From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============1870209626226317842==" MIME-Version: 1.0 From: Denis Kenzior Subject: Re: [PATCH 1/3] main: Manually initialize and clean up the main loop Date: Wed, 08 Jun 2016 13:31:36 -0500 Message-ID: <57586488.4010005@gmail.com> In-Reply-To: List-Id: To: ell@lists.01.org --===============1870209626226317842== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Mat, >> >> Can you do me a favor and split these into three patches, according to >> HACKING 'Submitting Patches'. > > What do you know, "patches should be split even if breaking compilation > is unavoidable". Curious about why this is (I'm not lobbying for a > change, it's just not self-evident). > We hardly ever use git bisect, so readability of commits trumps the = breaking of compilation. >> >> I wonder if returning bool here is useful since you never check the >> return value in any of the actual files that use this. > > All of the unit tests and examples exit out of main immediately after > calling l_main_exit, but that may not be the case for every program. > l_main_exit will do nothing if the main loop is still running. > > l_main_run and l_main_quit's return values are similarly ignored, but > are potentially useful in a similar way. > Fair enough, just curious. Some like to use void *_exit functions = because you can't do anything if stuff fails anyway. I've had this = argument with myself many times and don't really have a preference. Regards, -Denis --===============1870209626226317842==--