On 8/13/2022 1:35 PM, brian m. carlson wrote: > On 2022-08-13 at 17:26:05, Junio C Hamano wrote: >> Does FreeBSD offer choices of cron implementations other than Vixie, >> just like some Linux distributions? If somebody on a non-FreeBSD >> platform happens to choose to use Vixie, then they would presumably >> have the same problem, so a compile-time switch, whose default is >> hardcoded based on the target platform, would not work very well. >> The default will be wrong for some users, and users can later choose >> to switch between different cron implementations. > > I'm using Debian unstable, and I'm using Vixie cron. I believe that's > the default implementation. However, I could also well use cronie, > since that's available in Debian as well. So, yeah, I think this is a > thing to consider. > >> Configuration knob can be used as a workaround, but in this case, I >> am not sure if it is worth doing. What's the downside of securely >> opening a temporary file and write whatever we are currently piping >> to a spawned "crontab" command and then giving the path to that >> temporary file to the "crontab" command? Wouldn't that give us the >> maximal portability without that much code, no? > > I think we should try to provide an option which works across at least > the versions on a particular OS. The temporary file seems like a nice, > portable option, so I think we should just do that unless there's some > practical objection. > > If Derrick doesn't get to it this next week, I can send a patch. I agree that the tempfile approach makes the most sense in terms of what we can do within the Git codebase. I won't be able to get to this change this week, so I'd be happy to review one of yours, brian. Be careful to test manually when making this change, because our tests don't actually interact with the system's crontab and instead verify the interaction using replacement commands. Thanks, -Stolee
