wikitext 4.0.1 → 4.0.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/bin/wikitext +10 -110
- data/ext/wikitext/ary.c +116 -0
- data/ext/wikitext/ary.h +50 -0
- data/ext/wikitext/depend +32 -0
- data/ext/wikitext/parser.c +2595 -0
- data/ext/wikitext/parser.h +40 -0
- data/ext/wikitext/ruby_compat.h +34 -0
- data/ext/wikitext/str.c +109 -0
- data/ext/wikitext/str.h +64 -0
- data/ext/wikitext/token.c +125 -0
- data/ext/wikitext/token.h +117 -0
- data/ext/wikitext/wikitext.c +125 -0
- data/ext/wikitext/wikitext.h +39 -0
- data/ext/wikitext/wikitext_ragel.c +3211 -0
- data/ext/wikitext/wikitext_ragel.h +26 -0
- data/lib/wikitext/version.rb +1 -1
- metadata +17 -3
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: c61ab6467120d8def1be560fcb4604ee7c90454b
|
4
|
+
data.tar.gz: 1c7be94a73a1d038a18744775054910772a964f0
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: df484b7d09e76c9b01a53cf4c8cb7b7bcea8fba56d6e7d1443e01259b3b41934e4f517a87cf0a4593ac33646c47a9b9b4c96accf4f810db83013d32a248dd3f7
|
7
|
+
data.tar.gz: 66d3a468c2d787a2ec572b24ff53e8210cc520f37fd4caab68fc3536c0640f01deaddbe59e22f892eea8f2872788ef2f50bc66bba363298c1e5cee7c655cd444
|
data/bin/wikitext
CHANGED
@@ -1,116 +1,16 @@
|
|
1
1
|
#!/usr/bin/env ruby
|
2
|
-
# Copyright 2008-2013 Wincent Colaiuta. All rights reserved.
|
3
2
|
#
|
4
|
-
#
|
5
|
-
#
|
3
|
+
# This file was generated by Bundler.
|
4
|
+
#
|
5
|
+
# The application 'wikitext' is installed as part of a gem, and
|
6
|
+
# this file is here to facilitate running it.
|
6
7
|
#
|
7
|
-
# 1. Redistributions of source code must retain the above copyright notice,
|
8
|
-
# this list of conditions and the following disclaimer.
|
9
|
-
# 2. Redistributions in binary form must reproduce the above copyright notice,
|
10
|
-
# this list of conditions and the following disclaimer in the documentation
|
11
|
-
# and/or other materials provided with the distribution.
|
12
|
-
|
13
|
-
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
14
|
-
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
15
|
-
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
16
|
-
# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
|
17
|
-
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
|
18
|
-
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
|
19
|
-
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
20
|
-
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
21
|
-
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
22
|
-
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
23
|
-
# POSSIBILITY OF SUCH DAMAGE.
|
24
|
-
|
25
|
-
require 'wikitext'
|
26
|
-
require 'wikitext/version'
|
27
|
-
|
28
|
-
module Wikitext
|
29
|
-
module Tool
|
30
|
-
# Simple substitute for the HighLine library if it is not available.
|
31
|
-
class FakeHighLine
|
32
|
-
def color(str, _)
|
33
|
-
str
|
34
|
-
end
|
35
|
-
|
36
|
-
def output_cols
|
37
|
-
80
|
38
|
-
end
|
39
|
-
end
|
40
|
-
|
41
|
-
INPUT_FILES = []
|
42
|
-
|
43
|
-
def self.interactive?
|
44
|
-
STDOUT.tty? && STDIN.tty? && INPUT_FILES.empty?
|
45
|
-
end
|
46
|
-
|
47
|
-
def self.pretty_print tokens
|
48
|
-
tokens.each do |token|
|
49
|
-
puts <<-END
|
50
|
-
Token: type: #{token.token_type}
|
51
|
-
line: #{token.line_start}..#{token.line_stop} column: #{token.column_start}..#{token.column_stop}
|
52
|
-
pointer: #{token.start}..#{token.stop}
|
53
|
-
code_point: #{token.code_point}
|
54
|
-
string_value: #{token.string_value.inspect}
|
55
|
-
|
56
|
-
END
|
57
|
-
end
|
58
|
-
end
|
59
8
|
|
60
|
-
|
61
|
-
|
62
|
-
|
63
|
-
method = :tokenize
|
64
|
-
else
|
65
|
-
INPUT_FILES << arg
|
66
|
-
end
|
67
|
-
end
|
9
|
+
require 'pathname'
|
10
|
+
ENV['BUNDLE_GEMFILE'] ||= File.expand_path("../../Gemfile",
|
11
|
+
Pathname.new(__FILE__).realpath)
|
68
12
|
|
69
|
-
|
70
|
-
|
71
|
-
require 'highline'
|
72
|
-
rescue LoadError
|
73
|
-
begin
|
74
|
-
require 'rubygems'
|
75
|
-
require 'highline'
|
76
|
-
rescue LoadError
|
77
|
-
end
|
78
|
-
end
|
79
|
-
puts "wikitext #{Wikitext::VERSION}"
|
80
|
-
highline = (defined?(HighLine) ? HighLine : FakeHighLine).new
|
81
|
-
end
|
13
|
+
require 'rubygems'
|
14
|
+
require 'bundler/setup'
|
82
15
|
|
83
|
-
|
84
|
-
if INPUT_FILES.empty?
|
85
|
-
begin
|
86
|
-
while true
|
87
|
-
puts highline.color('(Ctrl+D to process, Ctrl+C to exit)>>', :bold) if interactive?
|
88
|
-
input = STDIN.read
|
89
|
-
puts '-' * highline.output_cols if interactive?
|
90
|
-
if method == :tokenize
|
91
|
-
pretty_print parser.tokenize(input)
|
92
|
-
else
|
93
|
-
puts parser.parse(input)
|
94
|
-
end
|
95
|
-
puts '-' * highline.output_cols if interactive?
|
96
|
-
exit unless interactive?
|
97
|
-
end
|
98
|
-
rescue Interrupt
|
99
|
-
end
|
100
|
-
else # we have INPUT_FILES
|
101
|
-
exit_status = 0
|
102
|
-
INPUT_FILES.each do |file|
|
103
|
-
begin
|
104
|
-
puts parser.parse(File.new(file).read)
|
105
|
-
rescue Errno::ENOENT
|
106
|
-
STDERR.puts "error: no such file or directory: #{file}"
|
107
|
-
exit_status |= 1
|
108
|
-
rescue Errno::EACCES
|
109
|
-
STDERR.puts "error: permission denied: #{file}"
|
110
|
-
exit_status |= 2
|
111
|
-
end
|
112
|
-
end
|
113
|
-
exit exit_status
|
114
|
-
end
|
115
|
-
end # module Tool
|
116
|
-
end # module Wikitext
|
16
|
+
load Gem.bin_path('wikitext', 'wikitext')
|
data/ext/wikitext/ary.c
ADDED
@@ -0,0 +1,116 @@
|
|
1
|
+
// Copyright 2008-2009 Wincent Colaiuta. All rights reserved.
|
2
|
+
//
|
3
|
+
// Redistribution and use in source and binary forms, with or without
|
4
|
+
// modification, are permitted provided that the following conditions are met:
|
5
|
+
//
|
6
|
+
// 1. Redistributions of source code must retain the above copyright notice,
|
7
|
+
// this list of conditions and the following disclaimer.
|
8
|
+
// 2. Redistributions in binary form must reproduce the above copyright notice,
|
9
|
+
// this list of conditions and the following disclaimer in the documentation
|
10
|
+
// and/or other materials provided with the distribution.
|
11
|
+
//
|
12
|
+
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
13
|
+
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
14
|
+
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
15
|
+
// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
|
16
|
+
// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
|
17
|
+
// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
|
18
|
+
// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
19
|
+
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
20
|
+
// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
21
|
+
// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
22
|
+
// POSSIBILITY OF SUCH DAMAGE.
