What you need can be done in many ways. One easy way is to use a regular expression to match that. The regular expression looks like this /:\d+]/ . When applied to a single line from the error stack will match this :lineNumber]
So for example ( if err is Error ) this will output the line number:

var lineNumber : uint;
var loc : String = err.getStackTrace().split(”at “)[1].split(”()”)[1];
var match : Array = loc.match(/:\d+]/gi);
if(match && match.length > 0)
{
lineNumber = parseInt(String(match[0]).substr(1,match[0].length-1));
trace( “lineNumber:”+ lineNumber);
}

But Great work.

Thanks

Breaking up View and Controller is harder, and seems the source of much of the over-engineering in MVC.

For me, keeping the model separate, taking advantage of the fact that business rules exist independent from most of your technical decisions even though they impact your technical decisions, is the big win.

Thanks,
George

Thanks

John

- 11 web pages. To be conservative I’ll count user controls as pages as well (most plug into a wizard page) so that bumps the count up to 21 actual web pages.
- 8 solutions
- 50 projects/layers
- 852 supporting *.cs files (interfaces, views, business objects, DOA objects) not counting the designer.cs or aspx.cs code-behind files
- 56 Nhibernate mapping files

So that’s roughly 40 supporting *.cs files per web page for a fairly simple website. Frankly that’s just dumb. And whenever requirements change (let’s face it, they do) the architect/lead developer spends such a long time trying to work out whether the existing bloat can accomodate the marginal changes without breaking too many things.

There are alternative design patterns to MVP that are faster to write, more flexible to change, more maintainable, more readable/comprehensible, and which abstract code and seperate concerns to a reasonable level.

I really hope the old school Java programmers stuck thinking in this pattern retire and stop propagating this stupid architectural pattern.

No, not at all. The thing is that I created this to suit my needs and will continue to update it as soon as my needs expand or somebody who uses it needs something added/fixed and I find it right to do so. Haven’t got the chance but I will check out what else is there and most importantly to see if the results are similar. Who knows, maybe they can go together or even joined… I’m always opened for suggestions.

yeah, me too. the official word I got was that the AIR framework would provide enough low level stuff for the community to then take and run with (which you have - many thanks for the hard work)

the interesting thing is that window managers have been provided by Windows all the way back (as far a I can remember) VisualBasic3.0.

does it worry you that there might be a duplication of effort in different implementations? I mean, for the rest of us we just need a window manager that is predictable to implement and simply works.

i’ve been looking for something like that. the fact, that the text is real (selectable) text makes the whole thing even more cooler! would be nice if you share the source code - i’d like to play arround with it …. getting lot of ideas when i see this!

keep on with you new blog!

best regards

jan v.
halle (s.)
germany

Scalability: In another project, you could easily add some more buttons and change the Model to emit relative time (eg: to create a stopwatch). Or in another, you could create three world clock times in your View that are in sync simply by adding two simple lines of code.

To each his own, I guess. I think using MVC becomes self-fulfilling after a while. After a year of using it solidly in almost every major project, I’ve found it harder to go back to prototyping in the Flash IDE the way I used to.