Re: [Dltk-dev] source path computing

["Jae Gangemi" <jgangemi@xxxxxxxxx>]

  perfect! that's just the type of information i was looking for. 

  i agree w/ both of your suggestions - i had been pondering the mapping issue, and that handles the issue perfectly. i'll start looking into this tonight - if i have further questions, i'll let you know.

  thx!

On 8/1/07, Dmitriy Kovalev <kds@xxxxxxxxx> wrote:

Hello, jae

Here are several my thoughts about remote debugging and working with
remote sources.

I think remote debugging implementation should support two possibilities:
1) Map remote sources to the sources on local machine. In this case it's
required to know the root url on a remote machine and root url on a
local machine. This information can be added to the launch configuration
tab or after starting the debugging session with a popup dialog with
request of source mapping.

2) Retrieve remote sources to the local machine. In this situation it's
also required to know the root url on remote machine and then during
debugging process add new files to the project using dbgp 'souce' command.

Eclipse uses several steps to map active stack frame to the editor:

Step 1. ScriptSourceLookupDirector maps between IStackFrame and any
Object. For now this any object is a simple File obtained from
dbgp url.

Step 2. ScriptDebugModelPresentation maps object returned from
ScriptSourceLookupDirector to IEditorInput, now it's a class that
implements IPathEditorInput.  Signature of this method: IEditorInput
getEditorInput(Object element)

Step 3. ScriptDebugModelPresentation maps IEditorInput to the editor id
(String). Signature of this method: String getEditorId(IEditorInput
input, Object element)

The third step always returns constant string editor identifier. You can
customize Step 1 with Step 2.
For example from Step 1 you can return some object with information
required to create new project file and on step 2
create this file and load sources using dbgp.

If you want to know additional details, please ask.

--
Dmitriy Kovalev


Jae Gangemi wrote:
>
>   could i please get a quick overview of how this works?
>
>   right now, remote debugging is kind of useless unless the file that
> is currently being debugged actually exists on the client machine in
> the same location as it does on the remote machine. what really should
> happen (i think) is if the incoming session is remote, then the source
> should be queried for using the dbgp 'source' command and then once
> the debugging session is finished, the user should be able to edit the
> file (i would actually like to have the option to edit it while the
> debugger is going, but that's for another day) and then save it to a
> location of their choosing.
>
>   i think it would be nice if the remote file could be mapped back
> into the project being debugged itself, but that may prove to be
> difficult if a file of the same name and namespace exists in multiple
> locations (this method is employed by the day job to simulate
> interface type implementations in perl), so for now,  just being able
> to retrieve the remote source would be nice, and it seems that the
> source path computer and/or lookup director is the place for this.
>
> --
> -jae
> --
> This message was scanned by ESVA and is believed to be clean.
> Click here to report this message as spam.
> <http://mail-gw-us.xored.com/cgi-bin/learn-msg.cgi?id=7978C28230.38463>
> ------------------------------------------------------------------------
>
> _______________________________________________
> dltk-dev mailing list
> dltk-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/dltk-dev
>


_______________________________________________
dltk-dev mailing list
dltk-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/dltk-dev

-- 
-jae