Re: Add compat to ERC

[Philip Kaludercic]

"J.P." <jp@neverwas.me> writes:

> Philip Kaludercic <philipk@posteo.net> writes:
>
>> I'd be glad to help, I just don't use ERC so I'd need a basic
>> configuration to contribute materially.
>
> Cool! Glad you're on board. Believe it or not, ERC thinks it can read
> minds and so only offers "ad hoc" connections. (And they wonder why
> we've had such a smooth go of things.) Something like this should work
> in theory:
>
>   (erc-tls
>    :server "testnet.ergo.chat"
>    :port 6697
>    :nick "testing123"
>    :full-name "Testing 123")
>
> Of course, now comes the part where you spend the next day reading
> EmacsWiki! (Seriously, though, please let me know if that doesn't work.)

Ok, I will keep this in mind (FYI, I had used ERC in the past so I am
not totally unfamiliar, but I switched to rcirc a while back).

>>> diff --git a/lisp/erc/erc-compat.el b/lisp/erc/erc-compat.el
>>> index 16cfb15a5a..899522f186 100644
>>> --- a/lisp/erc/erc-compat.el
>>> +++ b/lisp/erc/erc-compat.el
>>> @@ -26,6 +26,7 @@
>>>  ;; This mostly defines stuff that cannot be worked around easily.
>>>  
>>>  ;;; Code:
>>> +(require 'compat nil 'noerror)
>>
>> I guess this is the main part.  Have you tested it to see if it works?
>
> Well, tested it in the lazy/naive sense of
>
> 1. connected to various servers in the manner shown above

Using what version of Emacs?

> 2. installed a fake ELPA package and ran our test suite, which also
>    contains some shims for obscure stuff, like gv-place, ert, etc.

What do you mean by fake ELPA package?

>>> +(defmacro erc-compat--decoded-time-period (time)
>>> +  (declare (indent defun))
>>> +  (list (if (< emacs-major-version 28)
>>> +            'erc-compat--28-decoded-time-period
>>> +          'decoded-time-period)
>>> +        time))
>>
>> If you want, we can upstream these definitions into compat.  I would
>> guess others are interested in them too.
>>
>> My wish would be to one day replace all custom *-compat.el files in the
>> core with compat.
>
> That'd be superb. And there's no rush, obviously. Really, though, most
> of the stuff in ours is just cruft waiting to die. So, from us, it'd
> only be that one you highlighted: `decoded-time-period' from Emacs 28.

I have just pushed a commit with a definition for decoded-time-period,
so that will be part of the next release.