RE: [gmt-dev] Error Transformations

["Willink, Ed" <Ed.Willink@xxxxxxxxxxxxxxx>]

hi All

> b) what to do with the Error element.
...
> Is there an implicit behaviour that supplies an error output for
> every transform without an error output?
> 
> If there is such an implicit behaviour, should it be available for
> other categories of output as well? e.g. a CodeCoverage output
> or a Debug output.

Yes. An implicit communication is of course an exception. So we
just need an ability to

throw an exception
	- possibly by connecting it to the optional exception output.

catch an exception
	- possibly by connecting from the optional exception output

exception creation works by normal element evolution/preservation

exception match/try by normal type pattern matching

exception rethrowing works by just throwing the catch

default propagation works by an implicit rethrow of all exceptions
that cannot match any explicit handling 

So we need:

An ability to define an exception output.
	- maybe a special output port icon
		- avoids the problem of naming the port
	- maybe a boolean attribute on a standard port output
		- avoids an extra icon
	- maybe an <<exception>> stereotype on a standard port output
		- avoids an extra icon

An ability to define that a connection is from an exception outpyt
	- maybe use the reserved <<exception>> port name
	- maybe a boolean attribute/stereotype on the connection
		- cannot be done in GME
	- maybe an <<exception>> stereotype on the caught element
		- avoids a reserved name

The rest is 'just' compiler semantics.

I think I prefer a special output port with the reserved name
<<exception>>.

Comments?

	Regards
			
		Ed Willink

------------------------------------------------------------------------
E.D.Willink,                             Email: mailto:EdWillink@xxxxxxx
Thales Research and Technology (UK) Ltd, Tel:  +44 118 923 8278 (direct)
Worton Drive,                            or  +44 118 986 8601 (ext 8278)
Worton Grange Business Park,             Fax:  +44 118 923 8399
Reading,   RG2 0SB
ENGLAND          http://www.computing.surrey.ac.uk/personal/pg/E.Willink
------------------------------------------------------------------------
(formerly Racal Research and Thomson-CSF)

As the originator, I grant the addressee permission to divulge
this email and any files transmitted with it to relevant third parties.