From: Jeff King Date: Wed, 12 Aug 2026 21:34:38 GMT Subject: Re: [PATCH v3 00/12] coverity: fix unchecked returns Message-ID: <20260812213438.GC152730@coredump.intra.peff.net> In-Reply-To: On Wed, Aug 12, 2026 at 10:29:33AM -0700, Junio C Hamano wrote: > "Johannes Schindelin via GitGitGadget" > writes: > > > This is the next batch of fixes in response to issues reported by Coverity. > > > > Changes since v2: > > > > * Added a new commit to handle block-writer initialization errors (instead > > of ignoring them). > > * The bw->zstream attribute is now also deinitialized in the error case, as > > suggested by Junio. > > * The commit message of "reftable/block: check deflateInit() return value" > > was rephrased to stop suggesting that silent corruption by zlib would be > > possible before that patch: This turned out to be provably incorrect. > > * When aborting the bisect because dup2() failed, a left-over saved_stdout > > is now also cleaned up. > > Everything looks sensible. I am fine with declaring victory, but > does anyone want to second it? I cannot claim to have read all of the patches carefully, but this version addressed the sole concern I raised, and in the few other patches I glanced over I didn't see anything to complain about. So maybe consider that a weak second. :) -Peff