Ah, so that is what happened.  The first time I checked I was using extended 
syntax, the second time I was not.

I agree they should be optional.  But then nobody listens to me...  It's the 
same reason why I stopped talking to myself, I found I just wasn't listening. 
:-D



----- Original Message -----
> From: Andy Seaborne <[email protected]>
> To: [email protected]
> Cc: 
> Sent: Wednesday, June 20, 2012 10:36 AM
> Subject: Re: ARQ makes semi colons optional in SPARUL
> 
> On 20/06/12 17:13, Tim Harsch wrote:
>>  Scratch that.  My mistake.  ARQ requires the semi...
>> 
>> 
>>  ----- Original Message -----
>>>  From: Tim Harsch<[email protected]>
>>>  To: JENA-USERS<[email protected]>
>>>  Cc:
>>>  Sent: Wednesday, June 20, 2012 9:03 AM
>>>  Subject: ARQ makes semi colons optional in SPARUL
>>> 
>>>  I noticed that in ARQ I could choose not to provide a semi-colon 
> between SPARUL
>>>  commands and it would be accepted.  I just wanted to point it out in 
> case the
>>>  (a) the spec needs updating or (b) ARQ needs to be more pedantic.  If 
> it is an
>>>  intentional divergence from the spec then no matter...  but I'm 
> curious to
>>>  know what the intention was here.
>>> 
>>> 
>>>  Here is the grammar production from the spec:
>>>  [30]   Update   ::=   Prologue ( Update1 ( ';' Update )? )?
>>> 
> 
> Personally, I think they should be optional but that's consensus 
> standards for you.
> 
> At least trailing semi-colon is legal
> 
> INSERT DATA {} ;
> 
> It would better is repeated ; were legal.
> 
> INSERT DATA {} ;;
> 
> Extended syntax accepts both - none and multiple.
> 
>     Andy
>

Reply via email to