|
23
|
+
|
24
|
+
#include "ary.h"
|
25
|
+
|
26
|
+
ary_t *ary_new(void)
|
27
|
+
{
|
28
|
+
ary_t *ary = ALLOC_N(ary_t, 1);
|
29
|
+
ary->count = 0;
|
30
|
+
ary->max = DEFAULT_ENTRY_COUNT;
|
31
|
+
ary->entries = ALLOC_N(int, DEFAULT_ENTRY_COUNT);
|
32
|
+
return ary;
|
33
|
+
}
|
34
|
+
|
35
|
+
int ary_entry(ary_t *ary, int idx)
|
36
|
+
{
|
37
|
+
if (idx < 0)
|
38
|
+
idx = ary->count + idx;
|
39
|
+
return (idx >= 0 && ary->count > idx) ? ary->entries[idx] : INT_MAX;
|
40
|
+
}
|
41
|
+
|
42
|
+
void ary_clear(ary_t *ary)
|
43
|
+
{
|
44
|
+
ary->count = 0;
|
45
|
+
}
|
46
|
+
|
47
|
+
int ary_pop(ary_t *ary)
|
48
|
+
{
|
49
|
+
if (ary->count > 0)
|
50
|
+
{
|
51
|
+
ary->count--;
|
52
|
+
return 1;
|
53
|
+
}
|
54
|
+
return 0;
|
55
|
+
}
|
56
|
+
|
57
|
+
void ary_push(ary_t *ary, int val)
|
58
|
+
{
|
59
|
+
if (ary->count == ary->max)
|
60
|
+
{
|
61
|
+
ary->max += DEFAULT_ENTRY_COUNT;
|
62
|
+
REALLOC_N(ary->entries, int, ary->max);
|
63
|
+
}
|
64
|
+
ary->entries[ary->count] = val;
|
65
|
+
ary->count++;
|
66
|
+
}
|
67
|
+
|
68
|
+
int ary_includes(ary_t *ary, int val)
|
69
|
+
{
|
70
|
+
for (int i = 0, max = ary->count; i < max; i++)
|
71
|
+
{
|
72
|
+
if (ary->entries[i] == val)
|
73
|
+
return 1;
|
74
|
+
}
|
75
|
+
return 0;
|
76
|
+
}
|
77
|
+
|
78
|
+
int ary_includes2(ary_t *ary, int val1, int val2)
|
79
|
+
{
|
80
|
+
for (int i = 0, max = ary->count; i < max; i++)
|
81
|
+
{
|
82
|
+
if (ary->entries[i] == val1 ||
|
83
|
+
ary->entries[i] == val2)
|
84
|
+
return 1;
|
85
|
+
}
|
86
|
+
return 0;
|
87
|
+
}
|
88
|
+
|
89
|
+
int ary_includes3(ary_t *ary, int val1, int val2, int val3)
|
90
|
+
{
|
91
|
+
for (int i = 0, max = ary->count; i < max; i++)
|
92
|
+
{
|
93
|
+
if (ary->entries[i] == val1 ||
|
94
|
+
ary->entries[i] == val2 ||
|
95
|
+
ary->entries[i] == val3)
|
96
|
+
return 1;
|
97
|
+
}
|
98
|
+
return 0;
|
99
|
+
}
|
100
|
+
|
101
|
+
int ary_count(ary_t *ary, int item)
|
102
|
+
{
|
103
|
+
int count = 0;
|
104
|
+
for (int i = 0, max = ary->count; i < max; i++)
|
105
|
+
{
|
106
|
+
if (ary->entries[i] == item)
|
107
|
+
count++;
|
108
|
+
}
|
109
|
+
return count;
|
110
|
+
}
|
111
|
+
|
112
|
+
void ary_free(ary_t *ary)
|
113
|
+
{
|
114
|
+
free(ary->entries);
|
115
|
+
free(ary);
|
116
|
+
}
|
data/ext/wikitext/ary.h
ADDED
@@ -0,0 +1,50 @@
|
|
1
|
+
// Copyright 2008-2009 Wincent Colaiuta. All rights reserved.
|
2
|
+
//
|
3
|
+
// Redistribution and use in source and binary forms, with or without
|
4
|
+
// modification, are permitted provided that the following conditions are met:
|
5
|
+
//
|
6
|
+
// 1. Redistributions of source code must retain the above copyright notice,
|
7
|
+
// this list of conditions and the following disclaimer.
|
8
|
+
// 2. Redistributions in binary form must reproduce the above copyright notice,
|
9
|
+
// this list of conditions and the following disclaimer in the documentation
|
10
|
+
// and/or other materials provided with the distribution.
|
11
|
+
//
|
12
|
+
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
13
|
+
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
14
|
+
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
15
|
+
// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
|
16
|
+
// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
|
17
|
+
// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
|
18
|
+
// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
19
|
+
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
20
|
+
// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
21
|
+
// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
22
|
+
// POSSIBILITY OF SUCH DAMAGE.
|
23
|
+
|
24
|
+
#include "ruby_compat.h"
|
25
|
+
|
26
|
+
typedef struct
|
27
|
+
{
|
28
|
+
int count;
|
29
|
+
int max;
|
30
|
+
int *entries;
|
31
|
+
} ary_t;
|
32
|
+
|
33
|
+
// in the test suite array count goes no higher than 25 or 26
|
34
|
+
#define DEFAULT_ENTRY_COUNT 64
|
35
|
+
|
36
|
+
#define NO_ITEM(item) (item == INT_MAX)
|
37
|
+
|
38
|
+
ary_t *ary_new(void);
|
39
|
+
int ary_entry(ary_t *ary, int idx);
|
40
|
+
void ary_clear(ary_t *ary);
|
41
|
+
int ary_pop(ary_t *ary);
|
42
|
+
void ary_push(ary_t *ary, int val);
|
43
|
+
int ary_includes(ary_t *ary, int val);
|
44
|
+
int ary_includes2(ary_t *ary, int val1, int val2);
|
45
|
+
int ary_includes3(ary_t *ary, int val1, int val2, int val3);
|
46
|
+
|
47
|
+
// returns a count indicating the number of times the value appears in the collection
|
48
|
+
int ary_count(ary_t *ary, int item);
|
49
|
+
|
50
|
+
void ary_free(ary_t *ary);
|
data/ext/wikitext/depend
ADDED
@@ -0,0 +1,32 @@
|
|
1
|
+
# Copyright 2008-2010 Wincent Colaiuta. All rights reserved.
|
2
|
+
#
|
3
|
+
# Redistribution and use in source and binary forms, with or without
|
4
|
+
# modification, are permitted provided that the following conditions are met:
|
5
|
+
#
|
6
|
+
# 1. Redistributions of source code must retain the above copyright notice,
|
7
|
+
# this list of conditions and the following disclaimer.
|
8
|
+
# 2. Redistributions in binary form must reproduce the above copyright notice,
|
9
|
+
# this list of conditions and the following disclaimer in the documentation
|
10
|
+
# and/or other materials provided with the distribution.
|
11
|
+
#
|
12
|
+
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
13
|
+
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
14
|
+
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
15
|
+
# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
|
16
|
+
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
|
17
|
+
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
|
18
|
+
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
19
|
+
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
20
|
+
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
21
|
+
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
22
|
+
# POSSIBILITY OF SUCH DAMAGE.
|
23
|
+
|
24
|
+
# don't warn about unused params because many Ruby methods accept "self" but don't use it
|
25
|
+
CFLAGS += -std=gnu99 -Wall -Wextra -Wno-unused-parameter
|
26
|
+
|
27
|
+
ary.o : ary.c ary.h ruby_compat.h
|
28
|
+
parser.o : ary.c ary.h parser.c parser.h ruby_compat.h str.c str.h token.h wikitext.h wikitext_ragel.h
|
29
|
+
str.o : ruby_compat.h str.c str.h
|
30
|
+
token.o : ruby_compat.h token.c token.h wikitext.h
|
31
|
+
wikitext.o : parser.h ruby_compat.h token.h wikitext.c wikitext.h wikitext_ragel.h
|
32
|
+
wikitext_ragel.o : ruby_compat.h token.h wikitext.h wikitext_ragel.h wikitext_ragel.c
|
@@ -0,0 +1,2595 @@
|
|
1
|
+
// Copyright 2007-2013 Wincent Colaiuta. All rights reserved.
