I’ve been doing some testing between different environments, and I have a reoccurring problem with error reporting in an on-error clause (this is only related to our Natural for Ajax modules).
Basically, the line number produced from *ERROR-LINE is absolutely fine in our development environment.
However, on migrating to the middle environment, the *ERROR-LINE reported is double what is should be (ie 1320 becomes 2640). … Further down the line (Production), the same error gives a completely different number all together again (and unfortunately at the moment I can’t decipher how it come to it’s conclusion for assigning the error-line… at least the middle deployment was double!).
Is there something causing this? We have looked a few things but can’t find the cause.
Trentan,
Are you able to view the source code in each environment? Is it possible your installation process is leading to a renumbering of the source? Are you sure you have identified the correct object where the error occurred, i.e. are you sure the adapter (*.ns8) isn’t where the error occurred but you only handle it on rolling up to the calling object? Is your development environment NaturalONE?
If NaturalONE is involved, at one point there was a problem between NaturalONE and the source on the development server side where the header would get out of sync with the line increment value on the server side. This prevented CTRL+G from finding the correct line. The NaturalONE source would show a line increment of 10 for example, but within the SAG editor in Telnet, the line increment would be 5. This bug led to error-lines being off by a factor. SAG patched this problem and I don’t recall it happening recently. If using NaturalONE and it is not up-to-date, you may want to check the LineIncrement value in the header.
* >Natural Source Header 000000
* :Mode S
* :CP
* :LineIncrement 10
* <Natural Source Header
This probably isn’t much help but hopefully it triggers some ideas. Good luck.
Todd
Todd,
Thanks for the response,
Yes my dev environment is Natural One.
The only difference from our past work is the line increments by one (N1 standard).
I have tested multiple errors in different locations / programs / sub programs and they all return the same result (doubling the line number …)
We have analysed the source and it is the same in each separate environment, (always -4 when looking in Natural One for the Natural Source Header).
In our code, there is no line increment definition in the header eg.
* >Natural Source Header 000000
* :Mode S
* :CP
* <Natural Source Header
I upgraded to 9.5 to see if that would rectify it, but it hasn’t. On another note, the outline takes ages to load now in 9.5 (ie font colours not visible until outline is loaded… and no ability to save without freezing up eclipse until the colours are activated… not too sure why that is)