Re: SQLRPGLE preprocessed code mixing with expanded copy book

On 15/10/2008, at 7:09 AM, Tyler, Matt wrote:

FYI, Here's IBM official response.

The developer had the following update...

"I looked into this more on Friday and it is not a duplicate of
SE32782. This problem has been around forever. There are two
simple/easy circumventions. One they figured out is to have a blank
line. This line can even be in the file that is being brought in by
the copy. The second is to compile with RPGPPOPT(*LVL1 or *LVL2).
Since the problem is easily to see and there are two easy
circumventions I do not plan to fix this."

That's such crap! It is obvious that the pre-compiler function is broken because it should not be inserting code inside conditional compilation unless directed to do so by the programmer.

The pre-compiler is obviously processing the /COPY statement, including the code in the positive test, inserting the /ELSE, including the code in the negative test, and then screwing up completely by inserting the SQLCA before the closing /ENDIF.

Since a blank line "fixes" things the pre-compiler must be inserting the SQLCA before the previous line rather than after it. Duh! Why would anyone think that is acceptable?

Given there are two work-arounds it is reasonable to say no PTF but it should at least be fixed in a future release--preferably the next one although given the newness of VRM610 they should PTF it in that.

"I do not plan to fix this" implies, broken, can stay broken, I don't care. It's only a little bit broken so that's OK? Pah! I despair sometimes.

Perhaps the developer is just trying it on. If I were you I'd reject the answer explaining why it's unacceptable and what you would accept such as a PTF in 610 but work-around for previous releases.

Regards,
Simon Coulter.
--------------------------------------------------------------------
FlyByNight Software OS/400, i5/OS Technical Specialists

http://www.flybynight.com.au/
Phone: +61 2 6657 8251 Mobile: +61 0411 091 400 /"\
Fax: +61 2 6657 8251 \ /
X
ASCII Ribbon campaign against HTML E-Mail / \
--------------------------------------------------------------------