abnftt translates a grammar in an augmented ABNF syntax into a treetop
grammar that can be compiled into a PEG parser.
Installation: gem install abnftt
Usage:
abnftt mygrammar.abnftt
generates three files:
-
mygrammar.abnf-- a copy of the pure ABNF in the abnftt file, with the "semantic predicates" (treetop term) removed. -
mygrammar.treetop-- a translation of the augmented ABNF in the abnftt file to the treetop grammar format. -
mygrammar.yaml-- a YAML representation of the parse tree of the augmented ABNF file.Use
tt mygrammar.treetopto obtain amygrammar.rb.
An example for the use of abnftt is available in the cddlc gem.
abnfrob is intended as a general tool for processing and converting
ABNF grammars.
Its basic function is reading and parsing an ABNF file and writing an equivalent one based on the parse tree.
As of today, it provides only one processing function that can be performed on the parse tree before writing the output ABNF file: Squashinating.
If you have a cbor-edn-h.abnf like this:
app-string-h = S *(HEXDIG S HEXDIG S / ellipsis S)
["#" *non-lf]
ellipsis = 3*"."
HEXDIG = DIGIT / "A" / "B" / "C" / "D" / "E" / "F"
DIGIT = %x30-39 ; 0-9
blank = %x09 / %x0A / %x0D / %x20
non-slash = blank / %x21-2e / %x30-10FFFF
non-lf = %x09 / %x0D / %x20-D7FF / %xE000-10FFFF
S = *blank *(comment *blank )
comment = "/" *non-slash "/"
/ "#" *non-lf %x0A…calling
abnfrob --squash=h cbor-edn-h.abnf
gives you this cbor-edn-h-sq.abnf (manually line-broken for readability):
sq-h-app-string-h = %s"h'" h-app-string-h "'"
h-app-string-h = h-S *(h-HEXDIG h-S h-HEXDIG h-S / h-ellipsis h-S)
[("#" / %s"\u" ("0023" / "{" *("0") "23}")) *(h-non-lf)]
h-ellipsis = 3*("." / %s"\u" ("002E" / "{" *("0") "2E}"))
h-HEXDIG = h-DIGIT
/ %x41-46 / %s"\u" ("004" %x31-36 / "{" *("0") "4" %x31-36 "}")
/ %x61-66 / %s"\u" ("006" %x31-36 / "{" *("0") "6" %x31-36 "}")
h-DIGIT = %x30-39 / %s"\u" ("003" %x30-39 / "{" *("0") "3" %x30-39 "}")
h-blank = %x0a / %s"\n" / %s"\r" / %s"\t"
/ %s"\u" ("000" ("9" / %s"A" / %s"a") / "{" *("0") ("9" / %s"A" / %s"a") "}")
/ %s"\u" ("000D" / "{" *("0") "D}") / " " / %s"\u" ("0020" / "{" *("0") "20}")
h-non-slash = h-blank / %x21-26 / %x28-2e / "\'"
/ %s"\u" ("002" h-x1e / "{" *("0") "2" h-x1e "}")
/ %x30-5b / %x5d-d7ff / %xe000-10ffff / "\\"
/ %s"\u" (h-x1f 3(h-x0f) / "0" h-x1f 2(h-x0f) / "00" h-x3f h-x0f
/ "D" h-x8b 2(h-x0f) %s"\u" "D" h-xcf 2(h-x0f)
/ "{" *("0") ("10" 4(h-x0f) / h-x1f 4(h-x0f)
/ h-x1f 3(h-x0f) / h-x1f 2(h-x0f) / h-x3f h-x0f) "}")
h-non-lf = %s"\t" / %s"\u" ("0009" / "{" *("0") "9}") / %s"\r"
/ %s"\u" ("000D" / "{" *("0") "D}") / %x20-26 / %x28-5b
/ %x5d-d7ff / "\'" / "\/" / "\\"
/ %s"\u" ("D" %x30-37 2(h-x0f) / h-x1c 3(h-x0f)
/ "0" h-x1f 2(h-x0f) / "00" h-x2f h-x0f
/ "{" *("0") ("D" %x30-37 2(h-x0f) / h-x1c 3(h-x0f)
/ h-x1f 2(h-x0f) / h-x2f h-x0f) "}")
/ %xe000-10ffff
/ %s"\u" (h-xef 3(h-x0f) / "D" h-x8b 2(h-x0f) %s"\u" "D" h-xcf 2(h-x0f)
/ "{" *("0") ("10" 4(h-x0f) / h-x1f 4(h-x0f) / h-xef 3(h-x0f)) "}")
h-S = *(h-blank) *(h-comment *(h-blank))
h-comment = ("/" / "\/" / %s"\u" ("002F" / "{" *("0") "2F}"))
*(h-non-slash) ("/" / "\/" / %s"\u" ("002F" / "{" *("0") "2F}"))
/ ("#" / %s"\u" ("0023" / "{" *("0") "23}"))
*(h-non-lf) (%x0a / %s"\n" / %s"\u" ("000A" / "{" *("0") "A}"))
h-x0f = %x30-39 / %x41-46 / %x61-66
h-x1c = %x31-39 / %x41-43 / %x61-63
h-x1e = %x31-39 / %x41-45 / %x61-65
h-x1f = %x31-39 / %x41-46 / %x61-66
h-x2f = %x32-39 / %x41-46 / %x61-66
h-x3f = %x33-39 / %x41-46 / %x61-66
h-x8b = %x38-39 / %x41-42 / %x61-62
h-xcf = %x43-46 / %x63-66
h-xef = %x45-46 / %x65-66In the main EDN grammar, you can amend app-string to
app-string = sq-h-app-string-h / app-prefix sqstr
...in the main ABNF to make use of this.
If you don't want to do your own line-breaking, but your screen is narrow (or you want to include the result in an RFC), you might want to try:
abnfrob -a --squash=h cbor-edn-h.abnf
(-a == --asr33) and you get some (not so great) automatic linebreaking.