Bug #4852
Updated by Greg Shah almost 4 years ago
We have a situation where ErrorManager.recordThrowError does not set the Progress error code on the ErrorConditionException. The cause seems to be the use of a different constructor (error, code) instead of (code, error), this one is setting the 'key' but leave the progressErrorCode unset (-1).
Is that something done in purpose? How can we get along this if changing the code to use the right constructor is not an option? What we try to do is to do an assignment to validate the input date and if that fails add additional error messages so I'm catching that ErrorConditionException and then use recordThrowError with an array of error number/messages.
This is at line 1438 in ErrorManager, if we can switch the order of parameters @errmsg@ and @num@ then our issue is gone :)
Argh, it looks that way the error number is appended twice at the end, a solution is not to clear pending errors but only throw the new ones we need so not an issue anymore, still this might be 'fixed' if not by-design.
Is that something done in purpose? How can we get along this if changing the code to use the right constructor is not an option? What we try to do is to do an assignment to validate the input date and if that fails add additional error messages so I'm catching that ErrorConditionException and then use recordThrowError with an array of error number/messages.
This is at line 1438 in ErrorManager, if we can switch the order of parameters @errmsg@ and @num@ then our issue is gone :)
Argh, it looks that way the error number is appended twice at the end, a solution is not to clear pending errors but only throw the new ones we need so not an issue anymore, still this might be 'fixed' if not by-design.