Re: frame parameter `buffer-list' - bug? if not, what changes it?

[Stefan Monnier <monnier@iro.umontreal.ca>]

> Even if a particular frame has only ever displayed a single buffer, this
> parameter can apparently have various values. For example, here are two
> different values of `buffer-list' for the frame *Help*, which has never
> displayed any buffer other than buffer *Help*:

>  (buffer-list #<buffer *Backtrace*>)

>  (buffer-list #<buffer foobar.el> #<buffer *Help*> #<buffer toss-1>)

> What is going on here? Is this a bug? If this is the intended behavior, then
> just what is that behavior? What determines (changes) the value of
> `buffer-list'? I can find no doc on this.

I'll let you check the source, but my guess is that the value is inherited
from somewhere at frame creation.

It's only meant to be used for things like next-buffer so that you get the
"last buffer used in this frame" rather than the "last buffer used".

> Motive: I'm trying to test for equality (with `equal') between frame
> configurations, and the same frame with all other parameters equal is
> showing up with different values for `buffer-list'.  I don't want to have
> to create and pass a different equality predicate (that would ignore
> `buffer-list'); I want to be able to use `equal', if I can.

I guess you're out of luck.  Two possibilities: change the buffer-list
parameters before creating your configurations, use a hand-written equality
predicate, lobby on gnu.emacs.bug to change the definition of `equal', lobby
on gnu.emacs.bug to change the way the parameter is initialized&updated to
be "more deterministic".  Oops, that's more than 2, sorry.


        Stefan