Re: updated turtle data files

Seaborne, Andy wrote:
> 
> 
> Lee Feigenbaum wrote:
>> Hi everyone,
>>
>> I checked into CVS updated Turtle data files that should eliminate 
>> numeric short forms in favor of explicit lexical form + datatype 
>> representations under data-r2/.
>>
>> Could people who can run most of the test suite take a shot at the 
>> latest in CVS and report any errors to me? I've done some spot 
>> checking but am not currently setup to run the full suite (oops, 
>> better get that in line by next week).
>>
>> I guess AndyS would be the best one to catch any mistakes I've made, 
>> but any other reports are welcome!
> 
> I've run the tests from CVS and from the web successfully.
> 
> 
> I had a basic check with:
> 
> find . -name \*.ttl -a ! -name manifest\* -print | xargs grep -l ' [0-9]'
> 
> I visually checked data files this produced and they look OK.
> 
> 
> It did also yield many result files (RDF graphs in our vocabulary for 
> result sets) in Turtle with short-cut numbers.
> 
> The use of rs:index does not matter but, in an ideal world, the use of 
> rs:value would change as well.

Oops, sorry that I forgot the result files in the first go round. I've 
gone through and fixed them - could you please rerun the tests from CVS 
and let me know if there are any problems?

thanks,
Lee

PS - I ended up using the following to find problem files:

grep -r '[[:space:]][0-9]' * | grep -v ':[[:space:]]*#' | grep -v CVS | 
grep \.ttl | grep -v Binary | grep -v anifest | grep -v :index | less | 
cut -d ':' -f 1 | sort | uniq



>     Andy
> 
>>
>> thanks,
>> Lee
>>
>

Received on Thursday, 13 September 2007 20:37:39 UTC