Den mån 26 aug. 2024 kl 00:14 skrev Tom Smyth <tom.smyth@wirelessconnect.eu>:
> Folks,
> Im just wondering what other porters experience of scan-build for the projects that you are maintaining ?
> has it been useful in identifying bugs?... or is the analysis engine too basic or shallow to properly analyse code ?
When I run it on "openbsd" code, the things it finds are often super
deep, requiring 37 steps and that syscalls or libc calls return 0 when
they can't and then figuring out 36 steps later that if they did, then
a bug could appear. Like this I ran over got a long time ago.
http://c66.it.su.se:8080/obsd/2019-10-25/scan-build-2019-10-25-192004-30128-1/report-f36c08.html#EndPath
The creation of dirs is probably a "side-effect" that it can't track
and hence it thinks you could call "make_parent_dirs_something()",
return without err being set, and still fail later because it doesn't
know how it would work.
This is probably because obsd devs are mostly very experienced in
knowing the environment they run in and not producing crap code
generally, but I think scan-build would find a lot of super easy bugs
if we ran it over generic 1995 C code of the time.
It does find trivial things like
http://c66.it.su.se:8080/obsd/2019-10-25/scan-build-2019-10-25-192004-30128-1/report-36d1ed.html#EndPath
> are there particular types of bugs it is good at identifying without false positives ?
> are there particular types of bugs that it identifies that are probably a waste of time ?
> any advice on using scan-build-16 with ports would be appreciated...
I like browsing over the html pages and following the things it does
find. I'm sure over larger sources, it could pick up a lot of things
that need fixing.
The trivial thing above might just be writing to a counter for which
you later plan to actually use, but for now scan-build will think it
is a serious error you should handle.
--
May the most significant bit of your life be positive.
No comments:
Post a Comment