On 03/11/20 23:20 PM, Nicolas Goaziou wrote:
> Hello,
>
> Eric Abrahamsen <e...@ericabrahamsen.net> writes:
>
>> But if we use `make-obsolete-variable', the CURRENT-NAME arg can be a
>> simply explanatory string.
>
> You're right. However, I'm not sure what the CURRENT-NAME should be,
> besides "don't use this, you probably want something else"

I would have liked to know what the "something else" was! Or even "Org
link regexps have been rewritten", something like that.

> Also, my suggestions still holds: it is useful to warn upstream about
> it.

Absolutely. I don't think Org has a responsibility to maintain backwards
compatibility for these variables, and upstream packages should be
tracking changes. But the more help we can provide, the better.

>> I was also recently bit by the removal of a bunch of regexps (in my
>> case, link regexps), and it would have been useful to have some sort of
>> a pointer, either in the obsolescence message or in the docs, about what
>> we're supposed to do instead.
>
> I'm not sure about what "bunch of regexps" you are talking about.

Looks like commit "Move link-related core functions out of \"org.el\"".
Nearly a year ago -- I wasn't paying attention! But defining link
regexps as aliases of other link regexps meant that the regexps matched,
but the match groups were off: that led to silent failure, and took
quite a while to debug. I guess I would have preferred a loud failure.

Anyway, it's not a big deal, I only bring it up because someone else
did!

Eric

Reply via email to