hot-deploy and source scanner finetune`

[GoogleCodeExporter]

Actually we fully support JDT hot-deploy feature when user run in debug mode, 
but we also enable source scanner by default , 
when user try to update java source , they found the server restart.

In general development , it's not a must have for enabling source scanner ,
the reason we enable it by default is just for someone might want to scan the 
config files.

That might make some user confuse about the hot-deploy feature not working in 
RJR,
I found a lot of ppl not using hot-deploy feature because source scanner.

We didn't have a way to manage the watch list in past, 
but now in Issue 88 , 1.3.2 will provide a tree view to manage the watch list.

Now it's time to make this one more powerful.

New spec changed:

1.If user run in Debug Mode , we will ignore any source change for ".class" 
files , we give the job to JDT (Hot-deploy) , if hot-deploy failed , you have 
to restart by yourself.

2.We will NOT scan target output folder by default, 
but only some known config files will be in the list, 
like web.xml,hibernate.cfg.xml ...etc

  Sure , if user want , they could still enable it in the watch list manager.

We are trying to find a useful default settings. :)

Original issue reported on code.google.com by tonylovejava on 13 Dec 2011 at 2:32

[GoogleCodeExporter]

Original comment by tonylovejava on 13 Dec 2011 at 2:37

• Added labels: 1.3.2

[GoogleCodeExporter]

I changed my mind for point 2, it's fine to keep target output folder in the 
scan list.
Almost every static file in the target output folder require a restart.

I was thinking for web application resources , but they are not actually in the 
case.

Original comment by tonylovejava on 13 Dec 2011 at 2:39

[GoogleCodeExporter]

Issue 101 has been merged into this issue.

Original comment by tonylovejava on 14 Dec 2011 at 5:16

[GoogleCodeExporter]

One more details after implementation.

Currently when we create a empty folder or remove a empty folder,
the web application will restart.

But it's actually not necessary, we will only scan "file" change in 1.3.2 .

Original comment by tonylovejava on 19 Dec 2011 at 1:04

[GoogleCodeExporter]

A quick update for point 2 , to simple the rules and make it more useful,
I am just going to scan all the xml file under WEB-INF by default.

I believe that could cover most use case so far.

Sure , user could still disable scanner by uncheck those items.

Original comment by tonylovejava on 19 Dec 2011 at 4:41

[GoogleCodeExporter]

Ready since 2011/12/19 , 
will be in 1.3.2 .

Original comment by tonylovejava on 19 Dec 2011 at 5:02

• Changed state: Ready