Re: [Nmh-workers] pick -size

[Arun Bhalla]

Simon Burge writes:
> I was cleaning out an old folder that had lots of daily status messages,
> when it occurred to me that messages less than a certain size didn't
> contain any useful information.
> 
> I was thinking of using a rule similar to what find(1) uses, where
> "-size -NUM means "less than NUM", "-size NUM" means "exactly NUM", and
> "size +NUM" means "greather than NUM".  I think it would also be useful
> to allow say "l" and "b" suffixes to mean "lines" and "bytes", with
> bytes being the default.
> 
> Does anyone else see the usefulness of adding a size-based search rule
> to pick?  I'll probably start working on a patch to do this anyways...

I think that's a great idea; it's interesting that pick(1) doesn't
have this already.  Maybe also add the other common units, such as 'k'
for kilobytes, 'm' for megabytes.  Note that find(1) uses 'b' for
512-byte blocks and 'c' for bytes.

I wonder if it should count only the body or the full size (including
headers).  I could see it going either way, but the lines option would
imply body-only to me.  It might not hurt to have one predicate that
counts the body-only and another predicate that includes the headers.
Header size might be less interesting, but it's a possibility for a
third predicate.

Arun