January 2017 - dfdl-wg - lists.ogf.org

clarification needed: unparsing and hidden group choices and choices with dfdl:choiceDispatchKey
by Mike Beckerle 13 Dec '19

13 Dec '19

So interesting issue. Comes up as part of our experience implementing the advanced expression language features: hidden groups, and computed elements. We have hidden groups in DFDL. If a xs:choice is inside a hidden group, then by definition, when unparsing, elements corresponding to some branch of the choice are not in the Infoset. DFDL spec clearly states that dfdl:outputValueCalc is not allowed on the head of a choice branch. But elements inside hidden groups must have dfdl:outputValueCalc, or a default value, otherwise there's no way for them to be filled in as part of the augmented infoset. So the only way a choice can be inside a hidden group is if the head of the choice branches have default values. This is not sufficient. I do not think it was our intention to make hidden groups incapable of holding things complex enough to contain a choice. Rather, the point is to be able to hide things that have excess complexity needed only for parsing/unparsing, so that they need not appear in the final Infoset. So if a choice is allowed inside a hidden group, then a mechanism is needed by which when unparsing a choice branch is selected based on values of other parts of the Infoset, and then the elements of that branch are computed and added to the augmented infoset. Note that without hidden groups, this problem never arises, as the Infoset always has to take a position on choice branches by having an unambiguous element in it. There are 3 cases: 1. unparsing an xs:choice where for parsing, each branch has a discriminator, or is resolved by the simple "first successful parse" rule. 2. unparsing an xs:choice with dfdl:choiceDispatchKey 3. unparsing xs:choice with dfdl:initiatedContent='yes' For all cases, I suggest we must remove the restriction on dfdl:outputValueCalc being disallowed on choice branch head. For (1). First branch where the branch can be constructed successfully is added to the augmented infoset. This is a case where an unparse error is non-fatal, and causes a form of unparser backtracking through alternatives. Branches would be attempted in schema order. For (2) I suggest dfdl:unparseChoiceDispatchKey, evaluated at unparse time only. Produces a choice branch key. That choice branch is then constructed, and default values or dfdl:outputValueCalcs are used to fill in values. (BTW: this is the driving use case we have in Daffodil for this issue.) For (3).... I have no specific solution. I think it may not actually be a separate case and has to be treated as (1) or (2). Comments/Discussion?

2 1

Action 287: Find a way to handle a variable path step in DFDL expression - Re: DFDL4S use of wildcard/regex in length path expression
by Mike Beckerle 13 Nov '17

