That's a tough question. Note that Photoshop CS3 installer has been busted for months
Yep, the same problem busts the CS4 installer as well, so both CS3 and CS4 has gone from working(with tricks, practically flawless in CS4s case) to non-installable.
From what I have understood, this is not really a regression, but rather a redesign which have caused the application to fail to install.
I would not weigh Perrys problem against mine(not that I think CSx are low-profile apps), but rather say that regressions seems to be more ok when the cause is a redesign, and not when it is a common bug.
I would not agree with that, though. This project is, for good reason, very cautious about accepting patches of bugs and should also be so with new functionality. Now I am not aware of the reasons for the redesign, they might be valid, but I would still think that we could learn from this in some way. If not a policy, but maybe some kind of way to tell the community of things like this happening?
Yes, the development versions are development versions in the projects view, however, the community treats them as real releases. Come to think of it, the project does it as well. Bug reports regarding the release version are somewhat frowned upon. Isn't that a sign that major releases are a bit too far apart?
//Nicklas
PS. This is especially annoying since I've just had a positive conversation with Adobe about helping out with providing a download location for atmlib.dll so that Dan and Austin could include it into winetricks, removing the last really unsafe step of the installation. I just hope they don't watch the appdb. But then I am a developer, so I hardly notice frustration anymore. :-) DS.