treerack/format_test.go

package treerack

import (
	"bytes"
	"fmt"
	"testing"
)

type formatDefinitionTestItem struct {
	title      string
	definition string
	syntax     string
	output     string
}

func testDefinitionFormatItem(t *testing.T, treerack *Syntax, f formatFlags, test formatDefinitionTestItem) func(t *testing.T) {
	return func(t *testing.T) {
		syntax := test.syntax
		if test.definition != "" {
			syntax = fmt.Sprintf("def = %s", test.definition)
		}

		nodes, err := treerack.Parse(bytes.NewBufferString(syntax))
		if err != nil {
			t.Fatal(err)
		}

		s := &Syntax{}
		if err := define(s, nodes); err != nil {
			t.Fatal(err)
		}

		def, ok := s.registry.definition["def"]
		if !ok {
			t.Fatal("failed to register definition")
		}

		output := def.format(s.registry, formatOptions{mode: f})
		if output != test.output {
			t.Error("invalid definition format")
			t.Log("got:     ", output)
			t.Log("expected:", test.output)
		}
	}
}

func testDefinitionFormat(t *testing.T, f formatFlags, tests []formatDefinitionTestItem) {
	treerack, err := bootSyntax()
	if err != nil {
		t.Fatal(err)
	}

	for _, test := range tests {
		t.Run(test.title, testDefinitionFormatItem(t, treerack, f, test))
	}
}

func TestCharFormat(t *testing.T) {
	testDefinitionFormat(t, formatNone, []formatDefinitionTestItem{{
		title:      "empty",
		definition: "[]",
		output:     "[]",
	}, {
		title:      "one char",
		definition: "[a]",
		output:     `"a"`,
	}, {
		title:      "escaped char",
		definition: "[\\a]",
		output:     `"a"`,
	}, {
		title:      "escaped control char",
		definition: "[\\^]",
		output:     `"^"`,
	}, {
		title:      "escaped whitespace char",
		definition: "[\\n]",
		output:     `"\n"`,
	}, {
		title:      "escaped verbatim whitespace char",
		definition: "[\n]",
		output:     `"\n"`,
	}, {
		title:      "escaped range",
		definition: "[\\b-\\v]",
		output:     "[\\b-\\v]",
	}, {
		title:      "anything",
		definition: ".",
		output:     ".",
	}, {
		title:      "not something",
		definition: "[^abc]",
		output:     "[^abc]",
	}, {
		title:      "range",
		definition: "[a-z]",
		output:     "[a-z]",
	}, {
		title:      "range and char mixed",
		definition: "[a-z_\\-A-Z]",
		output:     "[_\\-a-zA-Z]",
	}})
}

func TestSequenceFormat(t *testing.T) {
	testDefinitionFormat(t, formatNone, []formatDefinitionTestItem{{
		title:  "empty char sequence",
		syntax: `def = ""`,
		output: `""`,
	}, {
		title:  "char sequence",
		syntax: `def = "abc"`,
		output: `"abc"`,
	}, {
		title:  "char sequence, escaped",
		syntax: `def = "\\n"`,
		output: `"\\n"`,
	}, {
		title:  "chars",
		syntax: `def = "abc" [a-z]`,
		output: `"abc" [a-z]`,
	}, {
		title:  "quantifiers, 0-or-more, single char",
		syntax: `def = "a"*`,
		output: `"a"*`,
	}, {
		title:  "quantifiers, 0-or-more",
		syntax: `def = "abc"*`,
		output: `"abc"*`,
	}, {
		title:  "quantifiers, 1-or-more, single char",
		syntax: `def = "a"+`,
		output: `"a"+`,
	}, {
		title:  "quantifiers, 1-or-more",
		syntax: `def = "abc"+`,
		output: `"abc"+`,
	}, {
		title:  "quantifiers, 0-or-one, single char",
		syntax: `def = "a"?`,
		output: `"a"?`,
	}, {
		title:  "quantifiers, 0-or-one",
		syntax: `def = "abc"?`,
		output: `"abc"?`,
	}, {
		title:  "quantifiers, exact number, single char",
		syntax: `def = "a"{3}`,
		output: `"a"{3}`,
	}, {
		title:  "quantifiers, exact number",
		syntax: `def = "abc"{3}`,
		output: `"abc"{3}`,
	}, {
		title:  "quantifiers, max, single char",
		syntax: `def = "a"{0, 3}`,
		output: `"a"{,3}`,
	}, {
		title:  "quantifiers, max",
		syntax: `def = "abc"{0, 3}`,
		output: `"abc"{,3}`,
	}, {
		title:  "quantifiers, min, single char",
		syntax: `def = "a"{3,}`,
		output: `"a"{3,}`,
	}, {
		title:  "quantifiers, min",
		syntax: `def = "abc"{3,}`,
		output: `"abc"{3,}`,
	}, {
		title:  "quantifiers, range, single char",
		syntax: `def = "a"{3, 9}`,
		output: `"a"{3,9}`,
	}, {
		title:  "quantifiers, range",
		syntax: `def = "abc"{3, 9}`,
		output: `"abc"{3,9}`,
	}, {
		title:  "symbols",
		syntax: `a = "a"; b = "b"; c = "c"; def = a b c`,
		output: "a b c",
	}, {
		title:  "choice in sequence, single char",
		syntax: `def = "a" ("b" | "c")`,
		output: `"a" ("b" | "c")`,
	}, {
		title:  "choice in sequence",
		syntax: `def = "abc" ("def" | "ghi")`,
		output: `"abc" ("def" | "ghi")`,
	}, {
		title:  "grouped quantifier, single char",
		syntax: `def = ("a" "b"){3}`,
		output: `("a" "b"){3}`,
	}, {
		title:  "grouped quantifier",
		syntax: `def = ("abc" "def"){3}`,
		output: `("abc" "def"){3}`,
	}})
}

