RE: Recursive Function
- From: Carel-Jan Engel <cjpengel.dbalert@xxxxxxxxx>
- To: oracle-l@xxxxxxxxxxxxx
- Date: Sun, 09 May 2004 22:05:58 +0200
At 09:48 PM 5/9/2004, you wrote:
>Inline comments >>
>
> >> I'm older; I started with machine-language. Makes me a trochlodyte?
Probably, I'm not that good in history/archeology/paleontology! apologies
> >> To check means more CPU time, so more wall-time, so less performance??
>But those days switching of boundary checks in the (USCD-P) compiler caused
>a significant gain in performance.
Boundary checking (this was Pascal), means that the compiler had to
generate additional code to test at runtime that copy operations in arrays
didn't try to write outside of the arrys. (Something that 'C' has 'switched
off' by default ;-) ). So, after running our (extensive) tests, we switched
this of for production. We went even further: declaring a set of parameters
in a certain order meant that they were stored on the stack in that order.
Disabling type-checking, and passing the address of the first parameter to
a procedure or function that had the same set of parameters declared (in
the same order!) saved pushing extra pointers on the stack. Just the first
address was passed, and the rest was derived automatically. Saved some
additional time at runtime. We did this at extremely low-level functions.
Forgive me that I do not recall the exact syntax.
> >>>> How would you pass these values 'back'?? Some global variables (in a
>pkg?, something like this?)
No, call by reference means that the parameter used in the function
actually uses the same address space as the variable that was passed to the
function at calling time. So the value is 'back' already when the
procedure/function ends. But, from the perspective of the procedure the
variables are within the scope of the procedure.
> >>>> If you don't have an error, what error-codes would you put back. And in
>reverse: if you got an error, what function value would you give back?
Like in 'C': 0 is OK, every non-zero value is a predefined error situation.
>What I particularly like to do is passing results (error-codes) back
>through the function, whilst the result of the function itself is passed
>through the variables.
>
> >>That may be old fashioned. But it works flawlessly. The first time I wrote
>some Oracle code I did it like this and I heard that I was old fashioned. I
>didn't give a ...
Amen!
>Of course raising exceptions is the mechanism for
>that now, so my way of working might be somewhat old-fashioned. Unskyld.
Regards, Carel-Jan
===
If you think education is expensive, try ignorance. (Derek Bok)
===
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
Other related posts:
