Re: [PATCH] Fix display-buffer-use-some-window to honor reusable-frames

[martin rudalics]

> Is this related to display-buffer-use-least-recent-window and/or
> window-bump-use-time?  Or is this another set of issues?

It's another set of issues loosely related to
'display-buffer-use-least-recent-window'.  I found it because I tried to
understand why that function did not always return the LRU window.  See
my other mail to Tom.

> Would it be possible to modify the effect of
> display-buffer-avoid-small-windows so that it works via the
> window-min-height alist entry?

No.  'window-min-height' so far is honored only in functions of the
'display-buffer-in-direction' type because there it's easy to tell which
window should have that property.  It's not used in 'get-lru-window'
because that function is not used for buffer display alone.  As a rule,
buffer display options should never show up in functions that can be
used in another context.

We could give 'get-lru-window' an additional ALIST argument which would
indicate that it was indirectly called by 'display-buffer'.  I wouldn't
like it and I don't think we would be able to figure out all the
consequences in due time.  So we need a 'get-lru-window' clone that can
be used by buffer display and takes only an ALIST as argument and
nothing else.

'display-buffer-avoid-small-windows' is an attempt to reintroduce a
concept we tried to abandon many years ago.  It's inherently like
'pop-up-windows' where applications and users fought each other by
setting or binding that variable to their like with undefined outcome.

We tried to stop such fighting by giving users options like
'display-buffer-alist' and 'display-buffer-base-action' and giving
applications the ACTION argument with clear precedence rules.  There's
no precedence rule for 'display-buffer-avoid-small-windows'.

martin