|
2
|
+
//
|
3
|
+
// Redistribution and use in source and binary forms, with or without
|
4
|
+
// modification, are permitted provided that the following conditions are met:
|
5
|
+
//
|
6
|
+
// 1. Redistributions of source code must retain the above copyright notice,
|
7
|
+
// this list of conditions and the following disclaimer.
|
8
|
+
// 2. Redistributions in binary form must reproduce the above copyright notice,
|
9
|
+
// this list of conditions and the following disclaimer in the documentation
|
10
|
+
// and/or other materials provided with the distribution.
|
11
|
+
//
|
12
|
+
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
13
|
+
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
14
|
+
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
15
|
+
// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
|
16
|
+
// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
|
17
|
+
// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
|
18
|
+
// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
19
|
+
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
20
|
+
// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
21
|
+
// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
22
|
+
// POSSIBILITY OF SUCH DAMAGE.
|
23
|
+
|
24
|
+
#include <stdbool.h>
|
25
|
+
|
26
|
+
#include "parser.h"
|
27
|
+
#include "ary.h"
|
28
|
+
#include "str.h"
|
29
|
+
#include "wikitext.h"
|
30
|
+
#include "wikitext_ragel.h"
|
31
|
+
|
32
|
+
#define IN(type) ary_includes(parser->scope, type)
|
33
|
+
#define IN_EITHER_OF(type1, type2) ary_includes2(parser->scope, type1, type2)
|
34
|
+
#define IN_ANY_OF(type1, type2, type3) ary_includes3(parser->scope, type1, type2, type3)
|
35
|
+
|
36
|
+
// output styles
|
37
|
+
enum { HTML_OUTPUT, XML_OUTPUT };
|
38
|
+
|
39
|
+
// poor man's object orientation in C:
|
40
|
+
// instead of passing around multiple parameters between functions in the parser
|
41
|
+
// we pack everything into a struct and pass around only a pointer to that
|
42
|
+
typedef struct
|
43
|
+
{
|
44
|
+
str_t *capture; // capturing to link_target, link_text, or NULL (direct to output, not capturing)
|
45
|
+
str_t *output; // for accumulating output to be returned
|
46
|
+
str_t *link_target; // short term "memory" for parsing links
|
47
|
+
str_t *link_text; // short term "memory" for parsing links
|
48
|
+
str_t *line_ending;
|
49
|
+
str_t *tabulation; // caching buffer for emitting indentation
|
50
|
+
ary_t *scope; // stack for tracking scope
|
51
|
+
ary_t *line; // stack for tracking scope as implied by current line
|
52
|
+
ary_t *line_buffer; // stack for tracking raw tokens (not scope) on current line
|
53
|
+
VALUE external_link_class; // CSS class applied to external links
|
54
|
+
VALUE external_link_rel; // rel attribute applied to external links
|
55
|
+
VALUE mailto_class; // CSS class applied to email (mailto) links
|
56
|
+
VALUE img_prefix; // path prepended when emitting img tags
|
57
|
+
int output_style; // HTML_OUTPUT (default) or XML_OUTPUT
|
58
|
+
int base_indent; // controlled by the :indent option to Wikitext::Parser#parse
|
59
|
+
int current_indent; // fluctuates according to currently nested structures
|
60
|
+
int base_heading_level;
|
61
|
+
bool pending_crlf;
|
62
|
+
bool autolink;
|
63
|
+
bool space_to_underscore;
|
64
|
+
} parser_t;
|
65
|
+
|
66
|
+
const char null_str[] = { 0 };
|
67
|
+
const char escaped_no_wiki_start[] = "<nowiki>";
|
68
|
+
const char escaped_no_wiki_end[] = "</nowiki>";
|
69
|
+
const char literal_strong_em[] = "'''''";
|
70
|
+
const char literal_strong[] = "'''";
|
71
|
+
const char literal_em[] = "''";
|
72
|
+
const char escaped_em_start[] = "<em>";
|
73
|
+
const char escaped_em_end[] = "</em>";
|
74
|
+
const char escaped_strong_start[] = "<strong>";
|
75
|
+
const char escaped_strong_end[] = "</strong>";
|
76
|
+
const char escaped_tt_start[] = "<tt>";
|
77
|
+
const char escaped_tt_end[] = "</tt>";
|
78
|
+
const char pre_start[] = "<pre>";
|
79
|
+
const char pre_end[] = "</pre>";
|
80
|
+
const char escaped_pre_start[] = "<pre>";
|
81
|
+
const char escaped_pre_end[] = "</pre>";
|
82
|
+
const char blockquote_start[] = "<blockquote>";
|
83
|
+
const char blockquote_end[] = "</blockquote>";
|
84
|
+
const char escaped_blockquote_start[] = "<blockquote>";
|
85
|
+
const char escaped_blockquote_end[] = "</blockquote>";
|
86
|
+
const char strong_em_start[] = "<strong><em>";
|
87
|
+
const char strong_start[] = "<strong>";
|
88
|
+
const char strong_end[] = "</strong>";
|
89
|
+
const char em_start[] = "<em>";
|
90
|
+
const char em_end[] = "</em>";
|
91
|
+
const char code_start[] = "<code>";
|
92
|
+
const char code_end[] = "</code>";
|
93
|
+
const char ol_start[] = "<ol>";
|
94
|
+
const char ol_end[] = "</ol>";
|
95
|
+
const char ul_start[] = "<ul>";
|
96
|
+
const char ul_end[] = "</ul>";
|
97
|
+
const char li_start[] = "<li>";
|
98
|
+
const char li_end[] = "</li>";
|
99
|
+
const char h6_start[] = "<h6>";
|
100
|
+
const char h6_end[] = "</h6>";
|
101
|
+
const char h5_start[] = "<h5>";
|
102
|
+
const char h5_end[] = "</h5>";
|
103
|
+
const char h4_start[] = "<h4>";
|
104
|
+
const char h4_end[] = "</h4>";
|
105
|
+
const char h3_start[] = "<h3>";
|
106
|
+
const char h3_end[] = "</h3>";
|
107
|
+
const char h2_start[] = "<h2>";
|
108
|
+
const char h2_end[] = "</h2>";
|
109
|
+
const char h1_start[] = "<h1>";
|
110
|
+
const char h1_end[] = "</h1>";
|
111
|
+
const char p_start[] = "<p>";
|
112
|
+
const char p_end[] = "</p>";
|
113
|
+
const char space[] = " ";
|
114
|
+
const char a_start[] = "<a href=\"";
|
115
|
+
const char a_class[] = "\" class=\"";
|
116
|
+
const char a_rel[] = "\" rel=\"";
|
117
|
+
const char a_start_close[] = "\">";
|
118
|
+
const char a_end[] = "</a>";
|
119
|
+
const char link_start[] = "[[";
|
120
|
+
const char link_end[] = "]]";
|
121
|
+
const char separator[] = "|";
|
122
|
+
const char ext_link_start[] = "[";
|
123
|
+
const char backtick[] = "`";
|
124
|
+
const char quote[] = "\"";
|
125
|
+
const char ampersand[] = "&";
|
126
|
+
const char quot_entity[] = """;
|
127
|
+
const char amp_entity[] = "&";
|
128
|
+
const char lt_entity[] = "<";
|
129
|
+
const char gt_entity[] = ">";
|
130
|
+
const char escaped_blockquote[] = "> ";
|
131
|
+
const char ext_link_end[] = "]";
|
132
|
+
const char literal_img_start[] = "{{";
|
133
|
+
const char img_start[] = "<img src=\"";
|
134
|
+
const char img_end_xml[] = "\" />";
|
135
|
+
const char img_end_html[] = "\">";
|
136
|
+
const char img_alt[] = "\" alt=\"";
|
137
|
+
const char pre_class_start[] = "<pre class=\"";
|
138
|
+
const char pre_class_end[] = "-syntax\">";
|
139
|
+
|
140
|
+
// Mark the parser struct designated by ptr as a participant in Ruby's
|
141
|
+
// mark-and-sweep garbage collection scheme. A variable named name is placed on
|
142
|
+
// the C stack to prevent the structure from being prematurely collected.
