Re: MUST or SHOULD in window gravity?

[Sasha_Vasko osca state mo us]

> Sasha_Vasko said:
> > It has to be MUST since all that comes not from this specs, but instead
> > from ICCCM.
>
> If the ICCCM is wrong, then this spec should override the ICCCM.

ICCCM is not wrong, it just obscure, that's why this section was added to
explain in greater details how things works.

> The spec already says that it overrides the ICCCM & it does so in a few
places.
> It certainly seems odd for this spec to mandate frame decoration
placement.

Here is some more explanation :

It really is not odd when you come to work with it in real life. This rules
are
the only way for clients to estimate its frame position as related to its
own
position, and how different configure requests will be handled based on the
gravity.
Another reason why that should be mandated is that if user specifyes
geometry as a
command line parameter - it expects it to be honored, and the only way for
it
to be honored is to mandate specified behaviour to Window Manager.
Of course Window Manager is free to disregard configure request entirely
or in parts, and this specification by no means interfer with this ability.
What is mandated is HOW request should be handled if IT IS honored.

Unfortunately this part of ICCCM is what is most widely misunderstood,
and this misunderstanding is causing most unpleasant problems in the real
life.

>
> However, if it's the opinion of this group that a window manager MUST
> place frame decorations this way (as stated by the ICCCM), then this
> "will" should become "MUST".  That way the text is no longer ambiguous,
> and it becomes consistent with the rest of the document (including this
> section).
>
>
> --- David A. Wheeler
>     dwheeler ida org

Cheers
Sasha.