My apologies. I have used ManuScript so long that I tend to see an error message and think: “this is the last line where it know what it was doing” rather than “this is where the error is”, so I really tend to think that the errors were being reported correctly.
The highlighted line is never the line that contains the error. Almost always the error is in the line statement, which may cover more than 1 physical line.
This could be one reason why I tend to use multiple short lines rather than long single lines…
If you use /* */ comments in your method the location is further confused, so I never do that. I have never gone to the trouble of figuring out if those lines are just not counted but if you have several in a method, then at beat you may have to count up all the comment lines, and when you are already trying to figure out an error, it is just not worth the convenience of block commenting to me.
I have code in my text editor that will add or remove // from the start of selected lines but I only bother with that if there is a really big block.
So my experience does match yours, and I don’t think about it at all, so that would be confusing to a new user.
I also use
str = “” & arr[i]; and
num = 0 + global_var;
without thinking about it, and this used to catch me up all the time.
Dictionaries and sparse arrays do not require you to force assignments to return a value, so I now use them wherever I can (plus dictionaries are absolutely spectacular data structures when you want to sort an array, or do fast lookup, for example) but dialogs can only use old-style arrays and global variables, so there is really no avoiding “” &
Thomas gives a lot of good information. One point, however:
One important thing to be aware of: Sibelius often (or always??) reports syntax errors on the wrong line.
Actually, it only does this if you use comments in the form /* comment */
For this reason I nearly always use the // form for comments despite the inconvenience.
I didn't systematically examine whether the number the error dialog shows is correct, but the line that is highlighted is wrong in my experience. And that's all I'm usually looking at because the editor does not show line numbers. A little experimentation suggests that the highlight is always off by one line if there are no block comments and no lines are auto-wrapped (i.e. the editor window needs to be wide enough). Does that match your experience?
Especially if errors occur at run-time, the wrapping problem is too bad because the editor window opens with a width that can't be influenced beforehand (or can it?) and might be too narrow. And re-sizing the window after it has opened does not change the highlight.