better error messages for runtime errors requested

[GoogleCodeExporter]

 In my opinion the error message is too spare in case of run-time errors.

Take also a look to the attached file.
The error message points to the top-most calling function only. I spent already 
hours to discover the correct location for such kind of error. This is not 
acceptable when we are using around 5000 lines of code and more (as in my case) 
-- especially by the lack of a debugger.

Original issue reported on code.google.com by analogc...@gmx.net on 9 Nov 2012 at 9:59

Attachments:

• bug1.tll

[GoogleCodeExporter]

Original comment by krustev....@gmail.com on 11 Nov 2012 at 9:53

• Changed state: Started

[GoogleCodeExporter]

Updated in r2307

This is one of those temporary patches. The update (I don't even call it fix) 
is certainly better than nothing, but for a really satisfying solution we need 
a debug functionality.
The original error message (which is confusing I know) remains. Now there is 
something in addition which could be called stack unroll, but only function 
names. No location info. And that's the main trouble - location info is 
maintained only in the parser - this is runtime error and none of the 
components keep it.

When we come to a debugger - then that'll be the first thing to add anyway.
Hope that will do for the moment...

Original comment by krustev....@gmail.com on 11 Nov 2012 at 10:03

• Changed state: Fixed

[GoogleCodeExporter]

that is already a real progress. Thanks.

Do you think you can put even more information into the message. 
My Imagination is something like: "Invalid-Index 5 (max length is 4) for list 
with variable name 'numbers'". If the list is not assigned to any variable it 
could also be " .. for unassigned list" --> even that is helpful. 

Original comment by analogc...@gmx.net on 11 Nov 2012 at 10:19

[GoogleCodeExporter]

That should be possible - I'll have another look to see how easy it can be ...

Original comment by krustev....@gmail.com on 12 Nov 2012 at 4:10

[GoogleCodeExporter]

The latest updates in r2308 should produce a better error message as requested 
in the previous comment.
Another problem was found (see issue 154) with list union which should be fixed 
shortly.

Original comment by krustev....@gmail.com on 13 Nov 2012 at 2:54

[GoogleCodeExporter]

thanks a lot for your ultra-fast fixes .. 
If I'll have problems again to find the position of some of my next runtime 
errors probably I reopen this issue

Original comment by analogc...@gmx.net on 13 Nov 2012 at 10:24

• Changed state: Verified