[Webkit-unassigned] [Bug 196855] Fix Covscan uninitialized after ctor

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=196855

--- Comment #5 from Eike Rathke <erack at redhat.com> ---
(In reply to Darin Adler from comment #4)
> This patch includes fixes for some real bugs, and some cases where Coverity
> effectively just prefers a different coding style. The Coverity coding style
> may be safer, but we don’t necessarily want to make that kind of a change
> without a discussion first.

There may be some false positives now because the actual patches stem from covscans of between 2.20.x to 2.22.7 backported to master.

However, I think there is nothing coding style involved but rather most are indeed some "left uninitialized" errors. Or could you point out an example that you think is only coding style?

> Generally we would not land a patch, especially from a non-contributor,
> without discussion.

Fine with me.

> > Source/WebCore/html/HTMLMediaElement.cpp:2105
> > -        if (actionIfInvalid == Complain)
> > +        if (actionIfInvalid == Complain) {
> >              FrameLoader::reportLocalLoadFailed(frame.get(), url.stringCenterEllipsizedToLength());
> >              ERROR_LOG(LOGIDENTIFIER, url , " was rejected by SecurityOrigin");
> > +        }
> >          return false;
> 
> Good fix, but a different Coverity error, not uninitialized data members.

Historically it has been in a general covscan fixes patch that later was renamed to uninit_ctor probably because by then it contained mostly such fixes..  Certainly the covscan error here indeed was something different.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190412/1874a298/attachment.html>