Re: Should diff.elisp.xfuncname match cl-lib macros?

["Basil L. Contovounesios" <contovob@tcd.ie>]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: "Basil L. Contovounesios" <contovob@tcd.ie>
>> Date: Sat, 16 Mar 2019 15:30:22 +0000
>> Cc: emacs-devel@gnu.org
>> 
>> >> Given the increasing presence of cl-defuns, cl-defgenerics, etc. in the
>> >> emacs.git sources, would it be welcome to additionally match cl-lib
>> >> macros?
>> >
>> > Yes, please,
>> 
>> I can think of three ways to achieve this given the lack of shy groups
>> in EREs:
>
> Is this feature even useful with Lisp?  I find it mostly useless,
> here's a typical example:
>
>   git log -L :next-line:lisp/simple.el
>
> Type this command in your repository, then watch the fun.  More fun is
> available when Git for some reason decides you asked about some
> variable, not a function (and in Lisp we frequently have variables by
> the same name as a function).

I don't think Git ever tries to distinguish between variables and
functions, despite the name of xfuncname.

> So I think until we fix this basic deficiency, extending that to more
> symbols will just add to the mess.  YMMV.

Which deficiency do you mean?  That 'git log -L' picks up
next-line-or-history-element instead of next-line in your example,
i.e. that it's currently inaccurate?  How do you usually use this
feature?  What kind of fix do you envision?  (Sorry, I don't have
experience using this Git feature.)

-- 
Basil