Change (setf (point) ...) to (goto-char ...)

[polaris64]

Emacs version 29.0.50 has deprecated the use of forms such as (setf (point) n). The warning emitted is: -

Warning (bytecomp): ‘point’ is an obsolete generalized variable; use ‘goto-char’ instead.

[alphapapa]

Frankly, I do not understand why that deprecation was made, and I do not like it. It's a useful paradigm, and I don't see any reason to throw it away.

Anyway, Emacs 29 is yet to be released. So while I appreciate your keeping up with things, I don't want to merge this yet. I'd like to find out why it was done and try to get the decision reversed.

[polaris64]

OK that's absolutely fine. The main reason for this PR was to make you aware of the potential issue with Emacs 29 (in case you weren't aware already) and to provide a fix if it ends up needing to be applied.

[alphapapa]

FYI, I just posted a message to emacs-devel asking for these generalized variable setters to be un-obsoleted. (Not that I expect they will do so, but at least we can talk about it, I hope.)

And BTW, it's generally better to rebase your PRs than to merge master back into them. It makes for a much simpler commit history.

[polaris64]

FYI, I just posted a message to emacs-devel asking for these generalized variable setters to be un-obsoleted. (Not that I expect they will do so, but at least we can talk about it, I hope.)

Thanks for letting me know, yes I just read the conversation so far this morning. I agree with you that the setf form is quite nice to use (makes it feel a bit more like Common Lisp) so it'll be interesting to see what they decide. I also didn't realise that marking something as "obsolete" was the first step in a discussion about whether something should be removed, I thought that had already been discussed and decided upon already!

And BTW, it's generally better to rebase your PRs than to merge master back into them. It makes for a much simpler commit history.

OK no problem, I'll rebase this PR onto the current master in a bit.