Bill Sommerfeld wrote: > Secondly: I think ALLOW_EXTENSIONS is not necessary. This is another > case where an extension gives meaning to what was formerly a syntax > error. > > There appears to be extensive implementation experience with this sort > of extension on other platforms. In the absence of evidence that there > is real crontab-manipulating code out there which will be broken by this > extension, I don't think we need the ALLOW_EXTENSIONS knob (except > *possibly* as a hedge to enable backport to an older release, and even > then I'm skeptical).
I agree with this and I particularly like how your rationale makes the distinction between full crontab file vs single entry. The only reason that I encouraged Chris to put in the knob was "standards fear fud". I think we need Don (or someone else intimate with the standards that cover cron/crontab) to comment on this. In particular do the standards tests actually put "garbage" in the crontab files and check that it fails ? knobless is best. -- Darren J Moffat