In article <3b7d8b76$1_2@dnews> "Kristofer Skaug" <✉skaug.demon.nl> wrote:
> what I was now envisioning was a sequence like this:
>
> try
> {1}
> try
> {2}
> try
> {3}
> try
> {4}
> finally
> <rollback 4>
> end; {4}
> finally
> <rollback 3>
> end; {3}
> finally
> <rollback 2>
> end; {2}
> finally
> <rollback 1>
> end; {1}
>
> I'm pretty sure any tight processing you want to do around the core part
> {4} is going to take a hit from all the extra try-finally infrastructure.
And how are you going to perform this differently in the absence of exceptions; that is, using a motley collection error codes and ragbag global variables, scattered hither and thither, every second statement a check for error conditions?
> Especially if you now extend this to one try/finally pair per state > parameter you need to roll back, and there may be 10 or more of those...
You speak of combinatorial explosion. There are ways to fold these.
-- Barry