|
143
|
+
#define GC_WRAP_PARSER(ptr, name) volatile VALUE name __attribute__((unused)) = Data_Wrap_Struct(rb_cObject, 0, parser_free, ptr)
|
144
|
+
|
145
|
+
parser_t *parser_new(void)
|
146
|
+
{
|
147
|
+
parser_t *parser = ALLOC_N(parser_t, 1);
|
148
|
+
parser->capture = NULL; // not a real instance, pointer to other member's instance
|
149
|
+
parser->output = str_new();
|
150
|
+
parser->link_target = str_new();
|
151
|
+
parser->link_text = str_new();
|
152
|
+
parser->line_ending = NULL; // caller should set up
|
153
|
+
parser->tabulation = str_new();
|
154
|
+
parser->scope = ary_new();
|
155
|
+
parser->line = ary_new();
|
156
|
+
parser->line_buffer = ary_new();
|
157
|
+
parser->external_link_class = Qnil; // caller should set up
|
158
|
+
parser->external_link_rel = Qnil; // caller should set up
|
159
|
+
parser->mailto_class = Qnil; // caller should set up
|
160
|
+
parser->img_prefix = Qnil; // caller should set up
|
161
|
+
parser->output_style = HTML_OUTPUT;
|
162
|
+
parser->base_indent = 0;
|
163
|
+
parser->current_indent = 0;
|
164
|
+
parser->base_heading_level = 0;
|
165
|
+
parser->pending_crlf = false;
|
166
|
+
parser->autolink = true;
|
167
|
+
parser->space_to_underscore = true;
|
168
|
+
return parser;
|
169
|
+
}
|
170
|
+
|
171
|
+
void parser_free(parser_t *parser)
|
172
|
+
{
|
173
|
+
// we don't free parser->capture; it's just a redundant pointer
|
174
|
+
if (parser->output) str_free(parser->output);
|
175
|
+
if (parser->link_target) str_free(parser->link_target);
|
176
|
+
if (parser->link_text) str_free(parser->link_text);
|
177
|
+
if (parser->line_ending) str_free(parser->line_ending);
|
178
|
+
if (parser->tabulation) str_free(parser->tabulation);
|
179
|
+
if (parser->scope) ary_free(parser->scope);
|
180
|
+
if (parser->line) ary_free(parser->line);
|
181
|
+
if (parser->line_buffer) ary_free(parser->line_buffer);
|
182
|
+
free(parser);
|
183
|
+
}
|
184
|
+
|
185
|
+
// for testing and debugging only
|
186
|
+
VALUE Wikitext_parser_tokenize(VALUE self, VALUE string)
|
187
|
+
{
|
188
|
+
if (NIL_P(string))
|
189
|
+
return Qnil;
|
190
|
+
string = StringValue(string);
|
191
|
+
VALUE tokens = rb_ary_new();
|
192
|
+
char *p = RSTRING_PTR(string);
|
193
|
+
long len = RSTRING_LEN(string);
|
194
|
+
char *pe = p + len;
|
195
|
+
token_t token;
|
196
|
+
next_token(&token, NULL, p, pe);
|
197
|
+
rb_ary_push(tokens, wiki_token(&token));
|
198
|
+
while (token.type != END_OF_FILE)
|
199
|
+
{
|
200
|
+
next_token(&token, &token, NULL, pe);
|
201
|
+
rb_ary_push(tokens, wiki_token(&token));
|
202
|
+
}
|
203
|
+
return tokens;
|
204
|
+
}
|
205
|
+
|
206
|
+
// for benchmarking raw tokenization speed only
|
207
|
+
VALUE Wikitext_parser_benchmarking_tokenize(VALUE self, VALUE string)
|
208
|
+
{
|
209
|
+
if (NIL_P(string))
|
210
|
+
return Qnil;
|
211
|
+
string = StringValue(string);
|
212
|
+
char *p = RSTRING_PTR(string);
|
213
|
+
long len = RSTRING_LEN(string);
|
214
|
+
char *pe = p + len;
|
215
|
+
token_t token;
|
216
|
+
next_token(&token, NULL, p, pe);
|
217
|
+
while (token.type != END_OF_FILE)
|
218
|
+
next_token(&token, &token, NULL, pe);
|
219
|
+
return Qnil;
|
220
|
+
}
|
221
|
+
|
222
|
+
VALUE Wikitext_parser_fulltext_tokenize(int argc, VALUE *argv, VALUE self)
|
223
|
+
{
|
224
|
+
// process arguments
|
225
|
+
VALUE string, options;
|
226
|
+
if (rb_scan_args(argc, argv, "11", &string, &options) == 1) // 1 mandatory argument, 1 optional argument
|
227
|
+
options = Qnil;
|
228
|
+
if (NIL_P(string))
|
229
|
+
return Qnil;
|
230
|
+
string = StringValue(string);
|
231
|
+
VALUE tokens = rb_ary_new();
|
232
|
+
|
233
|
+
// check instance variables
|
234
|
+
VALUE min = rb_iv_get(self, "@minimum_fulltext_token_length");
|
235
|
+
|
236
|
+
// process options hash (can override instance variables)
|
237
|
+
if (!NIL_P(options) && TYPE(options) == T_HASH)
|
238
|
+
{
|
239
|
+
if (rb_funcall(options, rb_intern("has_key?"), 1, ID2SYM(rb_intern("minimum"))) == Qtrue)
|
240
|
+
min = rb_hash_aref(options, ID2SYM(rb_intern("minimum")));
|
241
|
+
}
|
242
|
+
int min_len = NIL_P(min) ? 3 : NUM2INT(min);
|
243
|
+
if (min_len < 0)
|
244
|
+
min_len = 0;
|
245
|
+
|
246
|
+
// set up scanner
|
247
|
+
char *p = RSTRING_PTR(string);
|
248
|
+
long len = RSTRING_LEN(string);
|
249
|
+
char *pe = p + len;
|
250
|
+
token_t token;
|
251
|
+
token_t *_token = &token;
|
252
|
+
next_token(&token, NULL, p, pe);
|
253
|
+
while (token.type != END_OF_FILE)
|
254
|
+
{
|
255
|
+
switch (token.type)
|
256
|
+
{
|
257
|
+
case URI:
|
258
|
+
case MAIL:
|
259
|
+
case ALNUM:
|
260
|
+
if (TOKEN_LEN(_token) >= min_len)
|
261
|
+
rb_ary_push(tokens, TOKEN_TEXT(_token));
|
262
|
+
break;
|
263
|
+
default:
|
264
|
+
// ignore everything else
|
265
|
+
break;
|
266
|
+
}
|
267
|
+
next_token(&token, &token, NULL, pe);
|
268
|
+
}
|
269
|
+
return tokens;
|
270
|
+
}
|
271
|
+
|
272
|
+
// we downcase "in place", overwriting the original contents of the buffer
|
273
|
+
void wiki_downcase_bang(char *ptr, long len)
|
274
|
+
{
|
275
|
+
for (long i = 0; i < len; i++)
|
276
|
+
{
|
277
|
+
if (ptr[i] >= 'A' && ptr[i] <= 'Z')
|
278
|
+
ptr[i] += 32;
|
279
|
+
}
|
280
|
+
}
|
281
|
+
|
282
|
+
void wiki_append_entity_from_utf32_char(str_t *output, uint32_t character)
|
283
|
+
{
|
284
|
+
char hex_string[8] = { '&', '#', 'x', 0, 0, 0, 0, ';' };
|
285
|
+
char scratch = (character & 0xf000) >> 12;
|
286
|
+
hex_string[3] = (scratch <= 9 ? scratch + 48 : scratch + 87);
|
287
|
+
scratch = (character & 0x0f00) >> 8;
|
288
|
+
hex_string[4] = (scratch <= 9 ? scratch + 48 : scratch + 87);
|
289
|
+
scratch = (character & 0x00f0) >> 4;
|
290
|
+
hex_string[5] = (scratch <= 9 ? scratch + 48 : scratch + 87);
|
291
|
+
scratch = character & 0x000f;
|
292
|
+
hex_string[6] = (scratch <= 9 ? scratch + 48 : scratch + 87);
|
293
|
+
str_append(output, hex_string, sizeof(hex_string));
|
294
|
+
}
|
295
|
+
|
296
|
+
// Convert a single UTF-8 codepoint to UTF-32
|
297
|
+
//
|
298
|
+
// Expects an input buffer, src, containing a UTF-8 encoded character (which
|
299
|
+
// may be multi-byte). The end of the input buffer, end, is also passed in to
|
300
|
+
// allow the detection of invalidly truncated codepoints. The number of bytes
|
301
|
+
// in the UTF-8 character (between 1 and 4) is returned by reference in
|
302
|
+
// width_out.
