Well, i've started to filter the outcome of grep -R "8080" * in both
the tempo git clone and the generated tomcat bundle. After getting rid
of logs, tomcat examples and yui's color codes, i've found:
In the tempo clone:
1) Under "test" folder, a lot of xml config files under MODULE/src |
target/resources/* with references to localhost:8080, which i think is
ok. But, there are some java files wired with the 8080 reference, in
the ui-fw module, that would be a problem if someone needs to run the
tests in other port.
2) Under "config" xml files for each module (maybe from here the test
port should be picked?)
3) Liferay, jetty and tomcat config files. I suppose that if i change
some here, the tempo_builder would configure the container for the
generated bundle accordingly. Am i right?
4) In the tempo_builder_lib.rb script, a couple of methods:
set_tomcat_ports ({"8005"=>"9005","8080"=>"9080", "8443"=>"9443",
"8009"=> "9009"})
and
replace_all_with_map_in_folder({"localhost:8080"=>"
bpms.tempo.com:
8080", "
127.0.0.1:8080"=>"
bpms.tempo.com:8080"}, "#{@@server_folder}/
var/config")
What this calls do?
5) Finally, the following src/main/*.java references in the code:
fds/src/main/java/org/intalio/tempo/workflow/fds/
FormDispatcherConfiguration.java: private String _pxeBaseUrl =
"
http://localhost:8080/ode";
fds/src/main/java/org/intalio/tempo/workflow/fds/
FormDispatcherConfiguration.java: private String _fdsUrl = "http://
localhost:8080/fds/workflow/ib4p";
fds/src/main/java/org/intalio/tempo/workflow/fds/
FormDispatcherConfiguration.java: private String _tmsUrl = "http://
localhost:8080/axis2/services/TaskManagementServices";
tas-service/src/main/java/org/intalio/tempo/workflow/tas/core/
WDSStorageStrategy.java: * WDS endpoint (including the trailing
slash), such as <code>
http://localhost:8080/wds/</code>
tas-service/src/main/java/org/intalio/tempo/workflow/tas/sling/
SlingStorageStrategy.java: private String slingUrl = "http://
localhost:8080/sling";
ui-fw/src/main/java/org/intalio/tempo/uiframework/
Configuration.java: private String _tmpEndpoint = "
http://localhost:
8080/ode/processes/completeTask";
This seems important for the port changing issue.
In the Generated tomcat Bundle:
1) The binary coincidences of the .class corresponding to 5) above.
2) tomcat config files.
3) Tempo config files, guess the same ones in 2) above.
4) The Deployed WSDL and BPEL files has references to localhost:8080,
You put this reference in the designer? or is configured somewhere in
tempo?
5) One jsp file wired.
webapps/ode/fileupload.jsp: opts.setTo(new
EndpointReference("
http://localhost:8080/ode/processes/
DeploymentService"));
So, summarizing:
1) Where should be the port change? providing a script at the bundle
seems the best option, but the references in the .java files doesnt
affect?. the other option would be a switch in the tempo_builder.rb to
make a build with the selected port.
2) What can be done with the wired test cases?
I want to contribute, the codebase here is very advanced, and the SOA-
orientation makes it cooler. The problem is that the documentation
sums up to the install.txt file and your answers, which is somewhat
short to understand how to use and how to extend.
Because of that, i'm thinking about contributing with docs, and
modifying the rsc/scripts to make them more flexible. Lets see if my
boss agrees, he wants some workflows automated, not a bunch of ruby
scripts.
I'm more a python-svn guy than ruby-git, but i can give a try.
Sorry for the long statement, but i'm very interested in this.
On 25 nov, 21:07, Nicolas Modrzyk <
hellon...@gmail.com> wrote:
> Quick answer. Yes, the port change is the cause.
>
> A grep/replace script to replace all the ports is your best bet. There are a
> few configuration files you need to do that. (var/config) also deployed
> files have urls to change.
>
> if you have a nice ruby script you want to contribute, we are interested. :)
>
> For the record, I just checked tomcat on a regular port and it worked no
> problems.
>
> Nicolas,
>
> >
tempo-dev+...@googlegroups.com<
tempo-dev%2Bunsu...@googlegroups.com>
> > .