13 Nov '17

I have analyzed this use of (.*) notation in the Sentinel2X-bandTMISPData.xsd file. Below is my discussion of what this is, why it is used, the alternative to it, and why it is more problematic than it seems at first, as an addition to DFDL. There are two instances of (.*) in the Sentinel schema files. What it expresses is a partial step-name wildcard. This is not a variation on the XPath * notation, as that matches any step, and can't match a partial name. In the usage in the Sentinel schema, this wild card is a regex that matches against a step name, and it matches exactly a single name. It is not used in a way where it can result in a node-set instead of a single node. The question really is why is this (.*) needed - that is what is it trying to achieve, and whether there is an acceptable alternative for what it is doing. I find that this wildcard (.*) is used to achieve a parameterization of the types TypeISPData and TypeISPData_HKTM. These types are polymorphic, in that their exact behavior depends on aspects of their surrounding context. Each of these types incorporates a length of some content that is outside of its own definition. As the types are used now, the length comes from a thing outside of them in the schema that happens to have a particular suffix on its name, which is "Packet_Secondary_Header" or just "_Secondary_Header". So these names, while outside of the type, are in some sense being hard-coded in these types. Even though they are outside of these types, you cannot change the names of these elements without breaking the ability of the type to find them via this (.*) notation and a name suffix. These types, TypeISPData and TypeISPData_HKTM are placed inside other types, and those are then placed in context with various packet-header structures. Those structures have various distinguishing prefixes of sub-elemetnts such as MSI_Packet_Secondary_Header. There are a variety of different things instead of "MSI_", but there's always something with suffix "Packet_Secondary_Header" or "Secondary_Header" in it. The use of this (.*) name wildcard seems convenient, but doesn't offer anything that isn't better captured by a true parameterization mechanism which decouples the names used outside the type from those used inside it. DFDL provides a general mechanism for this sort of parameterization using variables and dfdl:newVariableInstance. Let's look at just one of the two instances, TypeISPData. To use variables, a variable is created which represents this parameter to the TypeISPData. It is declared in the schema file where TypeISPData is defined. <dfdl:defineVariable name="Additional_Content_Length" type="xs:int"/> The expression within TypeISPData that currently contains "contentLength(/Packet_Data_Field/(.*)Packet_Secondary_Header, 'bytes')" That part of the expression instead references the parameter variable $Additional_Content_Length At the point of use, where this packet secondary header element is combined with the element that contains the TypeISPData, at that location, a dfdl:newVariableInstance is created and bound like so: ... <xs:element name="MSI_Packet_Secondary_Header" type="TypePacketData_MSI"/> <xs:sequence> <xs:annotation><xs:appinfo source="http://www.ogf.org/dfdl"> <dfdl:newVariableInstance ref="Additional_Content_Length" value="{ dfdl:contentLength(../MSI_Packet_Secondary_Header, 'bytes') }"/> </xs:appinfo></xs:annotation> <xs:element name="MSI_User_Data_Field" type="TypeUserData"/> </xs:sequence> ... This binds the parameter to the needed length around the point of use of the type that needs it (which is down inside the TypeUserData) This seems bulky due to the XML and XSD-annotation notational overheads, but it is just parameter-binding as would occur in an ordinary programming language when passing an argument. Using variables decouples the names of the two parts of the schema entirely. For example, if MSI_Packet_Secondary_Header were to be renamed to MSI_PSH, it would change only the point of use where the variable binding occurs, and would not affect the type definition. This also facilitates testing the types in isolation. It eliiminates the need to have the complete data structure surrounding them. Use of (*.) name matching might seem convenient, but really it depends on an unstated invariant about the way names are chosen in the Sentinel schema. If the names weren't constructed so uniformly, this notational trick would be unable to make the necessary distinctions, and you'd have to fall back on using variables. So, the above explains an alternative, already in DFDL, that can achieve the parameterization of types that is needed. There is an additional issue that makes this (.*) notation problematic. The issue is thes semantics relative to QNames and ordinary XML namespace management for name conflict control, The Sentinel schema is all a "no namespace" schema. In that world a path step is just an identifier with no structure. However, more generally in DFDL, a path step might be a QName, and those have a namespace prefix part, and a name part. Name collisions and conflicts can be managed by use of different target namespaces and namespace prefixes. Using (.*) prefix regex-matching against QNames is problematic, as it subverts the ability to use namespace prefixes to eliminate name conflicts. E.g., does (.*)_Secondary_Header match "foo:My_Secondary_Header" given that it is in the foo namespace? Matching against things without taking namespace prefixes into consideration breaks XSD's ability to use prefixes to manage name collsions. So the answer of does "(.*)_Secondary_Header" match "foo:My_Secondary_Header" would have to be that it does not match, and that if one wanted to match that one would have to write "bar:(.*)_Secondary_Header" where bar prefix is bound to the same namespace as foo in the schema file where My_Secondary_Header is defined. This mixture of some plain-textual regex matching, and some name-prefix-qualified namespace-sensitive matching is not very attractive. There is certainly nothing like this in XPath. Last issue is implementation complexity: So, if one wants to do without the (.*) notation, in terms of making DFDL4S more complex to implement, you would need to add variables. Fortunately variables are really one of the simpler things in DFDL to implement at least for parsing, and is very simple if the dfdl:setVariable annotation is not implemented since it is not needed in this case. Each variable has a stack. the dfdl:newVariableInstance pushes a new entry onto the stack, and in this case, populates it immediately with a value. Reference to the variable from an expression always takes the value of the top-of-stack location. Exiting the scope of the annotation where the dfdl:newVariableInstance appeared pops the stack. In summary, given (a) the availability of a robust parameterization mechanism using parameter variables and newVariableInstance to bind them (b) the ease of implementing variables, and (c) the complexity of working out interactions with XML/XSD namespaces/prefixes and name management,..... given those reasons, I would be disinclined to advocate something like this (.