|
303
|
+
//
|
304
|
+
// Raises a RangeError if the supplied character is invalid UTF-8.
|
305
|
+
uint32_t wiki_utf8_to_utf32(char *src, char *end, long *width_out)
|
306
|
+
{
|
307
|
+
uint32_t dest = 0;
|
308
|
+
if ((unsigned char)src[0] <= 0x7f)
|
309
|
+
{
|
310
|
+
// ASCII
|
311
|
+
dest = src[0];
|
312
|
+
*width_out = 1;
|
313
|
+
}
|
314
|
+
else if ((src[0] & 0xe0) == 0xc0)
|
315
|
+
{
|
316
|
+
// byte starts with 110..... : this should be a two-byte sequence
|
317
|
+
if (src + 1 >= end)
|
318
|
+
// no second byte
|
319
|
+
rb_raise(eWikitextParserError, "invalid encoding: truncated byte sequence");
|
320
|
+
else if (((unsigned char)src[0] == 0xc0) ||
|
321
|
+
((unsigned char)src[0] == 0xc1))
|
322
|
+
// overlong encoding: lead byte of 110..... but code point <= 127
|
323
|
+
rb_raise(eWikitextParserError, "invalid encoding: overlong encoding");
|
324
|
+
else if ((src[1] & 0xc0) != 0x80 )
|
325
|
+
// should have second byte starting with 10......
|
326
|
+
rb_raise(eWikitextParserError, "invalid encoding: malformed byte sequence");
|
327
|
+
|
328
|
+
dest =
|
329
|
+
((uint32_t)(src[0] & 0x1f)) << 6 |
|
330
|
+
(src[1] & 0x3f);
|
331
|
+
*width_out = 2;
|
332
|
+
}
|
333
|
+
else if ((src[0] & 0xf0) == 0xe0)
|
334
|
+
{
|
335
|
+
// byte starts with 1110.... : this should be a three-byte sequence
|
336
|
+
if (src + 2 >= end)
|
337
|
+
// missing second or third byte
|
338
|
+
rb_raise(eWikitextParserError, "invalid encoding: truncated byte sequence");
|
339
|
+
else if (((src[1] & 0xc0) != 0x80 ) ||
|
340
|
+
((src[2] & 0xc0) != 0x80 ))
|
341
|
+
// should have second and third bytes starting with 10......
|
342
|
+
rb_raise(eWikitextParserError, "invalid encoding: malformed byte sequence");
|
343
|
+
|
344
|
+
dest =
|
345
|
+
((uint32_t)(src[0] & 0x0f)) << 12 |
|
346
|
+
((uint32_t)(src[1] & 0x3f)) << 6 |
|
347
|
+
(src[2] & 0x3f);
|
348
|
+
*width_out = 3;
|
349
|
+
}
|
350
|
+
else if ((src[0] & 0xf8) == 0xf0)
|
351
|
+
{
|
352
|
+
// bytes starts with 11110... : this should be a four-byte sequence
|
353
|
+
if (src + 3 >= end)
|
354
|
+
// missing second, third, or fourth byte
|
355
|
+
rb_raise(eWikitextParserError, "invalid encoding: truncated byte sequence");
|
356
|
+
else if ((unsigned char)src[0] >= 0xf5 &&
|
357
|
+
(unsigned char)src[0] <= 0xf7)
|
358
|
+
// disallowed by RFC 3629 (codepoints above 0x10ffff)
|
359
|
+
rb_raise(eWikitextParserError, "invalid encoding: overlong encoding");
|
360
|
+
else if (((src[1] & 0xc0) != 0x80 ) ||
|
361
|
+
((src[2] & 0xc0) != 0x80 ) ||
|
362
|
+
((src[3] & 0xc0) != 0x80 ))
|
363
|
+
// should have second and third bytes starting with 10......
|
364
|
+
rb_raise(eWikitextParserError, "invalid encoding: malformed byte sequence");
|
365
|
+
|
366
|
+
dest =
|
367
|
+
((uint32_t)(src[0] & 0x07)) << 18 |
|
368
|
+
((uint32_t)(src[1] & 0x3f)) << 12 |
|
369
|
+
((uint32_t)(src[1] & 0x3f)) << 6 |
|
370
|
+
(src[2] & 0x3f);
|
371
|
+
*width_out = 4;
|
372
|
+
}
|
373
|
+
else
|
374
|
+
rb_raise(eWikitextParserError, "invalid encoding: unexpected byte");
|
375
|
+
return dest;
|
376
|
+
}
|
377
|
+
|
378
|
+
// - non-printable (non-ASCII) characters converted to numeric entities
|
379
|
+
// - QUOT and AMP characters converted to named entities
|
380
|
+
// - if trim is true, leading and trailing whitespace trimmed
|
381
|
+
// - if trim is false, there is no special treatment of spaces
|
382
|
+
void wiki_append_sanitized_link_target(str_t *link_target, str_t *output, bool trim)
|
383
|
+
{
|
384
|
+
char *src = link_target->ptr;
|
385
|
+
char *start = src; // remember this so we can check if we're at the start
|
386
|
+
char *non_space = output->ptr + output->len; // remember last non-space character output
|
387
|
+
char *end = src + link_target->len;
|
388
|
+
while (src < end)
|
389
|
+
{
|
390
|
+
// need at most 8 bytes to display each input character (�)
|
391
|
+
if (output->ptr + output->len + 8 > output->ptr + output->capacity) // outgrowing buffer, must grow
|
392
|
+
{
|
393
|
+
char *old_ptr = output->ptr;
|
394
|
+
str_grow(output, output->len + (end - src) * 8); // allocate enough for worst case
|
395
|
+
if (old_ptr != output->ptr) // may have moved
|
396
|
+
non_space += output->ptr - old_ptr;
|
397
|
+
}
|
398
|
+
|
399
|
+
if (*src == '"')
|
400
|
+
{
|
401
|
+
char quot_entity_literal[] = { '&', 'q', 'u', 'o', 't', ';' }; // no trailing NUL
|
402
|
+
str_append(output, quot_entity_literal, sizeof(quot_entity_literal));
|
403
|
+
}
|
404
|
+
else if (*src == '&')
|
405
|
+
{
|
406
|
+
char amp_entity_literal[] = { '&', 'a', 'm', 'p', ';' }; // no trailing NUL
|
407
|
+
str_append(output, amp_entity_literal, sizeof(amp_entity_literal));
|
408
|
+
}
|
409
|
+
else if (*src == '<' || *src == '>')
|
410
|
+
rb_raise(rb_eRangeError, "invalid link text (\"%c\" may not appear in link text)", *src);
|
411
|
+
else if (*src == ' ' && src == start && trim)
|
412
|
+
start++; // we eat leading space
|
413
|
+
else if (*src >= 0x20 && *src <= 0x7e) // printable ASCII
|
414
|
+
{
|
415
|
+
*(output->ptr + output->len) = *src;
|
416
|
+
output->len++;
|
417
|
+
}
|
418
|
+
else // all others: must convert to entities
|
419
|
+
{
|
420
|
+
long width;
|
421
|
+
wiki_append_entity_from_utf32_char(output, wiki_utf8_to_utf32(src, end, &width));
|
422
|
+
src += width;
|
423
|
+
non_space = output->ptr + output->len;
|
424
|
+
continue;
|
425
|
+
}
|
426
|
+
if (*src != ' ')
|
427
|
+
non_space = output->ptr + output->len;
|
428
|
+
src++;
|
429
|
+
}
|
430
|
+
|
431
|
+
// trim trailing space if necessary
|
432
|
+
