Thomas Adam <[EMAIL PROTECTED]> wrote: > I'm unable to reproduce this bug whatsoever, and have tested it on > 2.5.19, which at this time of writing is the current FVWM unstable > version.
2.5.19 is concerned as well. Are you absolutely sure you left just(!)
enough room for the xterm, not leaving a horizontal gap where the root
window "shines through"? You should use "xmag -mag 8" to be able to
clearly see every pixel.
> Note that FVWM in its default state (with no user
> configuration file), defaults to using TileCascadePlacment
This should be obvious from my bug report. :-)
> -- and does
> indeed work as expected in that the xterm you mentioned that wasn't
> placed back in the top-right corner of the screen, actually is when
> asked to do so.
What do you mean by "when asked to do so"? At least the default
"TileCascadePlacement" should definitely put the window in the upper
right corner:
,----[ fvwm(1) ]---
|
| TileCascadePlacement automatically places new windows in a smart
| location - a location in which they do not overlap any other
| windows on the screen. If no such position can be found Cas-
| cadePlacement is used as a fall-back method.
|
`----
And in the setup described in my bug report, there is definitely such a
location where the xterm wouldn't overlap other windows. Instead, Fvwm
claims that tiling failed and falls back to cascading, which is not what
it should do (as there is enough room for tiling). Only when increasing
the empty space by one horizontal pixel (by decreasing the xclock's
width by one pixel), Fvwm correctly puts the xterm where it should be.
(At least for 2.5.18, the y-axis doesn't seem to be concerned by this
off-by-one-error)
See attached image for an actual example.
> I urge you to try 2.5.19 if you can, and report back here if you can
> reliably reproduce the bug.
I can reliably reproduce this bug with
2.5.17 (Debian version)
2.5.18 (Debian version 1:2.5.18-2)
2.5.19 (download from fvwm.org, configure && make install)
Please let me know if you need more information or I can test something
else that could help you narrowing down this bug.
Regards
Marc-Jano
placementbug.png
Description: PNG image
