please change default value of truncate-partial-width-windows to nil

please change default value of truncate-partial-width-windows to nil

[Drew Adams]

This new variable should not have been given a non-nil default value. Non-nil
overrides the user's preference as expressed via `truncate-lines'.

If a user has nil `truncate-lines' (which is still the _default_ value), then
lines should not be truncated. End of story. That's the meaning of
`truncate-lines' (always has been).

If you want the new variable `truncate-partial-width-windows' to override the
`truncate-lines' behavior, then let it do so for only a *non*-nil value of
`truncate-lines'.

Actually, the most reasonable approach might be to not introduce a new option
(with the confusion of variable interaction and override), but simply to let
`truncate-lines have three possible (buffer-local) values:

nil     - never  truncate lines in the current buffer
t       - always truncate lines in the current buffer
integer - do what integer `truncate-partial-width-windows' does today

(Customize sets the default value, not the buffer-local value, of course.)

Regardless of what you think of that suggestion, the important thing is this: If
`truncate-lines' is nil, then please just do what it says: do not truncate
lines, period.

Re: please change default value of truncate-partial-width-windows to nil

[Johan Bockgård]

"Drew Adams" <drew.adams@oracle.com> writes:

> This new variable

has existed since 1985.

RE: please change default value of truncate-partial-width-windowsto nil

[Drew Adams]

> > This new variable
> has existed since 1985.

You're right. And with value t, no less.
What is new is the default value of 50.
I stand corrected.

(I still think a non-nil default value is wrong, FWIW, for the reasons I gave.)

Re: please change default value of truncate-partial-width-windows to nil

[Stefan Monnier]

> Actually, the most reasonable approach might be to not introduce a new
> option (with the confusion of variable interaction and override), but
> simply to let `truncate-lines have three possible (buffer-local)
> values:

> nil     - never  truncate lines in the current buffer
> t       - always truncate lines in the current buffer
> integer - do what integer `truncate-partial-width-windows' does today

I think this would make a lot of sense.
I.e. first extend truncate-lines so it can take an integer value, then
mark truncate-partial-width-windows as obsolete.


        Stefan

Re: please change default value of truncate-partial-width-windows to nil

[Miles Bader]

Stefan Monnier <monnier@iro.umontreal.ca> writes:
>> nil     - never  truncate lines in the current buffer
>> t       - always truncate lines in the current buffer
>> integer - do what integer `truncate-partial-width-windows' does today
>
> I think this would make a lot of sense.
> I.e. first extend truncate-lines so it can take an integer value, then
> mark truncate-partial-width-windows as obsolete.

Eh, why is that better?

`truncate-partial-width-windows' has always[*] been essentially an
"exception for a funny situation", and wedging its functionality into
`truncate-lines' merely complicates things for typical code which really
only cares about the normal case, not the exceptional one.

Currently a mode only needs to worry about setting truncate-lines to t
(or toggling between nil and t in the case of Gnus, etc), if truncating
lines is somehow better for that mode.  I your suggestion is followed,
such modes will suddenly have to worry about such details as the proper
width at which point "emergency truncation" goes into effect -- and I
would argue that most modes just don't care about such detail (if a mode
_does_ care about that, it's easy enough to set both variables).

[*] and of course `truncate-partial-width-windows' has been around for
decades, with a non-nil default; I dunno why Drew is freaking out about
it _now_...

-Miles

-- 
Logic, n. The art of thinking and reasoning in strict accordance with the
limitations and incapacities of the human misunderstanding.
n

Re: please change default value of truncate-partial-width-windows to nil

[Stefan Monnier]

>> I think this would make a lot of sense.
>> I.e. first extend truncate-lines so it can take an integer value, then
>> mark truncate-partial-width-windows as obsolete.

> Eh, why is that better?

This way it also applies in the case where your narrow windows are the
result of your tiling window manager creating narrow frames rather than
a wide frame split into narrow windows.

> `truncate-partial-width-windows' has always[*] been essentially an
> "exception for a funny situation", and wedging its functionality into
> `truncate-lines' merely complicates things for typical code which really
> only cares about the normal case, not the exceptional one.

I can't see which complication you're referring to.

> Currently a mode only needs to worry about setting truncate-lines to t
> (or toggling between nil and t in the case of Gnus, etc), if truncating
> lines is somehow better for that mode.  I your suggestion is followed,

I only know of modes that set it to t, and for those there's
no difference.  Which code ever sets it to nil?

> [*] and of course `truncate-partial-width-windows' has been around for
> decades, with a non-nil default; I dunno why Drew is freaking out about
> it _now_...

Yes, I'm not worried about that indeed, I just feel like using 2 vars,
the second of which is only used in some cases but not all is
rather odd.

Maybe another option is to make truncate-partial-width-windows apply to
full-width windows as well when its value is an integer (so it also
applies when your window-manager makes your frames narrow).


        Stefan

Re: please change default value of truncate-partial-width-windows to nil

[Miles Bader]

Stefan Monnier <monnier@iro.umontreal.ca> writes:
> I only know of modes that set it to t, and for those there's
> no difference.  Which code ever sets it to nil?

Gnus has a command that toggles it between t and nil.

> Maybe another option is to make truncate-partial-width-windows apply to
> full-width windows as well when its value is an integer (so it also
> applies when your window-manager makes your frames narrow).

Sure, that seems to work; a new name would be called for I
think... (truncate-narrow-windows?)

-Miles

-- 
`...the Soviet Union was sliding in to an economic collapse so comprehensive
 that in the end its factories produced not goods but bads: finished products
 less valuable than the raw materials they were made from.'  [The Economist]

RE: please change default value of truncate-partial-width-windows to nil

[Drew Adams]

> and of course `truncate-partial-width-windows' has been around for
> decades, with a non-nil default; I dunno why Drew is freaking 
> out about it _now_...

Drew is not freaking out about anything.
_You_ seem to be freaking out about this.

Drew said the following, as soon as Johan reminded him that
`truncate-partial-width-windows' is not new:

> You're right. And with value t, no less.
> What is new is the default value of 50.
> I stand corrected.
> 
> (I still think a non-nil default value is wrong, FWIW, for 
> the reasons I gave.)

That was Drew's second, and last, input to this thread.

If that's "freaking out", then I hate to think what your alarmist and ad hominem
reaction is.