*) notation to DFDL unless it was first added to XPath, where all necessary details were worked out for us. Mike Beckerle | OGF DFDL Workgroup Co-Chair | Tresys Technology | www.tresys.com Please note: Contributions to the DFDL Workgroup's email discussions are subject to the OGF Intellectual Property Policy <http://www.ogf.org/About/abt_policies.php> On Tue, Sep 13, 2016 at 10:18 AM, Michele Zundo <michele.zundo(a)esa.int> wrote: > Dear Steve, > > sorry for the delay due to Summer break and other projects. > > Here is the .zip. > > Please note that we realised that one of our previous replies date 25 July > 2016 at 17:45:05 GMT+2 > was a misunderstanding from our part and not applicable. > > Regarding your questions please find answers interleaved below > > Michele > > This message and any attachments are intended for the use of the addressee or addressees only. > The unauthorised disclosure, use, dissemination or copying (either in whole or in part) of its > content is not permitted. > If you received this message in error, please notify the sender and delete it from your system. > Emails can be altered and their integrity cannot be guaranteed by the sender. > > Please consider the environment before printing this email. > > > > > Regards > > > > > > > *From:* Steve Hanson [*smh(a)uk.ibm.com* <smh(a)uk.ibm.com>] > *Sent:* Tuesday, July 26, 2016 3:20 AM > *To:* Michele Zundo > *Cc:* Mike Beckerle; *rui.mestre(a)deimos.com.pt* <rui.mestre(a)deimos.com.pt> > *Subject:* Re: Fwd: OGF DFDL WG Call Minutes 2016-07-05 > > > ….snip > > > dfdl:length="{/Packet_Primary_Header/Packet_Data_Length + 1 - > contentLength(/Packet_Data_Fi > eld/(.*)Packet_Secondary_Header, 'bytes') - 2}" > > Firstly, contentLength is a DFDL function so it needs to be in the DFDL > namespace, eg, dfdl:contentLength(). > > > Yes agree with you. We will add the* dfdl:* it in future releases and > modify the applications accordingly. > > > Secondly, the first argument to dfdl:contentLength() is a path, so you are > effectively still using regular expressions in path steps. > > > Yes. For now we are using it and expect this to become part of the > standard. > > > Regards > > Steve Hanson > *IBM Integration Bus* > <http://www-03.ibm.com/software/products/en/ibm-integration-bus>, > Hursley, UK > Architect, *IBM DFDL* > <http://www.ibm.com/developerworks/library/se-dfdl/index.html> > Co-Chair, *OGF DFDL Working Group* <http://www.ogf.org/dfdl/> > *smh(a)uk.ibm.com* <smh(a)uk.ibm.com> > tel:*+44-1962-815848* <%2B44-1962-815848> > mob:*+44-7717-378890* <%2B44-7717-378890> > > > > From: Michele Zundo <*michele.zundo(a)esa.int* > <michele.zundo(a)esa.int>> > To: Steve Hanson/UK/IBM@IBMGB > Cc: Mike Beckerle <*mbeckerle(a)tresys.com* <mbeckerle(a)tresys.com>> > Date: 25/07/2016 17:10 > Subject: Fwd: OGF DFDL WG Call Minutes 2016-07-05 > ------------------------------ > > > > > Dear Steve, > > Please find below the answer from our developers and example. > > Note that we have updated our implementation of DFDL to be as compliant as > we > can at this point in time with the exception noted below. > > Michele > > Begin forwarded message: > > *From: *"Rui Mestre (DME)" <*rui.mestre(a)deimos.com.pt* > <rui.mestre(a)deimos.com.pt>> > *Subject: Re: Fwd: OGF DFDL WG Call Minutes 2016-07-05* > *Date: *25 July 2016 at 17:45:05 GMT+2 > > Dear Michele, > > I believe that after our DFDL compliance effort the mentioned "use of a > regex in the path step of a DFDL expression" is no longer in place. > > Currently the only extension implemented in DFDL4S regarding the use of > regular expressions is that implementation of dfdl:contentLength is > extended to support also regular expressions when specifying the node. > > Please find attached a schema file example containing such extension in > the use of dfdl:contentLength. > > Best regards, > Rui > > > Begin forwarded message: > > *From: *Steve Hanson < <smh(a)uk.ibm.com>*smh(a)uk.ibm.com* <smh(a)uk.ibm.com>> > *Subject: OGF DFDL WG Call Minutes 2016-07-05* > *Date: *5 July 2016 at 17:49:13 GMT+2 > *To: * <dfdl-wg(a)ogf.org>*dfdl-wg(a)ogf.org* <dfdl-wg(a)ogf.org> > *Cc: *"Mike Beckerle" < <mbeckerle(a)tresys.com>*mbeckerle(a)tresys.com* > <mbeckerle(a)tresys.com>>, "Michele Zundo" <*michele.zundo(a)esa.int* > <michele.zundo(a)esa.int>> > > Please find minutes from the above call at > <https://redmine.ogf.org/dmsf_files/13537?download=> > *https://redmine.ogf.org/dmsf_files/13537?download=* > <https://redmine.ogf.org/dmsf_files/13537?download=> > > *@Michele - please can you send to the WG a schema that shows your use of > a regex in the path step of a DFDL expression ?* > > Next call *Aug 2nd* > > Regards > > Steve Hanson > Architect, IBM DFDL, > Co-Chair, *OGF DFDL Working Group* <http://www.ogf.org/dfdl/> > IBM SWG, Hursley, UK > *smh(a)uk.ibm.com* <smh(a)uk.ibm.com> > *tel:+44-1962-815848* <+44-1962-815848> > > ----------------------------------------- > Michele Zundo > > Head of Ground System Definition and Verification Office > EOP-PEP > European Space Agency, ESTEC > e-mail: <michele.zundo(a)esa.int>*michele.zundo(a)esa.int* > <michele.zundo(a)esa.int> > > > This message and any attachments are intended for the use of the addressee > or addressees only. > The unauthorised disclosure, use, dissemination or copying (either in > whole or in part) of its > content is not permitted. > If you received this message in error, please notify the sender and delete > it from your system. > Emails can be altered and their integrity cannot be guaranteed by the > sender. > > Please consider the environment before printing this email. > > ----------------------------------------- > Michele Zundo > > Head of Ground System Definition and Verification Office > EOP-PEP > European Space Agency, ESTEC > e-mail: *michele.zundo(a)esa.int* <michele.zundo(a)esa.int> > > > > > > > > > > > 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 > > > #### Sentinel2X-bandTMISPData.xsd moved to MyAttachments Repository V3.8 ( > Link > <notes:///802575AF0030E827/5DE5236E5AD1645685256EE0001BBADF/ABC83CA7B3700B0380258018002AE581>) > on 23 August 2016 by Steve Hanson. > > > 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 > > > ----------------------------------------- > Michele Zundo > > Head of Ground System Definition and Verification Office > EOP-PEP > European Space Agency, ESTEC > e-mail: michele.zundo(a)esa.int > > > > > > > > > >

