that is the general consensus.