func TestChoiceFormat(t *testing.T) {
	testDefinitionFormat(t, formatNone, []formatDefinitionTestItem{{
		title:  "choice of char sequences, single char",
		syntax: `def = "a" | "b" | "c"`,
		output: `"a" | "b" | "c"`,
	}, {
		title:  "choice of char sequences",
		syntax: `def = "abc" | "def" | "ghi"`,
		output: `"abc" | "def" | "ghi"`,
	}, {
		title:  "choice of inline sequences, single char",
		syntax: `def = "a" "b" | "c" "d" | "e" "f"`,
		output: `"a" "b" | "c" "d" | "e" "f"`,
	}, {
		title:  "choice of inline sequences",
		syntax: `def = "abc" "def" | "ghi" "jkl" | "mno" "pqr"`,
		output: `"abc" "def" | "ghi" "jkl" | "mno" "pqr"`,
	}, {
		title:  "choice of symbol",
		syntax: `a = "a"; b = "b"; c = "c"; def = a | b | c`,
		output: "a | b | c",
	}})
}

const testDoc = `/*
  foo
*/
// bar  
	        // bar
  //baz
     /* foo
  bar baz */// foo bar baz
wschar:alias  =// foo 
  /* bar */ " " | "\t" | "\n" | "\b" | "\f" | "\r" | "\v";
wsc:ws        = wschar | comment;

block-comment:alias:nows /* foo */ // bar
= "/*" ("*" [^/] | [^*])* "*/";
line-comment:alias:nows                     /*
foo
*/ = "//" [^\n]*;
comment-segment:alias:nows = // bar
line-comment | block-comment;
ws-no-nl:alias:nows        = " " | "\t" | "\b" | /* this one */ /* is a */ "\f" /* form feed */ // for sure
| "\r" | "\v";
comment:nows               = comment-segment /* segment is not the best name */ /* but */ (ws-no-nl* "\n"?  ws-no-nl* // fine
comment-segment)*;

any-char = "."; // equivalent to [^]

// caution: newline is accepted
/* class not */ class-not       = "^";
class-char:nows = [^\\\[\]\^\-] | "\\" . /* foo
bar */;
char-range:nows = class-char "-" class-char // foo
;
char-class:nows = "[" class-not? (class-char | char-range)* "]"; // foo
/* bar
baz */

// newline is accepted
sequence-char:nows = [^\\"] | "\\" .;
char-sequence:nows = "\"" sequence-char* "\"";

terminal:alias = any-char | char-class | char-sequence;

symbol:nows = [^\\ \n\t\b\f\r\v/.\[\]\"{}\^+*?|():=;]+;

group:alias = "(" expression ")";

number:alias:nows = [0-9]+;
count             = number;
count-quantifier  = "{" count "}";
range-from        = number;
range-to          = number;
range-quantifier  = "{" range-from? "," range-to? "}";
one-or-more       = "+";
zero-or-more      = "*";
zero-or-one       = "?";
quantity:alias    = count-quantifier
                  | range-quantifier
                  | one-or-more
                  | zero-or-more
                  | zero-or-one;

item:nows = (terminal | symbol | group) quantity?;
sequence  = item+;

option:alias = terminal | symbol | group | sequence;

// DOC: how the order matters
choice = option ("|" option)+;