3 5

Re: [DFDL-WG] clarification needed - choice with both dispatch and discriminator
by Andrew Edwards 27 Jan '17

27 Jan '17

1 0

clarification needed - choice with both dispatch and discriminator
by Mike Beckerle 27 Jan '17

27 Jan '17

Section 15.1.2 says you can have both dfdl:choiceBranchKey and a discriminator on a choice branch. But it doesn't clarify what the meaning of having both is. My assumption is that if you have both, the choice is resolved by the choiceDispatchKey/choiceBranchKey, so any discriminator would be about some surrounding point of uncertainty outside the choice. I wanted to confirm this is the intention. The alternative semantics would be that the discriminator is redundant, and so if it evaluates to true that means nothing at all since the choice branch has already been discriminated by the choiceDispatchKey/choiceBranchKey. If the discriminator evaluates to false, or errors when it evaluates, that would be a parse error that would cause the entire choice to fail. Comments? Mike Beckerle | OGF DFDL Workgroup Co-Chair | Tresys Technology | www.tresys.com Please note: Contributions to the DFDL Workgroup's email discussions are subject to the OGF Intellectual Property Policy <http://www.ogf.org/About/abt_policies.php>

3 2

suggest: need hexBinary with lengthUnits 'bits' with length not a multiple of 8.
by Mike Beckerle 27 Jan '17

