Re: [PATCH] Fix `early-init-file' value when file is missing

[Eli Zaretskii <eliz@gnu.org>]

> From: Radon Rosborough <radon.neon@gmail.com>
> Date: Sun, 10 Feb 2019 15:04:59 -0800
> Cc: emacs-devel <emacs-devel@gnu.org>
> 
> > you've modified the API of load-user-init-file
> 
> But we have never released a version of Emacs that includes this
> function -- and it's an internal function. There are only two callers
> of this function that exist, and they are both touched by this patch.

True, but that's not what bothered me.  What bothered me was that
changes in the interface and the implementation of the function might
adversely affect the other caller of this function in startup.el, and
verifying that no use case will become broken because of that is not a
trivial task.  For a significant improvement ion functionality, such
risks might be justified, but here we are talking about a minor
cleanup, so I'd prefer a safer change.

> > Or do that inside the function, but only if it processes early-init
> > file specifically?
> 
> I really thought that this was exactly what my patch did. There are
> exactly two places where this internal function is called, and
> checking whether the optional argument is provided is the way to tell
> whether the early init-file is being processed.

Such a test is too indirect, and will cease to DTRT if we ever need to
pass a second function in the case of early-init or not pass it in the
other case.  I hope we can make a more direct test.