bug#41340: 26.3; ido-completing-read doesn't respect completion-auto-help

[Ryan C. Thompson]

On 5/16/20 8:57 PM, Dmitry Gutov wrote:
> On 17.05.2020 03:12, Ryan C. Thompson wrote:
> > One minor wrinkle is that completion-auto-help has 3 valid values: 
> > nil, t, and lazy. I'm not sure I understand the difference between t 
> > and lazy, so I'm also unsure as to how to apply them to ido. In my 
> > testing, both t and lazy seem to result in the identical behavior in 
> > completing-read-default.
>
> When it's `lazy', you need to press TAB twice to see the completions 
> list. At first it'll just say "Next char is not unique".
>
> If you can honor both values, that would be best, but even making the 
> nil/non-nil distinction will be an improvement, sounds like.

Hmm, that's not the behavior I see, even in emacs -Q. Regardless of 
whether it's `t' or `lazy', the first TAB press completes any common 
prefix, and then the very next TAB press shows the completions buffer. I 
never get the "Next char is not unique" message (I checked the Messages 
buffer to be sure).

I don't have time to dig deeper into this right now, but I'll come back 
to it, probably on Monday.