1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
|
The clusterpush.pl script allows you to run a cluster regression using
code that has not been committed to svn. It's supposed to be a
replacement for clusterpush.sh.
Usage:
clusterpush.pl [type] ['bmpcmp'] ['lowres'|'hires'] [user]
Where type gs, pcl, xps, and/or ls (standing for language switch).
If you do not specify a type clusterpush.pl will guess what you want to
do based on the current directory.
bmpcmp indicates that a bitmap comparison of your output to that of the
last svn commit should be performed.
lores or hires indicates to only perform a low-resolution or
high-resolution cluster run; by default both are performed.
user is your cluster user name, if you don't specify one, clusterpush.pl
uses, in order of preference, the value of one of the following environment
variables: CLUSTER_USER, USER or USERNAME (USER and USERNAME being the normal
environment variables in Unix type shells). This makes sense on Linux and Mac
OS X and probably cygwin, but I'm not sure what happens when run under
Windows.
The order of the options is not important.
You need to run clusterpush.pl from the top level directory (the same
directory from where you normally run 'make'). Clusterpush prints and
error if you try to run it from a different directory.
Results will be sent via email after the run is completed, which takes
about 25 minuts, presuming the cluster isn't busy. Your code is compared
to the most recent baseline, so for meaningful results you will want
to sync your code with head if recent changes to head have resulted in
regression differences.
The first time you run clusterpush.pl it will take a really long time
to rsync your source to the master machine and then from the master to
the nodes. After that things will go much quicker (unless, like me,
you accidently leave giant .ppm files in the current directory).
bmpcmp mode:
Starting with r10876 the local cluster can now generate bitmaps that
show differences between clusterpush jobs and trunk.
To use the feature run clusterpush.pl with the option bmpcmp (i.e.
"toolbin/localcluster/clusterpush.pl bmpcmp"). This will queue a job
that looks at the most recent none-bmpcmp clusterpush job you ran and
generates bitmaps of the differences for the first 1000 files listed
in the report. It puts these files on the internets in the location
<http://www.ghostscript.com/~regression/$USER> and then sends you
an email telling you that it's done. Since some of the test files
are customer confidential this directory is password protected, I've
emailed out passwords, contact me if you have misplaced your password.
BTW, I consider the fact that can look at output from other users, e.g.
<http://www.ghostscript.com/~regression/marcos>, a feature.
Starting with r12230 you can add bmpcmp to a normal clusterpush
(i.e. 'clusterpush gs bmpcmp' or 'clusterpush gs pcl xps ls bmpcmp').
This will queueu both jobs in the correct sequence.
|