Static analyzers are very good things, and this is a good endorsement for PVS Studio.
As I read, though, I noticed a great number of the errors are also warnings in modern gcc & clang. The Tizen code under analysis must spew pages of compiler warnings every time it’s compiled. If the developers have chosen to ignore warnings, then they’ve already lost - if they want fewer defects then what they need first is a healthy development culture, and then perhaps a new static analysis tool.
“So I compiled this code base that didn’t have -Wall -Wextra -Werror -std=c++11 turned on, and it had errors!”
As someone who loved their N900, (and N800!) I’m pissed at what potential has been squandered by nothing being done with it after Nokia gave up. Samsung’s latest attempts are… not great.
What do you think about SailfishOS? It’s successor of MeeGo so in a way that’s continuation of N9/N900/N800 legacy.
They’re still living in 1491, so I haven’t seen any HW in the flesh, but from what I’ve been seeing, it isn’t terribly interesting. Jolla got lost in the woods with a tablet, a la Playbook.
I’m still looking for a useful replacement for the n810, which may have been the best portable computing device I’ve ever used; the n900 was just a little too bulky (but I still have mine).
Most of these errors would be hard errors in a language like Rust or Swift.
Or Perl, or PHP, or python too.
Not really sure of the value of people saying they need to rewrite a codebase from a time when Rust did not even exist into…Rust.
Maybe you can provide insight?
And some of them are already caught with better GCC/llvm flags. And some of that code is just bad C/C++, so it isn’t surprising that it also has errors.
Did I get it right, that the error count isn’t even that high? Just looking at them, I agree with the facepalming unicorn, and 1-5e/1kloc sounds horribly high, but apparently Tizen isn’t an entire trainwreck if it gets a mere 0.41.
Wonder what the number is on Panasonic’s fork of Firefox OS. Anyone know anything about Panasonic?