#36 closed editorial (fixed)
ABNF Update
Reported by: | mnot@… | Owned by: | julian.reschke@… |
---|---|---|---|
Priority: | normal | Milestone: | 06 |
Component: | non-specific | Severity: | Active WG Document |
Keywords: | Cc: |
Description
Update BNF to RFC4234.
Change History (82)
comment:1 Changed 15 years ago by julian.reschke@…
comment:2 Changed 15 years ago by julian.reschke@…
- Component set to non-specific
- Owner set to julian.reschke@…
comment:3 Changed 15 years ago by julian.reschke@…
comment:4 Changed 15 years ago by julian.reschke@…
comment:5 Changed 15 years ago by mnot@…
- Milestone set to unassigned
comment:6 Changed 15 years ago by mnot@…
- version set to 00-draft
comment:7 Changed 15 years ago by fielding@…
- Milestone changed from unassigned to 02
comment:8 Changed 15 years ago by julian.reschke@…
comment:9 Changed 15 years ago by julian.reschke@…
comment:10 Changed 15 years ago by julian.reschke@…
From [185]:
Get rid of duplicate BNF rule names ("host" -> "uri-host", "trailer" -> "trailer-part"), avoid underscore character in rule names ("http_URL" -> "http-URL", "abs_path" -> "path-absolute"), add rules for terms imported from URI spec ("absoluteURI", "authority", "path-absolute", "port", "query", "relativeURI", "host) -- these will have to be updated when switching over to RFC3986; addresses #36.
comment:11 Changed 15 years ago by julian.reschke@…
comment:12 Changed 15 years ago by julian.reschke@…
comment:13 Changed 15 years ago by julian.reschke@…
From [188]:
Exclude NUL from CHAR (see <http://www.w3.org/mid/52304CE2-0ADE-499A-9A5A-35284BD428BA@gbiv.com>); addresses #36.
comment:14 Changed 15 years ago by julian.reschke@…
comment:15 Changed 15 years ago by julian.reschke@…
comment:16 Changed 15 years ago by julian.reschke@…
comment:17 Changed 15 years ago by julian.reschke@…
comment:18 Changed 15 years ago by julian.reschke@…
comment:19 Changed 15 years ago by julian.reschke@…
comment:20 Changed 15 years ago by fielding@…
I would prefer that delta-seconds remain in P6, where it is most used, and the single use in P2 be replaced with 1*DIGIT. Alternatively, move it to P2 and refer to it there in P6, or simply define it more than once.
IMO, P1 should not define ABNF that it does not use.
comment:21 Changed 15 years ago by julian.reschke@…
comment:22 Changed 15 years ago by julian.reschke@…
comment:23 Changed 15 years ago by julian.reschke@…
comment:24 Changed 15 years ago by mnot@…
- Milestone changed from 02 to 03
comment:25 Changed 15 years ago by julian.reschke@…
comment:26 Changed 15 years ago by julian.reschke@…
TODO: use proper notation for case-sensitive char sequences, such as in HTTP-Version and method.
comment:27 Changed 15 years ago by julian.reschke@…
comment:28 Changed 15 years ago by julian.reschke@…
comment:29 Changed 15 years ago by julian.reschke@…
- Milestone changed from 03 to 04
comment:30 Changed 15 years ago by julian.reschke@…
comment:31 Changed 14 years ago by julian.reschke@…
comment:32 Changed 14 years ago by julian.reschke@…
- Milestone changed from 04 to unassigned
comment:33 Changed 14 years ago by julian.reschke@…
comment:34 Changed 14 years ago by julian.reschke@…
comment:35 Changed 14 years ago by julian.reschke@…
Plan for LWS:
- take out LWS
- put int OWS:
OWS = *WSP [obs-folding] obs-folding = CRLF 1*WSP
- explain when obs-folding is allowed (consumer vs producer)
- point out that this disallows continuations for producers, and only allows simple cont. for consumers (all-whitespace lines considered security risk)
- add appendix explaining for RFC2616 productions went
comment:36 Changed 14 years ago by julian.reschke@…
comment:37 Changed 14 years ago by julian.reschke@…
comment:38 Changed 14 years ago by julian.reschke@…
comment:39 Changed 14 years ago by julian.reschke@…
comment:40 Changed 14 years ago by julian.reschke@…
comment:41 Changed 14 years ago by julian.reschke@…