// DOC: not having 'not' needs some tricks sometimes

expression:alias = terminal | symbol | group | sequence | choice;

alias        = "alias";
ws                         = "ws";
nows               = "nows";
kw                         = "kw";
nokw                       = "nokw";
failpass             = "failpass";
root                       = "root";
flag:alias           = alias | ws | nows | kw | nokw | failpass | root;
definition-name:alias:nows = symbol (":" flag)*;
definition             = definition-name "=" expression;

definitions:alias = definition (";"+ definition)*;
syntax:root       = ";"* definitions? ";"*;
`

const testDocCheck = `/*
  foo
*/
// bar
// bar
// baz
/* foo
  bar baz */ // foo bar baz
wschar:alias               = // foo
                             /* bar */
                             " " | "\t" | "\n" | "\b" | "\f" | "\r" | "\v";
wsc:ws                     = wschar | comment;
block-comment:alias:nows   = /* foo */ // bar
                             "/*" ("*" [^/] | [^*])* "*/";
line-comment:alias:nows    = /*
                             foo
                             */
                             "//" [^\n]*;
comment-segment:alias:nows = // bar
                             line-comment | block-comment;
ws-no-nl:alias:nows        = " "
                           | "\t"
                           | "\b"
                           /* this one */ /* is a */
                           | "\f"
                           /* form feed */ // for sure
                           | "\r"
                           | "\v";
comment:nows               = comment-segment
                             /* segment is not the best name */ /* but */
                             ( ws-no-nl*
                               "\n"?
                               ws-no-nl*
                               // fine
                               comment-segment
                             )*;
any-char                   = "."; // equivalent to [^]

// caution: newline is accepted
/* class not */
class-not       = "^";
class-char:nows = [^\\\[\]\^\-] | "\\" .; /* foo
                  bar */
char-range:nows = class-char "-" class-char; // foo
char-class:nows = "[" class-not? (class-char | char-range)* "]"; // foo
                  /* bar
                  baz */

// newline is accepted
sequence-char:nows = [^\\"] | "\\" .;
char-sequence:nows = "\"" sequence-char* "\"";
terminal:alias     = any-char | char-class | char-sequence;
symbol:nows        = [^\\ \n\t\b\f\r\v/.\[\]\"{}\^+*?|():=;]+;
group:alias        = "(" expression ")";
number:alias:nows  = [0-9]+;
count              = number;
count-quantifier   = "{" count "}";
range-from         = number;
range-to           = number;
range-quantifier   = "{" range-from? "," range-to? "}";
one-or-more        = "+";
zero-or-more       = "*";
zero-or-one        = "?";
quantity:alias     = count-quantifier | range-quantifier | one-or-more | zero-or-more | zero-or-one;
item:nows          = (terminal | symbol | group) quantity?;
sequence           = item+;
option:alias       = terminal | symbol | group | sequence;

// DOC: how the order matters
choice = option ("|" option)+;

// DOC: not having 'not' needs some tricks sometimes

expression:alias           = terminal | symbol | group | sequence | choice;
alias                      = "alias";
ws                         = "ws";
nows                       = "nows";
kw                         = "kw";
nokw                       = "nokw";
failpass                   = "failpass";
root                       = "root";
flag:alias                 = alias | ws | nows | kw | nokw | failpass | root;
definition-name:alias:nows = symbol (":" flag)*;
definition                 = definition-name "=" expression;
definitions:alias          = definition (";"+ definition)*;
syntax:root                = ";"* definitions? ";"*;
`

const testDocURL = `// basd on RFC3986 and RFC6874

// char types:
digit:alias:failpass           = [0-9];
hex:alias:failpass             = [0-9a-fA-F];
alpha:alias:failpass           = [a-zA-Z];
delimiter:alias:failpass       = ":" | "/" | "?" | "#" | "[" | "]" | "@";
subdelimiter:alias:failpass    = "!" | "$" | "&" | "'" | "(" | ")" | "*" | "+" | "," | ";" | "=";
unreserved:alias:failpass      = alpha | digit | "-" | "." | "_" | "~";
reserved:alias:failpass        = delimiter | subdelimiter;
percent-encoded:alias:failpass = "%" hex{2};
path-char:alias:failpass       = unreserved | percent-encoded | subdelimiter | ":" | "@";

// scheme:
scheme = alpha (alpha | digit | [+-.])*;

// userinfo:
userinfo = (unreserved | percent-encoded | subdelimiter | ":")*;

// IPv4:
dec-byte:alias:failpass = digit | [1-9] digit | "1" digit{2} | "2" [0-4] digit | "25" [0-5];
ipv4:failpass     = dec-byte ("." dec-byte){3};

