Digitally signed JWS client - invoking email program on the client

Question

Hi All,
We have a digitally signed JWS application (non-sandboxed), I would like
to programatically invoke the email client outlook etc, with attached log file ( of our app)
basically an error handling mechanism such that any error on the client side can
be sent to the dev team by simply clicking a button say "Send Log File" on the application
Which should utimately opens end user's email client with attached log file, and users sends it to us from the email
Appreciate any pointers/tutorials in this regard
Many thanks,
Edited by: janaAb on Dec 11, 2007 6:20 AM

Answer 1

janaAb wrote: ...
Many thanks for the reply, Our applications are targetted 1.5 and before appreciate if you could clarify the following
appreciate if you could clarify
- forms an output stream direct back to a report page on the server, and prepares to write the entire report to that stream - logs, screenshots, system settings, ..the time, + you might prompt the user for a 'reason?', whatever.. back to the reporting page. - after writing the stream to the server, the server returns an URL in the output. by the time user has encountered the error it means, the exception logs are already written, now is this you have this in the memory and on user action write to the servlet outputstream ? , bit confused here is it whether normal IO operation to read the physical log files and write to servlet output stream ? ( i believe is the case on a successful operation, the log file just writes to disk )
In the back of my mind, I was thinking of the 'sandboxed' version I am developing, but strictly speaking, no.
The way I am using logging is this.
I create a custom handler that stores the last 500 (configurable) logged messages, then add that to an anonymous logger, which is used for all logging operations.
The logs might be written out somewhere, but my app. no longer cares either way. If a problem occurs, they (without any further elevation of privileges) have easy access to the last 500 logged messages.
Now, I'll break from what my sandboxed app. does here (too long, not relevant) and cut to..
If that were the case in your app., it would take no further reading of files on the local disk, in order to submit the report back to the servlet. We could form the entire log as a String within the app., then stamp it out to a ByteArrayOutputStream, and take it from there.
That does not make much sense for an all-permissions app. such as yours. It is more efficient to allow the logs to be written to disk as per normal, and only open (& submit) them if needed. Using a 'cycling' log name, you might also even offer the ability to submit the logs of earlier run(s). Remembering the logs of earlier runs is not a viable option for my sandboxed app.
janaAb wrote: in .NET world these kind of things seems pretty slick, ..
So (..shrugs) use .NET? Is this application only run on Windows? Does .NET support every platform you want to deploy on?
If this app. is only targeted at/supported on platforms for which .NET does the job, perhaps it should be considered as an alternative. I do not think it is necessary though (and I doubt it would work on my Windows machine).