bug#54536: 29.0.50; Improve ERC's handling of multiline prompt input

[Lars Ingebrigtsen]

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

> 1. What should happen when a user submits multiline input containing
>    empty lines? Should these be padded so they're not rejected by the
>    server? If so, where in the processing pipeline should that occur?
>    Should `erc-send-whitespace-lines' play a role here?
>
>    This patch says yes to the latter and interprets that option as
>    meaning "preserve whitespace-only lines and create them as necessary
>    from blank ones." As to where padding should happen, this patch punts
>    and retains the existing (unfortunate) practice of treating them at
>    the last minute.

Makes sense to me.

> 2. Should trailing blank lines be treated differently? If so, how?
>    Should they be auto-padded? Simply dropped? Or should encountering
>    them raise an error?
>
>    When `erc-send-whitespace-lines' is non-nil, this patch drops
>    trailing blanks by default, but it also provides an escape hatch.

Dropping trailing blank lines is a good thing, I think.

> 3. When `erc-send-whitespace-lines' is non-nil, should it auto-pad a
>    submission consisting of a single empty line? Should it allow a
>    whitespace-only singleton through?
>
>    This patch says no to the first and yes to the second.

Sounds good.

> 4. Should slash commands, like /MSG be allowed to lead a multiline
>    submission?
>
>    This patch says no, still choosing to interpret commands as always
>    consisting of a single line.

Ditto.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no