[ESS] Previous/Next error in ESS tracebug

[Sebastien Vauban]

Vitalie Spinu wrote:
> Sebastien Vauban on Thu, 21 Aug 2014 16:04:09 +0200 wrote:
>> stop("Test error")
>> source(file = "missing-file.R")
>> list(a=22, b=343)
>> ^
>> protocol <- NULL
>> paste("{**", })
>>
>> I have "No traceback available" when pressing C-c `...
>
> You will have a traceback whenever R provides it. In this very simple
> example it's a parsing error at level 1.

In fact, there are multiple errors:

- stop command,
- missing file which we try to source, and
- 2 parsing errors (the ^ and the })

> For non-parsing errors you should always get a trace.

Do you mean it should be the case for the "stop" and the "source" of
a missing file?

>> Not in such an iESS buffer, though:
>
> Same as above. Parsing errors don't have references with evaluation
> commands. I am pretty sure this could be fixed with some more tweaking
> but I don't have time for that right now.

OK.

>>     "Reference to '<path-to-file>' not found"
>>
>> Somehow, the "c" letter drive is not parsed correctly, and the file is
>> not found then... See http://screencast.com/t/dRGLA4kn5mO.
>
> This is a bug. I have just fixed it.

I don't see it yet on ELPA. I currently still have:

  ess                          20140821.1009  installed

As soon as it's there, I'll test it!

> I am really surprised nobody reported it so far. It means that error
> navigation never worked properly on windows.

As I really want to both use more R and use more ESS, I'll be happy to
interact more with (all of) you...

> Thanks a lot for reporting this.

You're welcome!  Thanks to YOU.

>> UPDATE: I just found out that if I eval my above buggy code with C-c
>> C-l (instead of C-c C-c),
>
> Yes. Loading the file or sourcing it directly usually work better for
> parsing errors.

That's an R feature -- not ESS, right?

Best regards,
  Seb

-- 
Sebastien Vauban