Matthew Dillon wrote:
> Yes, but the problem is that 99.9% of the time the cases where one
> isn't checking the return value are, in fact, not bugs.
Agreed with all of your reasoning and anecdotes except this 99.9%
figure. It might approach that here (DF) because there are a lot of
eyeballs watching each commit and the calibre of developers is unusually
high.
If I had to haphazard a guess, I'd say it's closer to 80% in our shop.
Sadly, there are people who will commit code simply because it compiles
and time/budget constraints limit the number of eyeballs watching each
commit. (And, joy, I get to help unspaghettify their code...)
Dave