I've found out that we have a major performance regression in remote X
performance starting with build 58.

Here are the results of a remote run of SiwngMarks:
java version "1.4.0-internal"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.0-internal-b57)
Java HotSpot(TM) Client VM (build 1.4.0-beta-b57, mixed mode)
-------- RUN --------
Starting SwingMark
SwingMark Test started at Thu Apr 19 16:01:00 PDT 2001
Program will automatically terminate after last run
Startup Time: 9434
Sub-Menus = 3276 (Paint = 0)
TextArea = 5838 (Paint = 149)
Sliders = 6478 (Paint = 501)
Lists = 4687 (Paint = 495)
Table Rows = 10553 (Paint = 176)
Tree = 44143 (Paint = 928)
Score: 85969
-------- DONE --------

And here is build 61:

java version "1.4.0-beta"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.0-beta-b61)
Java HotSpot(TM) Client VM (build 1.4.0-beta-b61, mixed mode)
-------- RUN --------
Starting SwingMark
SwingMark Test started at Thu Apr 19 15:49:35 PDT 2001
Program will automatically terminate after last run
Startup Time: 8475
Sub-Menus = 3084 (Paint = 0)
TextArea = 8860 (Paint = 466)
Sliders = 6422 (Paint = 501)
Lists = 4653 (Paint = 494)
Table Rows = 72138 (Paint = 176)
Tree = 41887 (Paint = 928)
Score: 147513
-------- DONE --------

Note the results for the Table Row test.

The bug is reproducible on solaris and linux.

dmitri.trembovetski@eng 2001-04-19