I'd like to suggest to relax the "mandatory-ness" (yeah, I know that's not a word) of platform updates.
My suggestion is clearly motivated by the desire to be able to downgrade from a buggy update (like the recent 10.1.1), or simply from an update that happens to suite my needs less well than the prior version - but without the penalty of being cut off from application updates like I would be now.
I am a software engineer and I believe I could contribute to design such a solution. If the developers are interested in a technical discussion, just point me to the right place for it.
Either way, I hope you'll reconsider my suggestion.
Thanks.
[This thread was split off from the Platform Bug Reports thread - mod JTH]
This discussion was moved from the platform bug report thread. Relevant points:
- The platform currently requires updating the platform before updating the apps to ensure integrity and ensure it works with the updater database file (which will likely change in the future)
- zoc is interested in altering this behavior for folks who want a specific version of the platform for specific reasons (bug introduction, etc)
- John doesn't quite have the time implement it right now, but is interested at some point in the future, possibly including a 'minimum required version' of the platform in the updater database to lockout older versions as the app database changes, etc but allow newer (compatible but not current) versions to still get apps
- zoc knows Delphi (which John is happy about). The updater is currently written in NSIS but will be moving to Delphi at some point in the future. The Platform is currently written in Delphi XE2 and will likely move to XE3 when it is released later this year (after testing and shakedown)
Did I miss anything zoc?
Sometimes, the impossible can become possible, if you're awesome!
Huh, sorry for the long delay, John.
A "minimum required version" is pretty much one of the things I had in mind, so that sounds really good. Any news in that area?
B/c if that is a simple "hack" in NSIS, and it turns out to work well in practice, it might help avoid the effort (and perhaps complexity) of a more "complete" solution.
This is not to say I'm backing off my offer to help. I still am interested in helping improve the platform, if and when my time permits. After all, I use portable apps almost exclusively (so I know both the pleasures and the pain of it)