On 10/13/14, 9:56 AM, "Erik de Bruin" <e...@ixsoftware.nl> wrote:
>> I don’t have any objection to making a few changes like this to the SDK >>to >> get past this issue. I’m definitely curious to know how much resulting >>JS >> there will be from compiling the SDK. >> > >I’m still working on getting ADVANCED OPTIMISATIONS working with the SDK + >Flash Player + application JS combo. However, with SIMPLE OPTIMISATIONS, >which includes all code and the kitchen sink - it’s basically whitespace >removal, it doesn’t recompile the JS - the output file is a whopping 2 Mb! >Cool, huh ;-) I expect some serious reduction in file size when the GCC is >able to actually parse an entire application and optimise the code >properly >… We’ll see. > > >> That said, what is the plan for when FalconJX compiles some users code >> that has tons of these patterns? Will we just output an error? Might >>be >> nice to have that error/warning in the AS compile. >> > >FalconJX should properly parse this construct. That will be challenging >though, as we’d need to look forward in the AST and then handle these >multiple catch blocks and turn them into a proper set of ‘if … then’ >statement in a single catch block. > >Also, how “heavy” is this problem? Are the catch nodes children of the >> try node? If so, could we just rewrite the AST there? >> > >Actually, hoping to get a blessing for rewriting the SDK, I didn’t look at >the AST at all :-) > >EdB > > > >-- >Ix Multimedia Software > >Jan Luykenstraat 27 >3521 VB Utrecht > >T. 06-51952295 >I. www.ixsoftware.nl
