| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 815.1 | VSII ? | DECWIN::FISHER | Burns Fisher 381-1466, ZKO3-4/W23 | Tue May 23 1989 15:04 | 20 |
| What hardware is this? I bet it's a VSII/GPX and not a VS2000/GPX, right?
That hardware has the cursor hardwired to 255 and 256 in the color maps.
What we do in VMS is to preallocate those entries in the default colormap,
but allow you to allocate them yourself in one that you create yourself.
We also return 256 as the number of color indices available.
In MITs server, they return 254 as the number of entries available and never
allow you to allocate the cursor colors.
It is possible that the Ultrix server was changed to match the MIT convention,
although I'm not sure. I thought that code was in the sources that were
common to both of us, but...
In any case, I think Ultrix plans to switch to the MIT server soon, so if you
don't get this behavior now, you probably will soon.
Can someone from Ultrix answer with more certainty?
Burns
|
| 815.2 | | FLUME::dike | | Tue May 23 1989 16:42 | 4 |
| The last I heard was that the UWS2.2 servers would be based on the MIT servers
rather than on the common server as has been the case in the past.
Jeff
|
| 815.3 | | ORPHAN::WINALSKI | Paul S. Winalski | Wed May 24 1989 05:04 | 8 |
| Does this mean that UWS2.2 will take a performance hit? The MIT server code
is a reference implementation, and as such is no speed demon. A lot of work
went into the DECwindows Xlib and server code to speed things up. This is
one of the reasons that DECwindows outperforms the competition on X performance:
they're all using the MIT code.
--PSW
|
| 815.4 | Not reference but example implementations | 7486::WITHROW | Robert Withrow | Wed May 24 1989 14:31 | 13 |
| RE: <<< Note 815.3 by ORPHAN::WINALSKI "Paul S. Winalski" >>>
> The MIT server code is a reference implementation,
The documentation clearly states that the MIT provided servers are *not*
reference implementations, but rather example implementations...
> and as such is no speed demon.
This is true, but because the example servers are intended to be very
portable, and emphasis was on making them correct and portable and not
speedy.
|
| 815.5 | Starting from the PMAX base... | 56579::thomas | The Code Warrior | Wed May 24 1989 15:30 | 3 |
| WSL in Palo Alto has done an remarkable job at speeding up the MIT servers.
I would guess that server taken from that base would be very quick...
|
| 815.6 | | 40470::PETTENGILL | mulp | Thu May 25 1989 02:39 | 3 |
| Does this imply that the VMS servers will no longer be constrained by common
Ultrix code and design and can be similarly optimized for VAX/VMS?
|
| 815.7 | That's not the only thing... | 19458::FISHER | Burns Fisher 381-1466, ZKO3-4/W23 | Thu May 25 1989 16:01 | 8 |
| Sharing code with Ultrix is but one of many constraints the server developers have/had.
Another is that as soon as we stray too far from MIT, we run into the problem of
there being thousands of developers for the MIT server making features faster than
we can, and we can't make use of their code. This has already happened. (How
come you &^&^$@ idiots can't do backing store like MIT does????)
Burns
|