On 10/11/2013 11:56 AM, Reed Underwood wrote: > I wonder if it would be alright to implement a > full Julian date in the date command's format options > (maybe a %J?). It could be really useful in shell > scripts, etc.
Can you give an example of what you mean be a full Julian date? Furthermore, can you compose a full Julian date using existing primitives already provided by date, or is there something fundamentally unique to your proposed format that requires burning a letter? We are really reluctant to burn a non-standardized letter without good cause (burning %N to display nanoseconds is such a case - it was information not available under any other existing format). > > I'm willing to implement the feature if it's > decided it's useful and worth the trouble. I'm not even sure what feature you are trying to add, and why it justifies burning a letter to add it. Scripts are unable to take advantage of the new letter until they have a new enough build of date that includes the extension (may take years); whereas if you can do it in longhand with existing options, they can use the longhand now. Furthermore, rather than burning a %letter, we have lately tended to add pre-canned formats: 'date --rfc-3339=seconds' being an example that generates a given format without you having to remember which % options build it up. For scripting, generating a common canned format via a command-line option rather than burning a new arcane %letter would be easier. -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org
signature.asc
Description: OpenPGP digital signature