if (trim && output->ptr + output->len != non_space)
|
433
|
+
output->len -= (output->ptr + output->len) - non_space;
|
434
|
+
}
|
435
|
+
|
436
|
+
// prepare hyperlink and append it to parser->output
|
437
|
+
// if check_autolink is true, checks parser->autolink to decide whether to emit a real hyperlink
|
438
|
+
// or merely the literal link target
|
439
|
+
// if link_text is Qnil, the link_target is re-used for the link text
|
440
|
+
void wiki_append_hyperlink(parser_t *parser, VALUE link_prefix, str_t *link_target, str_t *link_text, VALUE link_class, VALUE link_rel, bool check_autolink)
|
441
|
+
{
|
442
|
+
if (check_autolink && !parser->autolink)
|
443
|
+
wiki_append_sanitized_link_target(link_target, parser->output, true);
|
444
|
+
else
|
445
|
+
{
|
446
|
+
str_append(parser->output, a_start, sizeof(a_start) - 1); // <a href="
|
447
|
+
if (!NIL_P(link_prefix))
|
448
|
+
str_append_string(parser->output, link_prefix);
|
449
|
+
wiki_append_sanitized_link_target(link_target, parser->output, true);
|
450
|
+
|
451
|
+
// special handling for mailto URIs
|
452
|
+
const char *mailto = "mailto:";
|
453
|
+
long mailto_len = (long)sizeof(mailto) - 1; // don't count NUL byte
|
454
|
+
if ((link_target->len >= mailto_len &&
|
455
|
+
strncmp(mailto, link_target->ptr, mailto_len) == 0) ||
|
456
|
+
(!NIL_P(link_prefix) &&
|
457
|
+
RSTRING_LEN(link_prefix) >= mailto_len &&
|
458
|
+
strncmp(mailto, RSTRING_PTR(link_prefix), mailto_len) == 0))
|
459
|
+
link_class = parser->mailto_class; // use mailto_class from parser
|
460
|
+
if (link_class != Qnil)
|
461
|
+
{
|
462
|
+
str_append(parser->output, a_class, sizeof(a_class) - 1); // " class="
|
463
|
+
str_append_string(parser->output, link_class);
|
464
|
+
}
|
465
|
+
if (link_rel != Qnil)
|
466
|
+
{
|
467
|
+
str_append(parser->output, a_rel, sizeof(a_rel) - 1); // " rel="
|
468
|
+
str_append_string(parser->output, link_rel);
|
469
|
+
}
|
470
|
+
str_append(parser->output, a_start_close, sizeof(a_start_close) - 1); // ">
|
471
|
+
if (!link_text || link_text->len == 0) // re-use link_target
|
472
|
+
wiki_append_sanitized_link_target(link_target, parser->output, true);
|
473
|
+
else
|
474
|
+
str_append_str(parser->output, link_text);
|
475
|
+
str_append(parser->output, a_end, sizeof(a_end) - 1); // </a>
|
476
|
+
}
|
477
|
+
}
|
478
|
+
|
479
|
+
void wiki_append_img(parser_t *parser, char *token_ptr, long token_len)
|
480
|
+
{
|
481
|
+
str_append(parser->output, img_start, sizeof(img_start) - 1); // <img src="
|
482
|
+
if (!NIL_P(parser->img_prefix) && *token_ptr != '/') // len always > 0
|
483
|
+
str_append_string(parser->output, parser->img_prefix);
|
484
|
+
str_append(parser->output, token_ptr, token_len);
|
485
|
+
str_append(parser->output, img_alt, sizeof(img_alt) - 1); // " alt="
|
486
|
+
str_append(parser->output, token_ptr, token_len);
|
487
|
+
if (parser->output_style == XML_OUTPUT)
|
488
|
+
str_append(parser->output, img_end_xml, sizeof(img_end_xml) - 1); // " />
|
489
|
+
else
|
490
|
+
str_append(parser->output, img_end_html, sizeof(img_end_html) - 1); // ">
|
491
|
+
}
|
492
|
+
|
493
|
+
// will emit indentation only if we are about to emit any of:
|
494
|
+
// <blockquote>, <p>, <ul>, <ol>, <li>, <h1> etc, <pre>
|
495
|
+
// each time we enter one of those spans must ++ the indentation level
|
496
|
+
void wiki_indent(parser_t *parser)
|
497
|
+
{
|
498
|
+
if (parser->base_indent == -1) // indentation disabled
|
499
|
+
return;
|
500
|
+
int space_count = parser->current_indent + parser->base_indent;
|
501
|
+
if (space_count > 0)
|
502
|
+
{
|
503
|
+
char *old_end, *new_end;
|
504
|
+
if (parser->tabulation->len < space_count)
|
505
|
+
str_grow(parser->tabulation, space_count); // reallocates if necessary
|
506
|
+
old_end = parser->tabulation->ptr + parser->tabulation->len;
|
507
|
+
new_end = parser->tabulation->ptr + space_count;
|
508
|
+
while (old_end < new_end)
|
509
|
+
*old_end++ = ' ';
|
510
|
+
if (space_count > parser->tabulation->len)
|
511
|
+
parser->tabulation->len = space_count;
|
512
|
+
str_append(parser->output, parser->tabulation->ptr, space_count);
|
513
|
+
}
|
514
|
+
parser->current_indent += 2;
|
515
|
+
}
|
516
|
+
|
517
|
+
void wiki_append_pre_start(parser_t *parser, token_t *token)
|
518
|
+
{
|
519
|
+
wiki_indent(parser);
|
520
|
+
if ((size_t)TOKEN_LEN(token) > sizeof(pre_start) - 1)
|
521
|
+
{
|
522
|
+
str_append(parser->output, pre_class_start, sizeof(pre_class_start) - 1); // <pre class="
|
523
|
+
str_append(parser->output, token->start + 11, TOKEN_LEN(token) - 13); // (the "lang" substring)
|
524
|
+
str_append(parser->output, pre_class_end, sizeof(pre_class_end) - 1); // -syntax">
|
525
|
+
}
|
526
|
+
else
|
527
|
+
str_append(parser->output, pre_start, sizeof(pre_start) - 1);
|
528
|
+
ary_push(parser->scope, PRE_START);
|
529
|
+
ary_push(parser->line, PRE_START);
|
530
|
+
}
|
531
|
+
|
532
|
+
void wiki_dedent(parser_t *parser, bool emit)
|
533
|
+
{
|
534
|
+
if (parser->base_indent == -1) // indentation disabled
|
535
|
+
return;
|
536
|
+
parser->current_indent -= 2;
|
537
|
+
if (!emit)
|
538
|
+
return;
|
539
|
+
int space_count = parser->current_indent + parser->base_indent;
|
540
|
+
if (space_count > 0)
|
541
|
+
str_append(parser->output, parser->tabulation->ptr, space_count);
|
542
|
+
}
|
543
|
+
|
544
|
+
// Pops a single item off the parser's scope stack.
|
545
|
+
// A corresponding closing tag is written to the target string.
|
546
|
+
// The target string may be the main output buffer, or a substring capturing buffer if a link is being scanned.
