Re: [PATCH] epm.el: A CLI tool for package.el

[Max Nikulin <manikulin@gmail.com>]

On 04/05/2023 17:24, Ihor Radchenko wrote:
> Max Nikulin writes:
>> For those who are not familiar with convention any target may be
>> specified in docs, keeping usual meaning of the default "all" target.
> 
> I am mostly concerned about the existing users who are already settled
> on running the default "make". Their workflow will be broken with your
> suggestion.

I hope, it is possible to achieve helpful error message with a link 
explaining change.

>> An I am unsure concerning user prompt.
> 
> My main concern is for the users who changed their elpa directory
> location. They might run into issues unexpectedly.

My opinion is that an error is better for make than user prompt. Try to 
load from default `package-user-dir' but install there only if it is 
explicitly enabled.

>>>> #!/bin/sh
>>>> ":"; # -*- mode: emacs-lisp; lexical-binding: t; -*-
>>>> ":"; exec emacs --script "$0" "$@"
>>>
>>> Let's not lock to bash. AFAIK, our makefiles can currently work on
>>> Windows. Using /bin/sh will lead to regression.
> 
> I thought that more portable way is using "env".

Either env would invoke sh anyway (like in org publish scripts) or it is 
necessary to use env -S feature of relatively recent GNU env. For 
cmd.exe it is better to provide a .bat file running

     emacs --script epm.el ...(something for POSIX $@)

>> +(defun epm-nonempty-p (s)
>> +  (and s (not (string-empty-p s))))
> 
> Can just use `seq-empty-p'.

I am neutral in respect to such change. I just decided to minimize 
dependencies.

>> +  ;; TODO (load site-run-file 'no-error 'no-message)
>> +  ;; may be necessary to load elpa-* deb packages when -Q option
>> +  ;; is used. See Info node "(elisp) Init File".
> 
> Given the explanation in the top comment, is this necessary?

Do you have any arguments against switching from emacs -Q to emacs -q? I 
am still unsure if it is safe enough, but not doing it is not better. I 
am considering adding -q to this script when it is called directly. 
Unfortunately Emacs does not have an option to override effect of -q.