| Jim,
Sorry to not get back to you sooner, but I have been on vacation...
The error in .0 has been discussed extensively
in notes #275 and #352 in the CLT::GOBE_WIDGET conference. In some
notes users mention that the error occurs on some types of workstations
but not others (maybe because of differences in the servers?). Here
is the latest word on how to fix it.
...........................beth
===============================================================================
<<< CLT::DISK$CLT_LIBRARY3:[NOTES$LIBRARY]GOBE-WIDGET.NOTE;1 >>>
-< Discussion of GObE widget >-
================================================================================
Note 275.19 Unable to create tracking pixmap. 19 of 19
WIDGET::KLEIN 25 lines 25-JAN-1993 13:23
-< widget initial size >-
--------------------------------------------------------------------------------
I thought that there was a workaround for this discussed previously, but
maybe not.
My experience with this problem is that the tracking pixmap is created
very early in the creation of the widget and that if the widget does not
have a valid size at that time, creation of the tracking pixmap fails.
The "bug" is that the tracking pixmap should probably not be created until
the widget is realized, but the workaround is to give the widget an explicit
initial size other than the default (100x100 pixels works).
During the realization of the widget, the attachments will take effect and
size the widget correctly, but the explicit XmNwidth and XmNheight resources
give an initial size that is legal for the tracking pixmap creation logic.
One thing you might do if you get stuck is to use the debugger to trap
the CreateCallback for the widget. While in the CreateCallback, check
the w->core.width and w->core.height values to make sure that they are
large enough (100x100). I have found that by default, the widget is
10x10 which is not large enough for the tracking pixmap creation logic
to work properly.
Hope this helps.
-steve-
-------------------------------------------------
Now the bad news. We tried this and it did not work correctly. It crushed all
the icons in the domain into a 100x100 pixel area, so we are doing some more
investigation.
In the meantime you could try something for us. We do not have access to a
VAX 4500. Just to make sure everything else has been correctly installed, can
you set the display to another system (non-4500) (set display/node=xxx/cre/perm
and check that MCC comes up correctly there. Additionally, do you have MCC
installed on any other system. If so, does it work ok, and can you set display
to the 4500 and run that way.
If possible, we may want to test any fix we come up with on your system, since
we have no problem on any of our workstations.
Regards,
Verna
|
| Beth,
I read your memo, read the notes conf, but still don't understand
how to correct the problem. Is there some command from vms that i need
to do to make it work.
It does not display on VXT2000, or a Vaxstation 3100. I can however
get it to display on a PC using EXcursions. What gives ...
Thanks,
Jim
|
| Jim,
Thanks, that was a big help. As it turned out there is a logical
called DECW$SYSTEM_DEFAULTS. This logical points to the following
directories in this order.
SYS$COMMON:[DECW$DEFAULTS.USER]
SYS$COMMON:[DECW$DEFAULT.SYSTEM]
SYS$LIBARY:
Well the SYS$COMMON:[DECW$DEFAULT.USER] directory was corrupted and the
system could not access it. We corrected the problem with the directory
and it came up.
Thanks all for your help on this very strange problem. It is fixed.
JIM (Now, its Miller Time !)
|