|
547
|
+
void wiki_pop_from_stack(parser_t *parser, str_t *target)
|
548
|
+
{
|
549
|
+
int top = ary_entry(parser->scope, -1);
|
550
|
+
if (NO_ITEM(top))
|
551
|
+
return;
|
552
|
+
if (!target)
|
553
|
+
target = parser->output;
|
554
|
+
|
555
|
+
// for headings, take base_heading_level into account
|
556
|
+
if (top >= H1_START && top <= H6_START)
|
557
|
+
{
|
558
|
+
top += parser->base_heading_level;
|
559
|
+
// no need to check for underflow (base_heading_level is never negative)
|
560
|
+
if (top > H6_START)
|
561
|
+
top = H6_START;
|
562
|
+
}
|
563
|
+
|
564
|
+
switch (top)
|
565
|
+
{
|
566
|
+
case PRE:
|
567
|
+
case PRE_START:
|
568
|
+
str_append(target, pre_end, sizeof(pre_end) - 1);
|
569
|
+
str_append_str(target, parser->line_ending);
|
570
|
+
wiki_dedent(parser, false);
|
571
|
+
break;
|
572
|
+
|
573
|
+
case BLOCKQUOTE:
|
574
|
+
case BLOCKQUOTE_START:
|
575
|
+
wiki_dedent(parser, true);
|
576
|
+
str_append(target, blockquote_end, sizeof(blockquote_end) - 1);
|
577
|
+
str_append_str(target, parser->line_ending);
|
578
|
+
break;
|
579
|
+
|
580
|
+
case NO_WIKI_START:
|
581
|
+
// not a real HTML tag; so nothing to pop
|
582
|
+
break;
|
583
|
+
|
584
|
+
case STRONG:
|
585
|
+
case STRONG_START:
|
586
|
+
str_append(target, strong_end, sizeof(strong_end) - 1);
|
587
|
+
break;
|
588
|
+
|
589
|
+
case EM:
|
590
|
+
case EM_START:
|
591
|
+
str_append(target, em_end, sizeof(em_end) - 1);
|
592
|
+
break;
|
593
|
+
|
594
|
+
case TT:
|
595
|
+
case TT_START:
|
596
|
+
str_append(target, code_end, sizeof(code_end) - 1);
|
597
|
+
break;
|
598
|
+
|
599
|
+
case OL:
|
600
|
+
wiki_dedent(parser, true);
|
601
|
+
str_append(target, ol_end, sizeof(ol_end) - 1);
|
602
|
+
str_append_str(target, parser->line_ending);
|
603
|
+
break;
|
604
|
+
|
605
|
+
case UL:
|
606
|
+
wiki_dedent(parser, true);
|
607
|
+
str_append(target, ul_end, sizeof(ul_end) - 1);
|
608
|
+
str_append_str(target, parser->line_ending);
|
609
|
+
break;
|
610
|
+
|
611
|
+
case NESTED_LIST:
|
612
|
+
// next token to pop will be a LI
|
613
|
+
// LI is an interesting token because sometimes we want it to behave like P (ie. do a non-emitting indent)
|
614
|
+
// and other times we want it to behave like BLOCKQUOTE (ie. when it has a nested list inside)
|
615
|
+
// hence this hack: we do an emitting dedent on behalf of the LI that we know must be coming
|
616
|
+
// and then when we pop the actual LI itself (below) we do the standard non-emitting indent
|
617
|
+
wiki_dedent(parser, true); // we really only want to emit the spaces
|
618
|
+
parser->current_indent += 2; // we don't want to decrement the actual indent level, so put it back
|
619
|
+
break;
|
620
|
+
|
621
|
+
case LI:
|
622
|
+
str_append(target, li_end, sizeof(li_end) - 1);
|
623
|
+
str_append_str(target, parser->line_ending);
|
624
|
+
wiki_dedent(parser, false);
|
625
|
+
break;
|
626
|
+
|
627
|
+
case H6_START:
|
628
|
+
str_append(target, h6_end, sizeof(h6_end) - 1);
|
629
|
+
str_append_str(target, parser->line_ending);
|
630
|
+
wiki_dedent(parser, false);
|
631
|
+
break;
|
632
|
+
|
633
|
+
case H5_START:
|
634
|
+
str_append(target, h5_end, sizeof(h5_end) - 1);
|
635
|
+
str_append_str(target, parser->line_ending);
|
636
|
+
wiki_dedent(parser, false);
|
637
|
+
break;
|
638
|
+
|
639
|
+
case H4_START:
|
640
|
+
str_append(target, h4_end, sizeof(h4_end) - 1);
|
641
|
+
str_append_str(target, parser->line_ending);
|
642
|
+
wiki_dedent(parser, false);
|
643
|
+
break;
|
644
|
+
|
645
|
+
case H3_START:
|
646
|
+
str_append(target, h3_end, sizeof(h3_end) - 1);
|
647
|
+
str_append_str(target, parser->line_ending);
|
648
|
+
wiki_dedent(parser, false);
|
649
|
+
break;
|
650
|
+
|
651
|
+
case H2_START:
|
652
|
+
str_append(target, h2_end, sizeof(h2_end) - 1);
|
653
|
+
str_append_str(target, parser->line_ending);
|
654
|
+
wiki_dedent(parser, false);
|
655
|
+
break;
|
656
|
+
|
657
|
+
case H1_START:
|
658
|
+
str_append(target, h1_end, sizeof(h1_end) - 1);
|
659
|
+
str_append_str(target, parser->line_ending);
|
660
|
+
wiki_dedent(parser, false);
|
661
|
+
break;
|
662
|
+
|
663
|
+
case LINK_START:
|
664
|
+
// not an HTML tag; so nothing to emit
|
665
|
+
break;
|
666
|
+
|
667
|
+
case EXT_LINK_START:
|
668
|
+
// not an HTML tag; so nothing to emit
|
669
|
+
break;
|
670
|
+
|
671
|
+
case PATH:
|
672
|
+
// not an HTML tag; so nothing to emit
|
673
|
+
break;
|
674
|
+
|
675
|
+
case SPACE:
|
676
|
+
// not an HTML tag (only used to separate an external link target from the link text); so nothing to emit
|
677
|
+
break;
|
678
|
+
|
679
|
+
case SEPARATOR:
|
680
|
+
// not an HTML tag (only used to separate an external link target from the link text); so nothing to emit
|
681
|
+
break;
|
682
|
+
|
683
|
+
case P:
|
684
|
+
str_append(target, p_end, sizeof(p_end) - 1);
|
685
|
+
str_append_str(target, parser->line_ending);
|
686
|
+
wiki_dedent(parser, false);
|
687
|
+
break;
|
688
|
+
|
689
|
+
case END_OF_FILE:
|
690
|
+
// nothing to do
|
691
|
+
break;
|
692
|
+
|
693
|
+
default:
|
694
|
+
// should probably raise an exception here
|
695
|
+
break;
|
696
|
+
}
|
697
|
+
ary_pop(parser->scope);
|
698
|
+
}
|
699
|
+
|
700
|
+
// Pops items off the top of parser's scope stack, accumulating closing tags for them into the target string, until item is reached.
|
701
|
+
// If including is true then the item itself is also popped.
|
702
|
+
// The target string may be the main output buffer, or a substring capturing buffer when scanning links.
|
703
|
+
void wiki_pop_from_stack_up_to(parser_t *parser, str_t *target, int item, bool including)
|
704
|
+
{
|
705
|
+
int continue_looping = 1;
|
706
|
+
do
|
707
|
+
{
|
708
|
+
int top = ary_entry(parser->scope, -1);
|
709
|
+
if (NO_ITEM(top))
|
710
|
+
return;
|
711
|
+
if (top == item)
|
712
|
+
{
|
713
|
+
if (!including)
|
714
|
+
return;
|
715
|
+
continue_looping = 0;
|
716
|
+
}
|
717
|
+
wiki_pop_from_stack(parser, target);
|
718
|
+
} while (continue_looping);
|
719
|
+
}
|
720
|
+
|
721
|
+
void wiki_pop_all_from_stack(parser_t *parser)
|
722
|
+
{
|
723
|
+
for (int i = 0, max = parser->scope->count; i < max; i++)
|
724
|
+
wiki_pop_from_stack(parser, NULL);
|
725
|
+
}
|
726
|
+
|
727
|
+
void wiki_start_para_if_necessary(parser_t *parser)
|
728
|
+
{
|
729
|
+
if (parser->capture)
|
730
|
+
return;
|
731
|
+
|
732
|
+
// if no block open yet, or top of stack is BLOCKQUOTE/BLOCKQUOTE_START (with nothing in it yet)
|
733
|
+
if (parser->scope->count == 0 ||
|
734
|
+
ary_entry(parser->scope, -1) == BLOCKQUOTE ||
|
735
|
+
ary_entry(parser->scope, -1) == BLOCKQUOTE_START)
|
736
|
+
{
|
737
|
+
wiki_indent(parser);
|
738
|
+
str_append(parser->output, p_start, sizeof(p_start) - 1);
|
739
|
+
ary_push(parser->scope, P);
|
740
|
+
ary_push(parser->line, P);
|
741
|
+
}
|
742
|
+
else if (parser->pending_crlf)
|
743
|
+
{
|
744
|
+
if (IN(P))
|
745
|
+
// already in a paragraph block; convert pending CRLF into a space
|
746
|
+
str_append(parser->output, space, sizeof(space) - 1);
|
747
|
+
else if (IN(PRE))
|
748
|
+
// PRE blocks can have pending CRLF too (helps us avoid emitting the trailing newline)
|
749
|
+
str_append_str(parser->output, parser->line_ending);
|
750
|
+
}
|
751
|
+
parser->pending_crlf = false;
|
752
|
+
}
|
753
|
+
|
754
|
+
void wiki_emit_pending_crlf_if_necessary(parser_t *parser)
|
755
|
+
{
|
756
|
+
if (parser->pending_crlf)
|
757
|
+
{
|
758
|
+
str_append_str(parser->output, parser->line_ending);
|
759
|
+
parser->pending_crlf = false;
|
760
|
+
}
|
761
|
+
}
|
762
|
+
|
763
|
+
// Helper function that pops any excess elements off scope (pushing is already handled in the respective rules).
