[2908] in Release_7.7_team
update: Sun OpenGL
daemon@ATHENA.MIT.EDU (Alex T Prengel)
Wed Aug 15 14:51:29 2001
Message-Id: <200108151851.OAA01118@allegretto.mit.edu>
To: ghudson@MIT.EDU
cc: alexp@MIT.EDU, ajfox@MIT.EDU, release-team@MIT.EDU
Date: Wed, 15 Aug 2001 14:51:25 -0400
From: Alex T Prengel <alexp@MIT.EDU>
Hi, just got your mail as I was adding the info below the dotted line
I wrote a few minutes ago-
>Basically, we want to figure out what patches are required, how much
>impact they could have on the release, and then decide whether it's
>reasonable to apply them just before fall term starts.
This is fine with me- I'd like to stay as current with OpenGL as possible,
but as I note below, updating it may not fix this particular problem.
We have a work around for Maple that allows most Maple graphics
functionality to work even if we can't fix it.
>Looking at the OpenGL required patches list, I think the only patches
>we need to apply are 108604, 108605, and 108606, the driver upgrades
>for Elite3D, Creator3D, and PGX. (Maybe not the PGX patch, since I'm
>not sure we have any PGX machines.) Those should have relatively low
>impact on the release.
Makes sense- I installed some of these in doing the tests I ran.
>As for OpenGL 1.2 itself, miki will have to look into that to figure
>out exactly how to install it. Sun's shar file is convenient for the
>administrator of a single machine, but our release process can't take
>advantage of it; we deal in terms of Sun packages (preferrably from
>the Solaris 8 CD) and patches.
I didn't study it through in detail, but from the output it looked like
it was updating packages as it ran- so it may not be so bad after all;
Miki can check..
>> 2, In attempting to install the latest OpenOffice build on Suns, the
>> installer halts and won't proceed until a couple of required C++
>> patches are installed- these are 108434-01 and 108435-01, neither of
>> which we appear to have. Could these also be installed?
>I don't think this will be a problem, since C++ doesn't tend to impact
>core release functions.
Great- thanks!
Alex
***********************
An update to my mail earlier today on this subject:
I did a test installation of Sun's combined 32 and 64 bit OpenGL 1.2.2
package (archive file ogl122_rt32_64.shar) on two machines, an
Ultra 10/Elite 3D graphics machine (allegretto), and a Blade/Expert 3d
graphics machine (central). Unfortunately neither fixed the Maple OpenGL
problem we are seeing but I still think we should upgrade to OpenGL
1.2.2.
On the Ultra 10 the OpenGL installation (along with required patch
installation) went smoothly but the problem persisted unchanged, It
passed all the tests on Sun's installation test page
http://www.sun.com/software/graphics/OpenGL/Developer/verify_install.html
except that package SUNWxfb (Elite 3d Common Driver Routines) is not
present- however this might be due to the test page being out of date;
I noticed that the file checks return files that seem newer than the
ones listed on the test page.
On the blade a couple of things seemed to be off:
1. Before the update, I ran /usr/openwin/demo/GL/ogl_install_check
and was getting a frame rate of about 7 fps; on the Ultra 10 (before and
after) this was about 70 fps.
2. After the blade update, /usr/openwin/demo/GL/ogl_install_check was no
longer there; also the installer reported that installation of
SUNWglrt had partially failed- I'm not sure if these are related. Other
packages installed OK.
3. The Maple error on the blade was different after the update- originally
it was:
X-Error GLXBadDrawable...
after it was:
X-Error BadAlloc - insufficient resources for operation...
but this is about as far as I can go.
