[illumos-Developer] [REVIEW] 244: Need replacement for closed /usr/bin/tail and /usr/xpg4/bin/tail

[Garrett D'Amore]

On Mon, 2010-09-27 at 19:05 +0200, I. Szczesniak wrote:
> On Mon, Sep 27, 2010 at 6:51 PM, Garrett D'Amore <garrett at damore.org> wrote:
> > Can we talk about this?
> >
> > Chris has finished his work and is near to integration already.  Is
> > there some compelling reason why the ksh93 version is "better"?
> 
> Please clarify what you mean with "better".
> The AST tail code is significantly faster than the code from the
> webrev Chris posted and faster than GNU tail.

Have you tested this?  Quantified it?  One of the problems I understood
that existed with ksh93 tail was that it tended to grow unreasonably
large in memory consumption when given -f.

>  It is maintained and
> uses the upstream code in unmodified form, which should result in
> shorter maintenance cycles.

Possibly.  I don't put that much faith in upstreams in general, because
they have their own agendas, which may or may not match with our own.
I'm more interested in having code that is simple enough for us to
maintain on our own if we have to, then having very large complex code
bases that we *have* to rely on the upstream to maintain.

>  It has much more features, many of them
> derived from GNU *and* BSD.

Please quantify this.  What features are you talking about here?

For the record, I'm going to let cjlove proceed with his integration,
but he and I have agreed that there is a compelling reason to replace
that work with a ksh93 or other version, then that will be perfectly
fine.  I just want to have something quantifiable behind he "compelling
reason".

	- Garrett