bug#55696: 28.1; eshell fails to respect text-scale-increase

[Jim Porter <jporterbugs@gmail.com>]

On 6/10/2022 8:58 AM, Eli Zaretskii wrote:
>> Cc: 55696@debbugs.gnu.org, jeff.kowalski@gmail.com
>> From: Jim Porter <jporterbugs@gmail.com>
>> Date: Fri, 10 Jun 2022 08:44:22 -0700
>>
>> On 6/9/2022 10:52 PM, Eli Zaretskii wrote:
>>> What if there's an external command named 'listify' -- is there any
>>> chance that the test will invoke it?  If so, I think we should come up
>>> with a safer solution.
>>
>> No, that shouldn't be a problem. When Eshell executes something that
>> looks like a shell command[1], the order of precedence is this:
>>
>> 1. If an alias for the command exists in .emacs.d/eshell/aliases, use
>> that. (This won't happen in tests, since `eshell-test-command-result'
>> binds `eshell-directory-name' to a temp dir.)
>>
>> 2. If there's a Lisp function named `eshell/COMMAND', use that.
>>
>> 3. If there's an external command named `COMMAND', use that.
>>
>> 4. If there's a Lisp function named `COMMAND', use that.
> 
> So the problem with using 'list' was that there was no Lisp function
> eshell/list, is that right?

Correct. I think `listify' (aka `eshell/listify') is the preferred 
Eshell way of expressing that for the command-style syntax.

Maybe an `eshell/list' function would be good to have, but I'm not sure 
this problem would come up outside of the unit tests. They're really 
just trying to exercise all the different ways to expand variables, so 
the actual commands aren't very practical.