Re: What's cooking in git.git (Jan 2025, #06; Wed, 22)

From: Jeff King <peff@peff.net>
Date: Jan 31, 2025, 23:39 UTC
Message-ID: <20250131233902.GA3545002@coredump.intra.peff.net>
In-Reply-To: <20250131233452.GB3544301@coredump.intra.peff.net>

On Fri, Jan 31, 2025 at 06:34:53PM -0500, Jeff King wrote:

Show 13 quoted lines

> > Thanks.  A greedy me wonders if things like this can be caught by
> > them a bit earlier before they hit 'next', though ;-)
> 
> I've always been a little afraid to touch 'seen' since it does not
> necessarily even pass tests, and I don't want to waste too much time
> hunting problems in other people's topics. ;)
> 
> But I might give it a try. It will require some workflow changes, as I
> only run CI on what I integrate for my daily build, which is based on
> next. I do not ever look at or push what is in 'seen' at all to my repo.
> 
> Hmm. I wonder if I could just build my daily driver off of 'jch', which
> is a little less scary than 'seen' (and IIRC is your daily driver?).

Of course the ideal here would be for individual topic authors to run Coverity themselves. It's "easy" these days in that you just have to enable the CI job with a repo variable. But you have to sign up for an account with Coverity, and of course you'll have to sift through the false positives that appear in master from other people's topics.

I don't think they have any awareness of branches, or silencing noise that's in your upstream (although I am far from an expert, so maybe somebody interested can dig into their docs).

-Peff