Re: [dev] [st PATCH] X geometry changes

From: Christoph Lohmann <20h_AT_r-36.net>
Date: Fri, 25 Apr 2014 21:16:50 +0200

Greetings.

On Fri, 25 Apr 2014 21:16:50 +0200 suckless_AT_dev97.com wrote:
> From: Yuri Karaban <suckless_AT_dev97.com>
>
> Hello,
>
> Currently st does not follow X geometry specification. Implementation
> is incomplete and partially broken:
>
> - There is calculation of offset parameters, but these parameters are
> never saved in window manager hints. Therefore setting offsets has
> no effect.

This is the only part I may include, to have the offset set in the size
hints.

> - Instead of specifying initial geometry, current implementation locks
> window dimensions to specified geometry and window becomes
> unresizable.

Suckless is writing software for dwm. A specified geometry is fixed.
Otherwise st is not floating by default when forcing a size.

> - Current implementation interprets window dimensions as pixels, while
> common practice for text oriented applications (like terminals and
> editors) is to treat witdh and height as rows and columns.
>
> XWMGeometry(3) function is interpreting the width and height as
> multipliers for width increment and height increment hints.

This is basically wrong.

> Also xwininfo(1) is showing geometry in rows and columns if window
> have resize increment hints.

Then fix xwininfo and X11 to include both hints. Interpreting given ar‐
guments in different ways based on a subjective valuing of applications
is wrong in all senses.

> And finally width and height in pixels don't make sense for the
> window with discrete dimensions.

The pixel specification is needed in the tiling world, to have a correct
positioning. Here X11 is doing it wrong. Don’t overlap windows. St is
written to only take the parts it can really use.

Another note is, that you are not using the st style in your patch. Es‐
pecially in the switch and if statement. This won’t be included.

X11 is the enemy, introduced mass dynamic linking and inconsistencies
like you mentioned. I won’t include this.


Sincerely,

Christoph Lohmann
Received on Fri Apr 25 2014 - 21:16:50 CEST