[Edbrowse-dev] Issues reported by Clang static analyzer

[Edbrowse-dev] Issues reported by Clang static analyzer

[Paul Onyschuk]

Clang is an alternative C/C++ compiler, compatible with GCC and
packaged for most Linux distributions.  It also provides static analyzer
[1]. Using it is simple as running:

$ scan-build make

Then it generates reports in html format.  Here is output for
edbrowse [2].  Some of this reports can be false/bogus, still worth
looking at I guess.

btw.  What is prefered method for contributing to edbrowse?  Pull
request on github or simple diffs sent to mailing list are also
acceptable?  I don't have anything specific at the time, just asking for
the future.

[1] http://clang-analyzer.llvm.org/
[2] http://blinkkin.piasta.pl/clang/edbrowse/

-- 
Paul Onyschuk

Re: [Edbrowse-dev] Issues reported by Clang static analyzer

[Chris Brannon]

Karl Dahlke <eklhad@comcast.net> writes:

> I think it was fine the way it was, really,
> would only cause trouble on a situation that I know can never occur,

Yes, I'm all for fixing these, even when we can prove that they're
spurious.  I'm happy to work on more of them.

Oh, and Paul asked which method was best for sending contributions to
edbrowse.  Patches to the list are the best, so that either Karl, Adam,
or myself can review them.

Thanks,
-- Chris

[Edbrowse-dev] Issues reported by Clang static analyzer

[Karl Dahlke]

I think these should be investigated, and fixed.
I fixed the first one, an uninitialized pointer.
I think it was fine the way it was, really,
would only cause trouble on a situation that I know can never occur,
but the compiler can't know that, and it's good practice to initialize things,
so why not.
If any of you want to review some of these errors, feel free,
otherwise I'll explore them as time permits.
Likely fix them one at a time,
one this week, one next week, etc,
so perhaps several small deltas.

Karl Dahlke