Skip to main content


Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Language IDEs » ServerTools (WTP) » Breakpoints in JSP pages
Breakpoints in JSP pages [message #224199] Mon, 24 November 2008 20:03 Go to next message
Steven Curtis is currently offline Steven CurtisFriend
Messages: 6
Registered: July 2009
Junior Member
I have demonstrated to myself that JSP breakpoints are honored for JSP
pages located only within the root web content directory regardless of how
the debug profile's source lookup path is configured.

<wb-resource source-path>/index.jsp [line: 13] <- BREAKS SUCCESSFULLY
<wb-resource source-path>/jsp/index.jsp [line: 13] <- DOES NOT BREAK

I have also verified that the issue is not the responsibility of the OC4J
container as I have successfully triggered both breakpoints from within
JSwat pointing to the two JSP source directories shown above while
attaching remotely to the Java process.

It is also of note that a servlet called by both JSP pages successfully
breaks. In the case where Eclipse does not break on jsp/index.jsp line:
13, the Eclipse debugger stack trace shows "index.jsp line: 13" on the
stack just after the successful break in servlet code. The "jsp"
directory containing "index.jsp" has been manually added to the source
lookup path, and "Search for duplicate source files on the path" has been
selected to make sure there is no ambiguity between JSP pages.

Can my observations be verified by anyone else? Is there any advice on
how I can address this problem?

OC4J Extended 10.1.3.4.0
Eclipse Platform
Version: 3.4.1
Build id: M20080911-1700
WST 3.0.2
Re: Breakpoints in JSP pages [message #224217 is a reply to message #224199] Tue, 25 November 2008 00:43 Go to previous messageGo to next message
Steven Curtis is currently offline Steven CurtisFriend
Messages: 6
Registered: July 2009
Junior Member
Unfortunately, I have uncovered some confusing evidence. The scenario
previously described works in Eclipse if deployed to a JBoss server.
Since Eclipse correctly debugs the application deployed to a JBoss
container and the Swat debugger correctly debugs the application deployed
to an OC4J conainter, the problem has to be in the interaction between the
Eclipse debugger and an OC4J container and not explicitly in either one.
Any suggestions on how to investigate the interaction between the two
since each have shown proper behavior independently?

Steve
Re: Breakpoints in JSP pages [message #224225 is a reply to message #224217] Tue, 25 November 2008 04:01 Go to previous messageGo to next message
Eclipse UserFriend
Originally posted by: david_williams.linux.vnet.ibm.com

Steven Curtis wrote:
> Unfortunately, I have uncovered some confusing evidence. The scenario
> previously described works in Eclipse if deployed to a JBoss server.
> Since Eclipse correctly debugs the application deployed to a JBoss
> container and the Swat debugger correctly debugs the application
> deployed to an OC4J conainter, the problem has to be in the
> interaction between the Eclipse debugger and an OC4J container and not
> explicitly in either one. Any suggestions on how to investigate the
> interaction between the two since each have shown proper behavior
> independently?
>
> Steve
>
I don't have much concrete ... but first, you are intentionally studying
the problem with files that have the same name, right? There's some
(oldish) bugzilla entries that talk about problems with files that have
the same name (but in different directories). You might search bugzilla
to see if you can find them, and if so, if they might help explain the
issues. [I don't know if they'd be open or closed.] My guess it that it
is related to how the jsp name-mangling is done ... but, that is just a
guess.
Re: Breakpoints in JSP pages [message #224232 is a reply to message #224225] Tue, 25 November 2008 13:35 Go to previous messageGo to next message
Nick Sandonato is currently offline Nick SandonatoFriend
Messages: 126
Registered: July 2009
Senior Member
David Williams wrote:
> Steven Curtis wrote:
>> Unfortunately, I have uncovered some confusing evidence. The scenario
>> previously described works in Eclipse if deployed to a JBoss server.
>> Since Eclipse correctly debugs the application deployed to a JBoss
>> container and the Swat debugger correctly debugs the application
>> deployed to an OC4J conainter, the problem has to be in the
>> interaction between the Eclipse debugger and an OC4J container and not
>> explicitly in either one. Any suggestions on how to investigate the
>> interaction between the two since each have shown proper behavior
>> independently?
>>
>> Steve
>>
> I don't have much concrete ... but first, you are intentionally studying
> the problem with files that have the same name, right? There's some
> (oldish) bugzilla entries that talk about problems with files that have
> the same name (but in different directories). You might search bugzilla
> to see if you can find them, and if so, if they might help explain the
> issues. [I don't know if they'd be open or closed.] My guess it that it
> is related to how the jsp name-mangling is done ... but, that is just a
> guess.
>
>
>
Here's the (oldish) bug:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=99674

