johnf wrote:
> On Thursday 20 November 2008 08:37:28 am Paul McNett wrote:
>> Just curious: why do you want your users seeing the debug output?
> 
> I don't but in this case the user is the project manager.  He wants to see 
> the 
> command prompt because he want to see the errors.  He also has to report the 
> errors and  traceback or messages back to me.  At the moment he has been 
> reporting the error about the dataset.  He wanted to know why it has not been 
> fixed.  But of course it's not broken.  I have been considering trying to 
> find a work around.

What I think you should do for your distributed app is override 
sys.__excepthook__ to 
catch *all* unhandled errors and show a dialog with some information and then 
close 
the app. My app defines a HandleErrors property which defaults to True if the 
app is 
frozen (py2exe or py2app), and False if not (if I'm testing). That way the user 
sees 
the error dialog, and I see the traceback.

My error handler doesn't display anything technical, just a "We are sorry", and 
a 
place for the user to enter notes about what they were doing. It then connects 
to a 
webservice to send the technical information to my master error database, and 
closes 
the app.

Not allowing the app to continue when an unhandled error happens is important 
IMO. 
While it does frustrate users, the problem *must* be fixed by me ASAP and 
because it 
was unhandled, it is fairly risky to let the app continue on because now the 
state is 
undefined.

For the info and error logs, I reroute those to the user's AppData directory 
for my 
app, to files 'error.log' and 'info.log'. These logs reset with each app start. 
So a 
knowledgeable user could review them if they wanted.

Paul


_______________________________________________
Post Messages to: [email protected]
Subscription Maintenance: http://leafe.com/mailman/listinfo/dabo-users
Searchable Archives: http://leafe.com/archives/search/dabo-users
This message: http://leafe.com/archives/byMID/[EMAIL PROTECTED]

Reply via email to