Re: [PATCH/RFC] grep: add a grep.patternType configuration setting

[J Smith <dark.panda@gmail.com>]

On Fri, Aug 3, 2012 at 12:39 PM, Junio C Hamano <gitster@pobox.com> wrote:
>
> Much nicer than the drafts in earlier discussion.  One micronit is
> that probably it is preferrable to do '!=' instead of '>', as we are
> only trying to see if it is unspecified on the command line (hence
> need to pay attention to configuration) and do not care how the
> enums are ordered, but that is nothing that requires a re-roll.

True enough. Old habits I guess, eh?

> All the new tests in the script looked very well thought out.
>
> I noticed that this particular one will still succeed when somebody
> breaks your code to ignore the configuration (as -G "a+b" would give
> the expected match) or give higher precedence to the configuration
> (as fixed "a+b" also will give the expected match).  Not that it is
> wrong to have such a test that is unlikely to catch certain kinds of
> regressions in the suite, and the particular kind of breakage will
> be caught by the next test (snipped) anyway.

Yeah, that test is kind of iffy, but does no harm I suppose.

> Will queue.  Thanks.

Great, thanks. Been an informative experience for a first-time git
patcher. Cheers.