// IPv6:
h16:alias:failpass  = hex{1,4};
ls32:alias:failpass = (h16 ":" h16) | ipv4;
ipv6:failpass =                           (h16 ":"){6} ls32
     |                      "::" (h16 ":"){5} ls32
     |                 h16? "::" (h16 ":"){4} ls32
     | ((h16 ":")?    h16)? "::" (h16 ":"){3} ls32
     | ((h16 ":"){,2} h16)? "::" (h16 ":"){2} ls32
     | ((h16 ":"){,3} h16)? "::"  h16 ":"     ls32
     | ((h16 ":"){,4} h16)? "::"              ls32
     | ((h16 ":"){,5} h16)? "::"              h16 
     | ((h16 ":"){,6} h16)? "::"                  ;
zone-id:alias:failpass = (unreserved | percent-encoded)+;
ipv6-zone:failpass = ipv6 "%25" zone-id; // RFC6874

// host:
registry-name-rfc:failpass     = (unreserved | percent-encoded | subdelimiter)*; // all RFC chars allowed
dns-label:alias:failpass       = (alpha | digit) ("-"* (alpha | digit)+)*;
domain-name:failpass = dns-label ("." dns-label)* "."?; // DNS compatible
hostname-rfc        = ipv4 | "[" (ipv6 | ipv6-zone) "]" | registry-name-rfc;
hostname    = ipv4 | "[" (ipv6 | ipv6-zone) "]" | domain-name;
port            = digit*;
host-rfc:alias:failpass            = hostname-rfc (":" port)?;
host:alias:failpass        = hostname (":" port)?;

// path:
segment:alias:failpass                   = path-char*;
segment-non-zero:alias:failpass          = path-char+;
segment-non-zero-no-colon:alias:failpass = (unreserved | percent-encoded | subdelimiter | "@")+;
path-absolute-or-empty    = ("/" segment)*;
path-absolute             = "/" (segment-non-zero ("/" segment)*)?;
path-rootless             = segment-non-zero ("/" segment)*;
path-noscheme             = segment-non-zero-no-colon ("/" segment)*;
path-empty                = "";
query                     = (path-char | "/" | "?")*;
fragment                  = (path-char | "/" | "?")*;

// composed together:
authority-rfc:alias:failpass          = (userinfo "@")? host-rfc;
authority:alias:failpass      = (userinfo "@")? host;
hierarchy-part-rfc:alias:failpass     = "//" authority-rfc path-absolute-or-empty | path-absolute | path-rootless | path-empty;
hierarchy-part:alias:failpass = "//" authority path-absolute-or-empty | path-absolute | path-rootless | path-empty;
relative-part-rfc:alias:failpass      = "//" authority-rfc path-absolute-or-empty | path-absolute | path-noscheme | path-empty;
relative-part:alias:failpass  = "//" authority path-absolute-or-empty | path-absolute | path-noscheme | path-empty;
absolute-url-rfc       = scheme ":" hierarchy-part-rfc ("?" query)? ("#" fragment)?;
absolute-url   = scheme ":" hierarchy-part ("?" query)? ("#" fragment)?;
relative-url-rfc       = relative-part-rfc ("?" query)? ("#" fragment)?;
relative-url   = relative-part ("?" query)? ("#" fragment)?;

// supporting four possible types URLs:
url:root = absolute-url | absolute-url-rfc | relative-url | relative-url-rfc;
`

const testDocCheckURL = `// basd on RFC3986 and RFC6874

// char types:
digit:alias:failpass           = [0-9];
hex:alias:failpass             = [0-9a-fA-F];
alpha:alias:failpass           = [a-zA-Z];
delimiter:alias:failpass       = ":" | "/" | "?" | "#" | "[" | "]" | "@";
subdelimiter:alias:failpass    = "!" | "$" | "&" | "'" | "(" | ")" | "*" | "+" | "," | ";" | "=";
unreserved:alias:failpass      = alpha | digit | "-" | "." | "_" | "~";
reserved:alias:failpass        = delimiter | subdelimiter;
percent-encoded:alias:failpass = "%" hex{2};
path-char:alias:failpass       = unreserved | percent-encoded | subdelimiter | ":" | "@";

// scheme:
scheme = alpha (alpha | digit | [+-.])*;

// userinfo:
userinfo = (unreserved | percent-encoded | subdelimiter | ":")*;

// IPv4:
dec-byte:alias:failpass = digit | [1-9] digit | "1" digit{2} | "2" [0-4] digit | "25" [0-5];
ipv4:failpass           = dec-byte ("." dec-byte){3};

