| Internet-Draft | CDDL feature freezer | April 2021 |
| Bormann | Expires 23 October 2021 | [Page] |
In defining the Concise Data Definition Language (CDDL), some features have turned up that would be nice to have. In the interest of completing this specification in a timely manner, the present document was started to collect nice-to-have features that did not make it into the first RFC for CDDL, RFC 8610.¶
It is now time to discuss thawing some of the concepts discussed here. A number of additional proposals have been added.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 23 October 2021.¶
Copyright (c) 2021 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
In defining the Concise Data Definition Language (CDDL), some features have turned up that would be nice to have. In the interest of completing this specification in a timely manner, the present document was started to collect nice-to-have features that did not make it into the first RFC for CDDL [RFC8610].¶
It is now time to discuss thawing some of the concepts discussed here. A number of additional proposals have been added.¶
There is always a danger for a document like this to become a shopping list; the intention is to develop this document further based on real-world experience with the first CDDL standard.¶
Section 3.5.4 of [RFC8610] alludes to a new language feature, cuts, and defines it in a fashion that is rather focused on a single application in the context of maps and generating better diagnostic information about them.¶
The present document is expected to grow a more complete definition of cuts, with the expectation that it will be upwards-compatible to the existing one in [RFC8610], before this possibly becomes a mainline language feature in a future version of CDDL.¶
Some CBOR tags often would be most natural to use in a CDDL spec with a literal syntax that is tailored to their semantics instead of their serialization in CBOR. There is currently no way to add such syntaxes, no defined extension point either.¶
The text form of CoRAL [I-D.ietf-core-coral] defines literals of the form¶
for datetime items. (Similar advances should then probably be made in diagnostic notation.)¶
Regular expressions currently are notated as strings in CDDL, with all
the string escaping rules applied once. It might be convenient to
have a more conventional literal format for regular expressions,
possibly also providing a place to add modifiers such as /i.
This might also imply text .regexp ..., which with the proposal in
Section 4.1 then raises the question of how to indicate the regular
expression flavor.¶
A number of errata reports have been made around some details of text string and byte string literal syntax: [Err6527] and [Err6543]. These need to be addressed by re-examining the details of these literal syntaxes. Also, [Err6526] needs to be applied.¶
The ABNF used in [RFC8610] for the content of text string literals is rather permissive:¶
text = %x22 *SCHAR %x22 SCHAR = %x20-21 / %x23-5B / %x5D-7E / %x80-10FFFD / SESC SESC = "\" (%x20-7E / %x80-10FFFD)¶
This allows almost any non-C0 character to be escaped by a backslash,
but critically misses out on the \uXXXX and \uHHHH\uLLLL forms
that JSON allows to specify characters in hex. Both can be solved by
updating the SESC production to:¶
SESC = "\" ( %x22 / "/" / "\" / ; \" \/ \\
%x62 / %x66 / %x6E / %x72 / %x74 / ; \b \f \n \r \t
(%x75 hexchar) ) ; \u
hexchar = non-surrogate / (high-surrogate "\" %x75 low-surrogate)
non-surrogate = ((DIGIT / "A"/"B"/"C" / "E"/"F") 3HEXDIG) /
("D" %x30-37 2HEXDIG )
high-surrogate = "D" ("8"/"9"/"A"/"B") 2HEXDIG
low-surrogate = "D" ("C"/"D"/"E"/"F") 2HEXDIG
¶
Now that SESC is more restrictively formulated, this also requires an update to the BCHAR production used in the ABNF syntax for byte string literals:¶
bytes = [bsqual] %x27 *BCHAR %x27 BCHAR = %x20-26 / %x28-5B / %x5D-10FFFD / SESC / CRLF bsqual = "h" / "b64"¶
The updated version explicit allows \', which is no longer allowed
in the updated SESC:¶
BCHAR = %x20-26 / %x28-5B / %x5D-10FFFD / SESC / "\'" / CRLF¶
The ABNF used in [RFC8610] for the content of byte string literals lumps together byte strings notated as text with byte strings notated in base16 (hex) or base64 (but see also updated BCHAR production above):¶
bytes = [bsqual] %x27 *BCHAR %x27 BCHAR = %x20-26 / %x28-5B / %x5D-10FFFD / SESC / CRLF¶
Errata report 6543 proposes to handle the two cases in separate productions (where, with an updated SESC, BCHAR obviously needs to be updated as above):¶
bytes = %x27 *BCHAR %x27
/ bsqual %x27 *QCHAR %x27
BCHAR = %x20-26 / %x28-5B / %x5D-10FFFD / SESC / CRLF
QCHAR = DIGIT / ALPHA / "+" / "/" / "-" / "_" / "=" / WS
¶
This potentially causes a subtle change, which is hidden in the WS production:¶
WS = SP / NL SP = %x20 NL = COMMENT / CRLF COMMENT = ";" *PCHAR CRLF PCHAR = %x20-7E / %x80-10FFFD CRLF = %x0A / %x0D.0A¶
This allows any non-C0 character in a comment, so this fragment becomes possible:¶
foo = h' 43424F52 ; 'CBOR' 0A ; LF, but don't use CR! '¶
The current text is not unambiguously saying whether the three apostrophes
need to be escaped with a \ or not, as in:¶
foo = h' 43424F52 ; \'CBOR\' 0A ; LF, but don\'t use CR! '¶
... which would be supported by the existing ABNF in [RFC8610].¶
Controls are the main extension point of the CDDL language. It is relatively painless to add controls to CDDL. Several candidates have been identified that aren't quite ready for adoption, of which one shall be listed here.¶
There are many variants of regular expression languages. Section 3.8.3 of [RFC8610] defines the .regexp control, which is based on XSD [XSD2] regular expressions. As discussed in that section, the most desirable form of regular expressions in many cases is the family called "Perl-Compatible Regular Expressions" ([PCRE]); however, no formally stable definition of PCRE is available at this time for normatively referencing it from an RFC.¶
The present document defines the control operator .pcre, which is
similar to .regexp, but uses PCRE2 regular expressions.
