In the course of my beta-testing I've noticed that Doors CS's custom error handler seems to occasionally make mistakes when taking you to a specific line in the editor. It doesn't seem to be off by too much generally, but it's still annoying, and I'd like you all to present me with some cases where it works correct, and some where it works incorrectly, in the hopes that I can find a pattern and solve it. Please write the incorrect line of code and the line that Goto goes to at least; full code if you can.

I should note that this is the last issue (at this point) to be resolved before Doors CS 6.9 beta, the very last beta, is released, so the faster this is solved, the faster 6.9 beta and DCS 7 RC 1 can arrive!

I'd like to also point this page out: http://dcs.cemetech.net/index.php?title=Error_Codes

Also, I'll get right on this

Ah, thank you for that, much obliged. Hopefully that page will be slightly less necessary, now that I added the three-character error codes.

but it keeps a list that we can look off of to decide what is left

Edit:
:sub(str1,0,0 :111 - stops at the Sub( token (this assumes str1 does not exist)
:sub(str1,0,0 :111 - stops on the second line (Str1 exists)

Btw, should we give the 3 letter error code? or maybe the number of the error? Or just like how I've been saying it?

The three-line error code would help. Just to clarify, are you saying that you tried with Str1 existing and not existing, with this code?


Code:
?

So here's an error with expr(, as you said in the chat, there's nothing you can do about it, but you get a post anyways.
prgmA:
Code:
prgmB
Code:
It'll take you to the first line in A.

Thanks. Note to future Christopher: the reason I can't do anything about this is that the TI-OS only tells me the address of the temporary string it copied "^" into to try to evaluate it, and nothing about the program it was actually in, let alone what line it was parsing at.

A lot of thinking and testing has let me to the (hopefully correct) realization that the problem occurs when the BASIC error is traceable to the last token in a particular line, in which case it ends up on the next line. I'm looking through my code now to see if I can find any evidence to support this hypothesis.