Feature #964: === GUI ===
GUI: find the way for update notifier to not to rely on CHANGELOG
|Status:||Rejected||Start date:||19 Jul 2016|
Our GUI's update notifier works as following:
- on every GUI start, if the user has "online update notification" allowed, the Notifier goes to internet and downloads our CHANGELOG
- then he read top line of the file
BornAgain-1.6.1, released 2016.07.15
- parses version number, and compares current GUI's version with version number in the CHANGELOG
- if names are different, a simple http-label containing whole line "BornAgain-1.6.1, released 2016.07.15" is shown on WelcomeView
The disadvantages is that every time 50k CHANGELOG gets downloaded. It might become a problem with the rise of a number of GUI's running around the world, and size of CHANGELOG. Also if somebody change CHANGELOG top line format, update notifier will stop working.Within this item
- study alternatives
- consider somewhat lighter VERSION file located in same directory
#2 Updated by wuttke almost 4 years ago
50k are very little by today's standards. So the current solution is not really problematic.
I do not like very much the idea of adding an extra VERSION file just to save a few nanoseconds.
As for other alternatives:
It would be sufficient to download just the first 80 bytes of CHANGELOG. However, partial download is not supported by wget. I guess this is due to fundamentals of the HTTP. So this is not an option.
We could download the directory index http://apps.jcns.fz-juelich.de/src/BornAgain/, and parse it for BornAgain-*.tar.gz. This would limit the download to <7k, not growing with time. Nice, but quite some coding effort.
We could restrict downloads to one per day, by storing the last look-up date in .config/BornAgain/BornAgain.conf.
#3 Updated by pospelov almost 4 years ago
May be you are right, and there is no problem at all, at least now. So we can continue with current approach, for the moment. As for restricting downloads, I wouldn't do this because this http calls appears in our httpd log and allows us to judge (to some extent) how many people are using BornAgain, and deduce (to some extent) daily usage statistics. Anonymously, of course.