bug#34765: 26.1; with-temp-buffer should not run buffer-list-update-hook

[Basil L. Contovounesios]

martin rudalics <rudalics@gmx.at> writes:

> If you think that patch is of some use, please try to
> polish it up and push it.  Honestly, I've forgotten all about it.

Sure, I'll give it a go.

I'm slightly confused about the conclusion of this thread, though.
You and Eli said it was too radical to set inhibit_buffer_hooks based
on whether a buffer name starts with a space, as there may be legitimate
cases where those hooks should run in a temporary buffer.

I accept this, but I don't see how setting inhibit_buffer_hooks based on
a new optional argument to Fgenerate_new_buffer (and/or
Fget_buffer_create) solves the problem entirely.

If I create a temporary buffer with the proposed

  (generate-new-buffer "foo" t)

then how do I later tell Emacs that this buffer's hooks should run?

In other words, can we be sure that the buffers we choose to create with
inhibit_buffer_hooks set will *never* need to later unset it?

Should we expose a getter or setter for this buffer member, or is that
opening ourselves up to abuse?

I suppose if someone *really* wanted to un-temporarify a buffer with
inhibit_buffer_hooks set, they could create a new, non-temporary buffer
and swap buffer contents, or something to that effect?

Am I wildly misunderstanding something here?

> And also tell us why 'buffer-list-update-hook' would be still needed now
> that we have 'window-selection-change-functions'.

I've yet to look into these hooks and their precise semantics, but are
you suggesting that one might obsolete the other, in which case we
should deprecate the old one?

Thanks,

-- 
Basil