crontab entry environment variables [PSARC/2007/503 FastTrack timeo

[Hugh McIntyre]

Don Cragun wrote:
> Bill,
> 	I'm not sure what you mean.  The current knob in the spec is
> essentially "enable POSIX non-compliance in all crontab files".  I'm
> suggesting that that only be done as long as there are warnings in the
> appropriate man pages that this "feature", if enabled, should not be
> used in shared crontab files.

I assume the suggestion here (but which is not necessarily spelt out in 
the proposal) is "none of the distributions such as Solaris should ship 
any crontabs with these extensions in place out of the box".  So the 
only way to end up with a crontab file (shared or not) with these 
extensions is via a user or sysadmin editing the file.  The user would 
then be assumed to have read all of the warnings.

Some comment text at the top of the shipped crontab files for root and 
friends might not go amiss, along the lines of "environment variables 
are supported, but they may break POSIX or other scripted additions to 
the crontab files.  Also, such crontab files won't work if copied to an 
older Solaris release.".  Then people can be considered to be warned if 
they break things.

As long as such warnings can be considered sufficient, I agree that 
avoiding a config knob is good.

Hugh.



> 	I think the what you're suggesting is that non-compliance be
> the default and have a knob to enable conformance.  No matter which way
> the knob defaults, if you're going to allow use of this non-conforming
> behavior on some systems, the man pages need to document that shared
> crontabs (at least those that may be modified by POSIX conforming
> applications) should not use these non-conforming extensions.
> 	I was trying to help by supplying what I though were simple
> measures that would allow strictly conforming applications to continue
> working correctly, unchanged even though non-shared crontab files are
> using this new feature.  (No knob required.)
> 	As stated before, other ways of doing this are to have these
> environment settings affect only the single following crontab entry.  I
> suggested that long ago, but it was rejected.  John tried a different
> suggestion this afternoon, but I haven't seen a response to that yet.
> Either of these would allow strictly conforming applications to share
> crontabs with users using the extension without affect the conforming
> application behavior.  (Again, no knob required.)
> 
>  - Don
> 
>> Date: Wed, 12 Sep 2007 17:37:51 -0400
>> From: Bill Sommerfeld <sommerfeld at sun.com>
>>
>> On Wed, 2007-09-12 at 14:17 -0700, Don Cragun wrote:
>>> OK.  Now we're back to a standards conformance problem.
>>>
>>> If you want to inflict unrequested changes in behavior on strictly
>>> conforming POSIX applications that use crontab -e to add crontab
>>> entries to the end of a crontab file; then you need a knob that enables
>>> this extension only on individual crontab files at the request of the
>>> owner of that crontab file.
>> Why can't that compliance knob be "don't include lines which are syntax
>> errors under the POSIX spec"?
>>
>> 						- Bill
> 
> _______________________________________________
> opensolaris-arc mailing list
> opensolaris-arc at opensolaris.org