Updated to removed unresolvable choice possibility: *Productions* Document = Element Element = SimpleElement | ComplexElement SimpleElement = LeftSimpleFraming SimpleContent RightFraming ComplexElement = LeftComplexFraming ComplexContent RightFraming LeftSimpleFraming = Prefix *LeftPadding* LeftComplexFraming = Prefix RightFraming = *RightPadOrFill * Suffix ComplexContent = Sequence | Choice Sequence = Prefix SequenceContent Suffix SequenceContent = [ *PrefixSeparator* SequenceItem [ *Separator* SequenceItem ]* * PostfixSeparator* ] SequenceItem = Element | Array | ComplexContent Choice = Prefix ChoiceContent Suffix ChoiceContent = Element | Array | ComplexContent Array = [ Element [ *Separator* Element ]* [ *Separator* StopValue] ] StopValue = SimpleElement Prefix = LeadingAlignment *Initiator* Suffix = *Terminator* TrailingAlignment LeadingAlignment = *LeadingSkipBytes* *AlignmentFill* TrailingAlignment = *TrailingSkipBytes* On Wed, Jun 17, 2009 at 6:37 AM, Alan Powell <alan_powell@uk.ibm.com> wrote:
Mike
I think you have prefix/postfix separator in the wrong place
You definition has LeadingAlignment Initiator Prefixseparator and it should be Prefixseparator LeadingAlignment Initiator so
LeftSequenceFraming = Prefix *PrefixSeparator* should be LeftSequenceFraming = *PrefixSeparator *Prefix
Similarly for postfix.
Also I don't think ChoiceUnresolvableItem = OpaqueTypeContent is correct. Unresolveable choice content is the same as resolvable choice content as you need to be able to work out the length of the choices (which must be the same).
I have change *AlignmentPad* to *AlignmentFill *as the pad character is not used during alignment
Small stylistic comment: Should LeftPadding be just Padding and RightPadorFill be just PadorFill?
Alan Powell
MP 211, IBM UK Labs, Hursley, Winchester, SO21 2JN, England Notes Id: Alan Powell/UK/IBM email: alan_powell@uk.ibm.com Tel: +44 (0)1962 815073 Fax: +44 (0)1962 816898
From: Mike Beckerle <mbeckerle.dfdl@gmail.com> To: Steve Hanson/UK/IBM@IBMGB, Alan Powell/UK/IBM@IBMGB Date: 11/06/2009 03:23 Subject: please review revised DFDL grammar ------------------------------
I'll take this to the broader group once you guys have had a look at this.
Please review this revised grammar table. I've tried to include the necessary pieces to handle padding on left and padding or fill on right, and in a manner consistent with our most recent policies where dfdl:length for complex types can only be on the element itself, not any model group.
I am uncertain if this grammar has enough features to capture everything about say, CalendarTypeContent. E.g., are there no properties about framing that are special to calendars? This grammar suggests that padding/fill is all handled outside the content for calendars, as for all simple types.
I also added prefix/postfix separators back in. (We excised them a while back, but I'm pretty sure decided to put them back in.)
The issue about proper interpretation of leading and trailing skip bytes is outstanding still.
There are no changes to the simple content grammar. I'm including it here for making this message self-contained.
I believe that the blowout pictures in the grammar section are to some degree outliving their usefulness. They are in essence equivalent to this grammar. By arranging the grammar to use the terms left framing, right framing, and content, I believe the information about what start of content is, vs. start of item can be defined without most of the pictures. I think the one which defines item start, content start, content end, and item end is useful. The others not so useful.
-------------------------
First we present the simple content types.
*Production* *See Section(s)* SimpleContent = NumberTypeContent | StringTypeContent |
CalendarTypeContent | BooleanTypeContent | OpaqueTypeContent NumberTypeContent = *TextNumber* | *BinaryIntegerNumber* | * BinaryFloatNumber* TBD StringTypeContent = *TextString* TBD CalendarTypeContent = *TextCalendar* | *BinaryIntegerCalendar* TBD BooleanTypeContent = * TextBoolean* | *BinaryIntegerBoolean* TBD OpaqueTypeContent = * BinaryOpaque* TBD
There are additional NumberTypeContent terminals specified in the DFDL specification supplements (for bcd, zoned and packed decimal).
These simple types are the basic data content pieces found in DFDL-described data sources.
The rest of this grammar describes how these basic pieces are composed into arrays and more complex types.
*Productions* * * Document = Element Element = SimpleElement | ComplexElement
SimpleElement = LeftSimpleFraming SimpleContent RightFraming
ComplexElement = LeftComplexFraming ComplexContent RightFraming
LeftSimpleFraming = Prefix *LeftPadding*
LeftComplexFraming = Prefix
RightFraming = *RightPadOrFill * Suffix
ComplexContent = Sequence | Choice Sequence = LeftSequenceFraming SequenceContent RightSequenceFraming
SequenceContent = [ SequenceItem [ *Separator* SequenceItem ]* ]
SequenceItem = Element | Array | ComplexContent
LeftSequenceFraming = Prefix *PrefixSeparator*
RightSequenceFraming = *PostfixSeparator *Suffix Choice = LeftChoiceFraming ChoiceContent RightChoiceFraming
ChoiceContent = ChoiceResolvableItem | ChoiceUnresolvableItem*[MJB1]*<#1225559d07ce240a_121f6734f907cd24_121edcae928b9ee0__msocom_1>
ChoiceResolvableItem = Element | Array | ComplexContent
ChoiceUnresolvableItem = OpaqueTypeContent*[MJB2]*<#1225559d07ce240a_121f6734f907cd24_121edcae928b9ee0__msocom_2>
LeftChoiceFraming = Prefix
RightChoiceFraming = Suffix Array = [ Element [ *Separator* Element ]* [ *Separator* StopValue] ]
StopValue = SimpleElement Prefix = LeadingAlignment *Initiator*
Suffix = *Terminator* TrailingAlignment LeadingAlignment = *LeadingSkipBytes* *AlignmentPad*
TrailingAlignment = *TrailingSkipBytes*
------------------------------
<#1225559d07ce240a_121f6734f907cd24_121edcae928b9ee0__msoanchor_1>
------------------------------
* *
*Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU *