[ghemical-devel] A stupid question

[Tommi Hassinen]

On Tue, 17 Feb 2004, Jean Br=E9fort wrote:

> Hi,
>
> I have not used a lot ghemical so far, but I plan to. I tried work with
> all available versions,and have the same problem with all of them. When
> I add an object (electrostatic plane, orbital or other) some calculus
> occurs, but nothing is displayed (some times, atoms in the background
> become somewhat darker).

Hmmm... This is interesting, since in a few weeks I have got several
questions like this. But the problem does not appear in my system.

I think the problem is that many of the objects are rendered transparent,
and this transparent rendering suddenly is not working now in some systems
(althought the code has not been changed in a few years). I took an
example molecule, and made an electrostatic plot with default settings,
using the default command string (using the 1.01+ version):

=09add plane esp rb1 138.0 1.0 50 1 0.75

Here is what it looks like here:

=09http://www.uku.fi/~thassine/other/Screenshot1.png

As you can see from the screenshot, the plane object is transparent, and
the colurs range from red-green-blue. All electrostatic plots should look
more or less like this. If you wish to get a solid plot, you can modify
the command string like this:

=09add plane esp rb1 138.0 1.0 50 0 0.75
=09                              ^^^
That is, the two last values are the transparency flag (1=3Dtransparent) an=
d
the transparency (or "alpha") value (0.75 means the plane is 75% opaque,
or in other words 25% transparent). Switching transparency off I get:

=09http://www.uku.fi/~thassine/other/Screenshot2.png

So now the atoms behind the plane are not seen, but please try this and
see whether you can now see the plane.

> Is it a configuration problem or something else? What do I miss?

I'm not using any hardware acceleration in my work box, so it's
software-only rendering. In my home box I have a matrox g400 with DRI
hardware acceleration, but so far I have not seen problems there either.
Both setups are debian "testing".

But this really is already 2nd report of the same problem. That makes the
configuration problem less probable. Perhaps there is a newly-introduced
bug in libmesa or something like that? The plot rendering code is
identical in all versions of ghemical, so I expect all versions behave the
same way.

Please report whether you can see the solid plot.

Regards,

=09Tommi