More specifically, a .pcre control indicates that the text string
given as a target needs to match the PCRE regular expression given as
a value in the control type, where that regular expression is anchored
on both sides.
(If anchoring is not desired for a side, .* needs to be inserted
there.)¶
Similarly, .es2018re could be defined for ECMAscript 2018 regular
expressions with anchors added.¶
How useful would it be to have another variant of .bits that counts bits like in RFC box notation? (Or at least per-byte? 32-bit words don't always perfectly mesh with byte strings.)¶
Provide a way to specify bitfields in byte strings and uints to a higher level of detail than is possible with .bits. Strawman:¶
Field = uint .bitfield Fieldbits Fieldbits = [ flag1: [1, bool], val: [4, Vals], flag2: [1, bool], ] Vals = &(A: 0, B: 1, C: 2, D: 3)¶
Note that the group within the controlling array can have choices, enabling the whole power of a context-free grammar (but not much more).¶
While there are no co-occurrence constraints in CDDL, many actual use cases can be addressed by using the fact that a group is a grammar:¶
postal = {
( street: text,
housenumber: text) //
( pobox: text .regexp "[0-9]+" )
}
¶
However, constraints that are not just structural/tree-based but are predicates combining parts of the structure cannot be expressed:¶
session = {
timeout: uint,
}
other-session = {
timeout: uint .lt [somehow refer to session.timeout],
}
¶
As a minimum, this requires the ability to reach over to other parts of the tree in a control. Compare JSON Pointer [RFC6901] and JSON Relative Pointer [I-D.handrews-relative-json-pointer]. Stefan Goessner's jsonpath is a JSON variant of XPath that has not been formally standardized [jsonpath].¶
More generally, something akin to what Schematron is to Relax-NG may be needed.¶
CDDL rules could be packaged as modules and referenced from other modules. There could be some control of namespace pollution, as well as unambiguous referencing ("versioning").¶
This is probably best achieved by a pragma-like syntax which could be carried in CDDL comments, leaving each module to be valid CDDL (if missing some rule definitions to be imported).¶
A convention for mapping CDDL-internal names to external ones could be developed, possibly steered by some pragma-like constructs. External names would likely be URI-based, with some conventions as they are used in RDF or Curies. Internal names might look similar to XML QNames. Note that the identifier character set for CDDL deliberately includes $ and @, which could be used in such a convention.¶
For CDDL, alternative representations e.g. in JSON (and thus in YAML)
could be defined, similar to the way YANG defines an XML-based
serialization called YIN in Section 11 of [RFC6020].
One proposal for such a syntax is provided by the cddlc tool [cddlc]; this
could be written up and agreed upon.¶
cddlj = ["cddl", +rule] rule = ["=" / "/=" / "//=", namep, type] namep = ["name", id] / ["gen", id, +id] id = text .regexp "[A-Za-z@_$](([-.])*[A-Za-z0-9@_$])*" op = ".." / "..." / text .regexp "\\.[A-Za-z@_$](([-.])*[A-Za-z0-9@_$])*" namea = ["name", id] / ["gen", id, +type] type = value / namea / ["op", op, type, type] / ["map", group] / ["ary", group] / ["tcho", 2*type] / ["unwrap", namea] / ["enum", group / namea] / ["prim", ?(0..7, ?uint)] group = ["mem", null/type, type] / ["rep", uint, uint/false, group] / ["seq", 2*group] / ["gcho", 2*group] value = ["number"/"text"/"bytes", text]¶
This document makes no requests of IANA.¶
Many people have asked for CDDL to be completed, soon. These are usually also the people who have brought up observations that led to the proposals discussed here. Sean Leonard has campaigned for a regexp literal syntax.¶