27 Jan '17

We have users who have binary blobs the size of which is given in bits, and these blobs are not a multiple of 8 long. Today the DFDL spec doesn't allow hexBinary to have lengthUnits 'bits'. I am wondering if this restriction should be lifted. XSD constrains hexBinary to always have an even number of Hex digits, so we would have to do the same. So for an example, a 17 bit long hexBinary containing all 1 bits would be FFFF80 Erratum 5.15 extends the types that are allowed to have length in bits to include packed calendars. So there is precedent for opening this restriction up if need arises. I claim we need to (a) allow length units bits for all types (b) restrict the length to have to be 32-bits or 64-bits only, for types xs:float and xs:double when representation 'binary' (c) restrict packed decimal to have lengths be a multiple of 4 bits (when specified in units 'bits') All other restrictions should be lifted as those restrictions just cause problems in some formats. For example 12.3.7.2.5 Specifies that binary calendars must be 4 bytes or 8 bytes exactly, and cannot be specified in units 'bits'. This is just a mistake in DFDL. I have even seen binary calendars with 33 bits length. (seconds since 1-1-1970 representation aka binarySeconds) That additional bit extends the end time substantially. These restrictions were put into DFDL because our experience of many bit-granularity formats was limited. What we've found is that there are plenty of data formats where the notion of a "byte" is simply absent. Nothing uses multiples of 8 bits for anything, and nothing is measured in those units. It's always measured in bits. Even for things like float and double, which have impliicit lengths of 4 and 8 bytes respectively, many specifications will express those as 32 bits or 64 bits. Having to divide by 8 just makes the DFDL schema awkward. Similarly in these formats strings are given length in bits. 448 bits worth of 7-bit packed ascii characters is 64 characters, occupying 56 bytes, but the spec uses 448. These changes are all backward-compatible. They make legal property settings that previously had no meaning and caused SDEs. Discussion? Mike Beckerle | OGF DFDL Workgroup Co-Chair | Tresys Technology | www.tresys.com Please note: Contributions to the DFDL Workgroup's email discussions are subject to the OGF Intellectual Property Policy <http://www.ogf.org/About/abt_policies.php>

2 2

OGF DFDL WG Call 2017-01-24 - cancelled
by Steve Hanson 24 Jan '17

24 Jan '17

Steve not available so cancelling. Regards Steve Hanson Architect, IBM DFDL Co-Chair, OGF DFDL Working Group Hursley, UK smh(a)uk.ibm.com tel:+44-1962-815848

1 0

Cancelled: OGF DFDL Working Group call
by Steve Hanson 24 Jan '17

24 Jan '17

1 0

OGF DFDL WG Call Minutes 2017-10-01
by Steve Hanson 16 Jan '17

16 Jan '17

Please find minutes from the above call at https://redmine.ogf.org/dmsf_files/13548?download= Next call Jan 24th. Regards Steve Hanson Architect, IBM DFDL, Co-Chair, OGF DFDL Working Group IBM SWG, Hursley, UK smh(a)uk.ibm.com tel:+44-1962-815848

1 0

Announcement: Daffodil now runs NACHA DFDL Schema
by Mike Beckerle 12 Jan '17

12 Jan '17

(I posted this to the news items on the OGF DFDL Workgroup pages, but here it is for those following dfdl-wg also) Tresys Technology is pleased to announce that the Daffodil <https://exchange10.columbia.tresys.com/owa/redir.aspx?C=abb4b9a398b24e08ac9…> open source DFDL implementation is now able to run the NACHA Electronic Payment messages DFDL schema <https://exchange10.columbia.tresys.com/owa/redir.aspx?C=abb4b9a398b24e08ac9…>. This demonstrates an increase in interoperability for Daffodil with DFDL schemas created and tested on the IBM DFDL implementation <https://exchange10.columbia.tresys.com/owa/redir.aspx?C=abb4b9a398b24e08ac9…> . The version of Daffodil that has this functionality is still pre-release, but the forthcoming Daffodil 2.0.0 release will incorporate this capability. (Daffodil 2.0.0 development snapshots newer than 2017-01-11 can also be used.) The NACHA Electronic Payment messages DFDL schema <https://exchange10.columbia.tresys.com/owa/redir.aspx?C=abb4b9a398b24e08ac9…> was originally created and tested by IBM in 2014. Mike Beckerle | OGF DFDL Workgroup Co-Chair | Tresys Technology | www.tresys.com Please note: Contributions to the DFDL Workgroup's email discussions are subject to the OGF Intellectual Property Policy <http://www.ogf.org/About/abt_policies.php>

1 0

Announcement: NITF DFDL Schema
by Steve Lawrence 12 Jan '17

12 Jan '17

Tresys Technology is pleased to announce the release of a DFDL schema to describe MIL-STD-2500C: National Imagery Transmission Format, more commonly known as NITF. The schema and test files are available for download on the DFDLSchemas github page [1]. The version of Daffodil that has the functionality to parse and unparse NITF files using this schema is still pre-release, but the forthcoming Daffodil 2.0.0 release will incorporate the capability. (Daffodil 2.0.0 development snapshots newer than 2017-01-12 can also be used.) [1] https://github.com/DFDLSchemas/NITF

1 0