Provide machinery for continuing execution after exceptions

[marekjm]

Machine should provide support for continuing execution after exception has been handled.

Example situation:

• an exception is thrown, with its type specifying what has gone wrong,
• handler block is reached (possibly up the stack),
• the environment is restored to the point of exception,
• handler block fixes the environment,
• execution continues,

The simplest solution would involve adding some more instructions:

• restore, for restoring the environment from the point when exception has been thrown (otherwise, the frames from the throw-point up to the catch-point are lost),
• retry, that would leave the block and try to reexecute the instruction that threw the exception,
• continue, that would leave the handler block and continue execution with the instruction just after the one that threw the exception,

NOTE: this is only a first draft of the idea.

[marekjm]

I'm closing this issue. Maybe I'll revisit it later.