• Previous message: R. Timothy Edwards: "extraction magscale bug"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hi,

I had happily been using Magic 7.1 on a Linux Mandrake installation using
XFree86 3.3.6.  After trying different display depths I found that Magic
worked well with a depth of 16 bits.  (Good redraw rates and no problems
with swapping color palettes between windows.)

I stopped using Magic for a while, then recently tried running again on a
machine that had been upgraded to XFree86 4.0.3.  The color scheme now
looks dreadful, with a lot of opaque stippled white layers.  I had been
used to polysilicon looking kind of salmon colored, with pfet looking like
a stippled version of the same thing.  On the upgraded machine, poly still
looks the same color, but pfet is white with diagonal stippling.

The interesting thing is that if I use the box tool to successively select
the pfet regions, they refresh back to the stippled salmon color that I
expected.  I can walk over the pfet regions and refresh them box by box to
get the correct color on all of them.  However, as soon as the window does
an entire redraw (by dragging the window or various expose events) all the
pfet regions return to the stippled white.  So, an individual redraw of a
tile comes back with the color I'd expect, a redraw of the entire windows
doesn't.

As luck would have it, I have a second machine sitting in my office that
hadn't been upgraded and is still running XFree86 3.3.6.  It displays
things with the correct color scheme when running locally.  What's more,
if I use this as a remote X server for Magic running on the upgraded
machine, the colors also look correct.  This is consistent with the
reverse experiment.  If I use the upgraded machine as an X server running
a copy of Magic from the non-upgraded machine, the colors do NOT draw
correctly.

The problem seems specific to the X server, not to the copy of Magic that
is running.

In all cases, Magic starts up with the message:

   Using TrueColor, VisualID 0x22 depth 16

Anyone have any insights into the problem?  Any known bugs or
idiosyncrasies in the interaction between Magic and Xlib in the newer
releases of XFree86?

Thanks,

-Rob Penny (rob AT ucsd DOT edu)

• Next message: Stefan Robert: "Re: Compiling problems for MacOS X"
• Previous message: R. Timothy Edwards: "extraction magscale bug"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]