Hope this helps to clear things up a little.
Re: Breakpoints in JSP pages [message #224240 is a reply to message #224232] Tue, 25 November 2008 14:36 Go to previous messageGo to next message
Steven Curtis is currently offline Steven CurtisFriend
Messages: 6
Registered: July 2009
Junior Member
Nick Sandonato wrote:

[snip]

>> I don't have much concrete ... but first, you are intentionally studying
>> the problem with files that have the same name, right? There's some
>> (oldish) bugzilla entries that talk about problems with files that have
>> the same name (but in different directories). You might search bugzilla
>> to see if you can find them, and if so, if they might help explain the
>> issues. [I don't know if they'd be open or closed.] My guess it that it
>> is related to how the jsp name-mangling is done ... but, that is just a
>> guess.
>>
>>
>>
> Here's the (oldish) bug:
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=99674

> Hope this helps to clear things up a little.

Thanks for the info, but I guessed that might be the problem as well.
I've been able to reproduce the issue using two JSP files with different
file names. To your point, when two files with the same file name are in
different source paths in an OC4J debug session, current Eclipse behavior
is to prompt the user to resolve the ambiguity with a dialog box asking
for which source file to display. Unfortunately, in the jsp/index.jsp
case, the dialog doesn't show because the breakpoint isn't being triggered.

Another curiosity is the fact that the JBoss debug session has absolutely
no problem distinguishing between two files in different directories with
the same name while the OC4J debug session prompts to resolve ambiguities
and will occasionally display the wrong file. Is this due to an
incomplete or buggy JSR 45 implementation in OC4J? Incorrect "jsp
name-mangling" (Not sure what that is. Have any more details?) by OC4J?
I've read about line number maps that are supposed to be passed back to
the IDE. Maybe the maps coming from OC4J are malformed or incomplete?
Still researching. Suggestions appreciated.

Thanks,
Steve
Re: Breakpoints in JSP pages [message #224469 is a reply to message #224240] Wed, 26 November 2008 16:29 Go to previous message
Nitin Dahyabhai is currently offline Nitin DahyabhaiFriend
Messages: 4425
Registered: July 2009
Senior Member

Steven Curtis wrote:
> Another curiosity is the fact that the JBoss debug session has
> absolutely no problem distinguishing between two files in different
> directories with the same name while the OC4J debug session prompts to
> resolve ambiguities and will occasionally display the wrong file. Is
> this due to an incomplete or buggy JSR 45 implementation in OC4J?
> Incorrect "jsp name-mangling" (Not sure what that is. Have any more
> details?) by OC4J? I've read about line number maps that are supposed
> to be passed back to the IDE. Maybe the maps coming from OC4J are
> malformed or incomplete? Still researching. Suggestions appreciated.

Generally that's the case. I don't believe the JSR is specific
enough in this area.

https://bugs.eclipse.org/bugs/show_bug.cgi?id=99674#c8

--
---
Nitin Dahyabhai
Eclipse WTP Source Editing
IBM Rational


_
Nitin Dahyabhai
Eclipse Web Tools Platform
Previous Topic:Programmatic use of XMLCorePlugin and ICatalog APIs
Next Topic:Task Tags in JSP files are not listed in Tasks tab
Goto Forum:
  


Current Time: Tue Mar 19 03:25:17 GMT 2024

Powered by FUDForum. Page generated in 0.02700 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software

Back to the top