Inconsistent bottomLeft behavior

[zryip theSlug]

2010/1/14 J. Landman Gay <jacque at hyperactivesw.com>:
> zryip theSlug wrote:
>>
>> Larry obtains  "card id 1002": execution error at line n/a (Object:
>> coordinate is not a point) near "0,1020".
>>
>> Maybe I'm wrong but it is possible that the compiler is limited to the
>> first line of the content of a variable when it returns the error.
>
> The line number should point to the line of his script that has the error,
> not the line in the variable. If the compiler doesn't know which script line
> has the error, it seems like a compiler problem. But...

I wanted to point here that the content of the variable returned by
the compiler is truncated. I had not in mind that the compiler was not
able to determine where the error occured.

More I think to that more I believe that the two things are linked.
Surely I'm totally wrong in the interpretation of how the compiler
could operate, but it possible that because the variable which cause
problem is composed by more than one line, it's confused the scope of
the compiler because it can found the correct line between two lines.

It could be interresting that the compiler indicates somehow a
variable is truncated to help the debugging. Maybe an improvement to
bring in a future version of RR.

>
>> In fact RR would say to Larry that "0,1020&return&1024,0,2034,800" is
>> not a point. But the compiler try to keep the secret.
>
> True, although "item 4" would be "1020&return&0" -- which isn't a valid
> point either. We need to know if he has multiple monitors.

Yes it's true. Thanks to correct me Jacque. However when I tested it's
well 1020&return&1024 that I obtained.
It seems that "the working screenRects" takes care the direction in
which the first monitor is extended to the second.

> But the line number in the error message should not be "n/a" in any case.
> There is something wrong there.

Really

-- 
-Zryip TheSlug- wish you the best! 8)
http://www.aslugontheroad.co.cc