comment:42 Changed 14 years ago by julian.reschke@…
comment:43 Changed 14 years ago by julian.reschke@…
comment:44 Changed 14 years ago by julian.reschke@…
comment:45 Changed 14 years ago by julian.reschke@…
comment:46 Changed 14 years ago by julian.reschke@…
comment:47 Changed 14 years ago by julian.reschke@…
comment:48 Changed 14 years ago by julian.reschke@…
comment:49 Changed 14 years ago by julian.reschke@…
comment:50 Changed 14 years ago by julian.reschke@…
comment:51 Changed 14 years ago by julian.reschke@…
comment:52 Changed 14 years ago by julian.reschke@…
comment:53 Changed 14 years ago by julian.reschke@…
comment:54 Changed 14 years ago by julian.reschke@…
comment:55 Changed 14 years ago by mnot@…
- Milestone changed from unassigned to 05
comment:56 Changed 14 years ago by julian.reschke@…
comment:57 Changed 14 years ago by julian.reschke@…
comment:58 Changed 14 years ago by julian.reschke@…
comment:59 Changed 14 years ago by julian.reschke@…
comment:60 Changed 14 years ago by ylafon@…
Regarding list production: #foo => *(OWS "," [OWS foo]) 1#foo => *(OWS ",") OWS foo *(OWS "," [OWS foo])
or better
#foo => OWS *( "," OWS [foo OWS]) 1#foo => OWS *( "," OWS) foo OWS *( "," OWS [foo OWS])
comment:61 Changed 14 years ago by ylafon@…
If we need to keep the empty list elements (for consumers):
#foo => [ ( "," / foo) *(OWS "," [OWS foo]) ]
1#foo => *( "," OWS) foo *( OWS "," [OWS foo])
n#foo => *( "," OWS) foo (n-1)*( OWS "," OWS foo ) *(OWS "," [OWS foo])
Without the empty list elements (for producers):
#foo => [1#foo]
1#foo => foo *(OWS "," OWS foo)
n#foo => foo (n-1)*(OWS "," foo) *(OWS "," OWS foo)
comment:62 Changed 14 years ago by julian.reschke@…
comment:63 Changed 14 years ago by julian.reschke@…
- Milestone changed from 05 to 06
- Status changed from new to assigned
Left to do:
- rewrite general ABNF so that it refers to the header *value* ABNFs instead to the header ABNFs, then get rid of the header ABNFs.
- automatically generate appendix sections with expanded list productions.
comment:64 Changed 14 years ago by julian.reschke@…
comment:65 Changed 14 years ago by julian.reschke@…
comment:66 Changed 14 years ago by julian.reschke@…
comment:67 Changed 14 years ago by julian.reschke@…
comment:68 Changed 14 years ago by julian.reschke@…
comment:69 Changed 14 years ago by julian.reschke@…
comment:70 Changed 14 years ago by julian.reschke@…
comment:71 Changed 14 years ago by julian.reschke@…
comment:72 Changed 14 years ago by julian.reschke@…
comment:73 Changed 14 years ago by julian.reschke@…
comment:74 Changed 14 years ago by julian.reschke@…
comment:75 Changed 14 years ago by julian.reschke@…
comment:76 Changed 14 years ago by julian.reschke@…
- Resolution set to fixed
- Status changed from assigned to closed
comment:77 Changed 14 years ago by julian.reschke@…
With [543], all planned activities with respct to the ABNF syntax change are done. Please open new tickets in case new ABNF related issues come up.
comment:78 Changed 13 years ago by julian.reschke@…
comment:79 Changed 11 years ago by fielding@…
comment:80 Changed 11 years ago by julian.reschke@…
comment:81 Changed 11 years ago by naruse@…
Collectted ABNF in HTTP/1.1, part 5: Range Requests and Partial Responses draft-ietf-httpbis-p5-range-19 has typo.
WITH LIST RULE EXTENSION:
byte-range-set = 1#( byte-range-spec / suffix-byte-range-spec )
WRONG:
byte-range-set = ( *( "," OWS ) byte-range-spec ) / ( suffix-byte-range-spec *( OWS "," [ ( OWS byte-range-spec ) / suffix-byte-range-spec ] ) )
CORRECT:
byte-range-set = *( "," OWS ) ( byte-range-spec / suffix-byte-range-spec ) *( OWS "," [ OWS ( byte-range-spec / suffix-byte-range-spec ) ] )
comment:82 Changed 11 years ago by julian.reschke@…
- Priority set to normal
- Severity set to Active WG Document
Indeed. It seems that there's a bug in the conversion code that we are using. See #358. Thanks for the report.
RFC4234 will be obsoleted by RFC5234 very soon.