// IPv6:
h16:alias:failpass     = hex{1,4};
ls32:alias:failpass    = h16 ":" h16 | ipv4;
ipv6:failpass          = (h16 ":"){6} ls32
                       | "::" (h16 ":"){5} ls32
                       | h16? "::" (h16 ":"){4} ls32
                       | ((h16 ":")? h16)? "::" (h16 ":"){3} ls32
                       | ((h16 ":"){,2} h16)? "::" (h16 ":"){2} ls32
                       | ((h16 ":"){,3} h16)? "::" h16 ":" ls32
                       | ((h16 ":"){,4} h16)? "::" ls32
                       | ((h16 ":"){,5} h16)? "::" h16
                       | ((h16 ":"){,6} h16)? "::";
zone-id:alias:failpass = (unreserved | percent-encoded)+;
ipv6-zone:failpass     = ipv6 "%25" zone-id; // RFC6874

// host:
registry-name-rfc:failpass = (unreserved | percent-encoded | subdelimiter)*; // all RFC chars allowed
dns-label:alias:failpass   = (alpha | digit) ("-"* (alpha | digit)+)*;
domain-name:failpass       = dns-label ("." dns-label)* "."?; // DNS compatible
hostname-rfc               = ipv4 | "[" (ipv6 | ipv6-zone) "]" | registry-name-rfc;
hostname                   = ipv4 | "[" (ipv6 | ipv6-zone) "]" | domain-name;
port                       = digit*;
host-rfc:alias:failpass    = hostname-rfc (":" port)?;
host:alias:failpass        = hostname (":" port)?;

// path:
segment:alias:failpass                   = path-char*;
segment-non-zero:alias:failpass          = path-char+;
segment-non-zero-no-colon:alias:failpass = (unreserved | percent-encoded | subdelimiter | "@")+;
path-absolute-or-empty                   = ("/" segment)*;
path-absolute                            = "/" (segment-non-zero ("/" segment)*)?;
path-rootless                            = segment-non-zero ("/" segment)*;
path-noscheme                            = segment-non-zero-no-colon ("/" segment)*;
path-empty                               = "";
query                                    = (path-char | "/" | "?")*;
fragment                                 = (path-char | "/" | "?")*;

// composed together:
authority-rfc:alias:failpass      = (userinfo "@")? host-rfc;
authority:alias:failpass          = (userinfo "@")? host;
hierarchy-part-rfc:alias:failpass = "//" authority-rfc path-absolute-or-empty
                                  | path-absolute
                                  | path-rootless
                                  | path-empty;
hierarchy-part:alias:failpass     = "//" authority path-absolute-or-empty
                                  | path-absolute
                                  | path-rootless
                                  | path-empty;
relative-part-rfc:alias:failpass  = "//" authority-rfc path-absolute-or-empty
                                  | path-absolute
                                  | path-noscheme
                                  | path-empty;
relative-part:alias:failpass      = "//" authority path-absolute-or-empty
                                  | path-absolute
                                  | path-noscheme
                                  | path-empty;
absolute-url-rfc                  = scheme ":" hierarchy-part-rfc ("?" query)? ("#" fragment)?;
absolute-url                      = scheme ":" hierarchy-part ("?" query)? ("#" fragment)?;
relative-url-rfc                  = relative-part-rfc ("?" query)? ("#" fragment)?;
relative-url                      = relative-part ("?" query)? ("#" fragment)?;

// supporting four possible types URLs:
url:root = absolute-url | absolute-url-rfc | relative-url | relative-url-rfc;
`

func TestDocFormat(t *testing.T) {
	for _, test := range []struct{ title, in, out string }{{
		title: "format",
		in:    testDoc,
		out:   testDocCheck,
	}, {
		title: "check",
		in:    testDocCheck,
		out:   testDocCheck,
	}, {
		title: "format url",
		in:    testDocURL,
		out:   testDocCheckURL,
	}, {
		title: "check url",
		in:    testDocCheckURL,
		out:   testDocCheckURL,
	}} {
		t.Run(test.title, func(t *testing.T) {
			in := bytes.NewBufferString(test.in)
			s := &Syntax{}
			if err := s.ReadSyntax(in); err != nil {
				t.Fatal(err)
			}

			out := bytes.NewBuffer(nil)
			if err := s.Format(out); err != nil {
				t.Fatal(err)
			}

			if out.String() != test.out {
				t.Log(test.out)
				t.Log(out.String())
				t.Fatal()
			}
		})
	}
}