Re: [STUMP] Just a bug fix

[Manuel Giraud]

Shawn Betts <address@hidden> writes:

> This hunk looks good except for the bypass-maximize part. It's in CVS.

Ok, I see that you say below that the *bypass-maximize* list is
empty so it is useless. But, personally I appended "mplayer" to this
list in my personal stumpwm init file ... I set this list to nil by
default because I don't want to force people.

> [SNIP - vertical/horizontal lists hunk]
>
> In CVS, the current window in the window list is highlighted with an
> inverted bar that stretches the length of the window. But your code
> only highlights the window name.
>
> Also, I decided not to implement a horizontal window list because I've
> found the vertical list is much more readable and can display more
> windows without running off the edge of the screen.
>
> So I haven't applied this hunk to CVS. Are there other reasons to
> have a horizontal list that I've overlooked?
>

Hum, I don't use the horizontal list very much myself. But there's one
thing I think you have overlooked in this patch : for very very long
vertical list an automatic split is done. I've noticed the bug of
inverted box: maybe it should be corrected but it is bit more
complicated with a splitted vertical list (you can't say invert window
width wide).

> Why have you wrapped the function in (when f ..)? Was it to fix a bug?
> Which one?

Not really good bug fix. This is my last change : when you do a
'focus-frame-sibling' when there's no frame at all this will hang the
underlying lisp.

> Rather than have an editable-prompt-p, why not have an initial-input
> string you can pass in which gets put into the input area? That way
> you can still have a prompt string.

Yes it could be better. I haven't think of it because I usually use it
for interaction string without a "real" prompt like : "firefox http://";

> The reason I put 100 there was so the input window didn't shrink to a
> tiny box. I'd rather make it customizable than get rid of
> it. Comments?

Don't remember why I remove the 100 here but i think it is part of the
split and horizontal/vertical window list hack.

> No window will be left unmaximized.

See above :)

> I figured if we're going to parse the display string we should do it
> completely. So I wrote one that pulls out the host and display. I used
> port:getenv instead of xlib::getenv since xlib's fn isn't not
> exported.

Cool, this was a ugly and not really portable hack I've done.

> [SNIP - remove set-default-bindings fn]
>
> Why remove the function? I think it's convenient to have a function
> that restores the default bindings.

I remove this because it annoys me :) In fact, I modify the keybinding
hashtable *before* I call stumpwm:stumpwm so it will erase my personnal
keybindings when call.

>> -     (date-string (multiple-value-bind (sec min hour dom mon year dow)
>> +     (date-string (multiple-value-bind (sec min hour dom mon year dow dst_p 
>> timezone)
>
> Why did you add dst_p and timezone? For completeness? :)

I think that one big day I decided to internationalize this function
function a bit ... and then give up ;)

> I think this could be done better with my initial-input comment above.

Maybe.

>> -oeutnh(l (tree-accum-fn sib (lambda (x y) x) (lambda (x) x))))
>> +                   (screen-current-frame screen)))
>> +     (l (tree-accum-fn sib (lambda (x y) x) (lambda (x) x))))
>
> That's nice and embarrassing :).

Dvorak typist? :)

-- 
Manuel Giraud (CNRS/CETP)