|
764
|
+
// For example, given input like:
|
765
|
+
//
|
766
|
+
// > > foo
|
767
|
+
// bar
|
768
|
+
//
|
769
|
+
// Upon seeing "bar", we want to pop two BLOCKQUOTE elements from the scope.
|
770
|
+
// The reverse case (shown below) is handled from inside the BLOCKQUOTE rule itself:
|
771
|
+
//
|
772
|
+
// foo
|
773
|
+
// > > bar
|
774
|
+
//
|
775
|
+
// Things are made slightly more complicated by the fact that there is one block-level tag that can be on the scope
|
776
|
+
// but not on the line scope:
|
777
|
+
//
|
778
|
+
// <blockquote>foo
|
779
|
+
// bar</blockquote>
|
780
|
+
//
|
781
|
+
// Here on seeing "bar" we have one item on the scope (BLOCKQUOTE_START) which we don't want to pop, but we have nothing
|
782
|
+
// on the line scope.
|
783
|
+
// Luckily, BLOCKQUOTE_START tokens can only appear at the start of the scope array, so we can check for them first before
|
784
|
+
// entering the for loop.
|
785
|
+
void wiki_pop_excess_elements(parser_t *parser)
|
786
|
+
{
|
787
|
+
if (parser->capture)
|
788
|
+
return;
|
789
|
+
for (int i = parser->scope->count - ary_count(parser->scope, BLOCKQUOTE_START), j = parser->line->count; i > j; i--)
|
790
|
+
{
|
791
|
+
// special case for last item on scope
|
792
|
+
if (i - j == 1)
|
793
|
+
{
|
794
|
+
// don't auto-pop P if it is only item on scope
|
795
|
+
if (ary_entry(parser->scope, -1) == P)
|
796
|
+
{
|
797
|
+
// add P to the line scope to prevent us entering the loop at all next time around
|
798
|
+
ary_push(parser->line, P);
|
799
|
+
continue;
|
800
|
+
}
|
801
|
+
}
|
802
|
+
wiki_pop_from_stack(parser, NULL);
|
803
|
+
}
|
804
|
+
}
|
805
|
+
|
806
|
+
// trim parser->link_text in place
|
807
|
+
void wiki_trim_link_text(parser_t *parser)
|
808
|
+
{
|
809
|
+
char *src = parser->link_text->ptr;
|
810
|
+
char *start = src; // remember this so we can check if we're at the start
|
811
|
+
char *left = src;
|
812
|
+
char *non_space = src; // remember last non-space character output
|
813
|
+
char *end = src + parser->link_text->len;
|
814
|
+
while (src < end)
|
815
|
+
{
|
816
|
+
if (*src == ' ')
|
817
|
+
{
|
818
|
+
if (src == left)
|
819
|
+
left++;
|
820
|
+
}
|
821
|
+
else
|
822
|
+
non_space = src;
|
823
|
+
src++;
|
824
|
+
}
|
825
|
+
if (left != start || non_space + 1 != end)
|
826
|
+
{
|
827
|
+
// TODO: could potentially avoid this memmove by extending the str_t struct with an "offset" or "free" member
|
828
|
+
parser->link_text->len = (non_space + 1) - left;
|
829
|
+
memmove(parser->link_text->ptr, left, parser->link_text->len);
|
830
|
+
}
|
831
|
+
}
|
832
|
+
|
833
|
+
VALUE Wikitext_parser_sanitize_link_target(VALUE self, VALUE string)
|
834
|
+
{
|
835
|
+
str_t *link_target = str_new_from_string(string);
|
836
|
+
GC_WRAP_STR(link_target, link_target_gc);
|
837
|
+
str_t *output = str_new();
|
838
|
+
GC_WRAP_STR(output, output_gc);
|
839
|
+
wiki_append_sanitized_link_target(link_target, output, true);
|
840
|
+
return string_from_str(output);
|
841
|
+
}
|
842
|
+
|
843
|
+
// Encodes the parser link_target member (in-place) according to RFCs 2396 and 2718
|
844
|
+
//
|
845
|
+
// Leading and trailing whitespace trimmed. Spaces are converted to
|
846
|
+
// underscores if the parser space_to_underscore member is true.
|
847
|
+
static void wiki_encode_link_target(parser_t *parser)
|
848
|
+
{
|
849
|
+
char *src = parser->link_target->ptr;
|
850
|
+
char *start = src; // remember this so we can check if we're at the start
|
851
|
+
long len = parser->link_target->len;
|
852
|
+
if (!(len > 0))
|
853
|
+
return;
|
854
|
+
char *end = src + len;
|
855
|
+
long dest_len = len * 2;
|
856
|
+
char *dest = ALLOC_N(char, dest_len);
|
857
|
+
char *dest_ptr = dest; // hang on to this so we can pass it to free() later
|
858
|
+
char *non_space = dest; // remember last non-space character output
|
859
|
+
static char hex[] = { '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f' };
|
860
|
+
for (; src < end; src++)
|
861
|
+
{
|
862
|
+
// worst case: a single character may grow to 3 characters once encoded
|
863
|
+
if ((dest + 3) > (dest_ptr + dest_len))
|
864
|
+
{
|
865
|
+
// outgrowing buffer, must reallocate
|
866
|
+
char *old_dest = dest;
|
867
|
+
char *old_dest_ptr = dest_ptr;
|
868
|
+
dest_len += len;
|
869
|
+
dest = realloc(dest_ptr, dest_len);
|
870
|
+
if (dest == NULL)
|
871
|
+
{
|
872
|
+
// would have used reallocf, but this has to run on Linux too, not just Darwin
|
873
|
+
free(dest_ptr);
|
874
|
+
rb_raise(rb_eNoMemError, "failed to re-allocate temporary storage (memory allocation error)");
|
875
|
+
}
|
876
|
+
dest_ptr = dest;
|
877
|
+
dest = dest_ptr + (old_dest - old_dest_ptr);
|
878
|
+
non_space = dest_ptr + (non_space - old_dest_ptr);
|
879
|
+
}
|
880
|
+
|
881
|
+
// pass through unreserved characters
|
882
|
+
if ((*src >= 'a' && *src <= 'z') ||
|
883
|
+
(*src >= 'A' && *src <= 'Z') ||
|
884
|
+
(*src >= '0' && *src <= '9') ||
|
885
|
+
*src == '-' ||
|
886
|
+
*src == '_' ||
|
887
|
+
*src == '.' ||
|
888
|
+
*src == '~')
|
889
|
+
{
|
890
|
+
*dest++ = *src;
|
891
|
+
non_space = dest;
|
892
|
+
}
|
893
|
+
else if (*src == ' ' && src == start)
|
894
|
+
start++; // we eat leading space
|
895
|
+
else if (*src == ' ' && parser->space_to_underscore)
|
896
|
+
*dest++ = '_';
|
897
|
+
else // everything else gets URL-encoded
|
898
|
+
{
|
899
|
+
*dest++ = '%';
|
900
|
+
*dest++ = hex[(unsigned char)(*src) / 16]; // left
|
901
|
+
*dest++ = hex[(unsigned char)(*src) % 16]; // right
|
902
|
+
if (*src != ' ')
|
903
|
+
non_space = dest;
|
904
|
+
}
|
905
|
+
}
|
906
|
+
|
907
|
+
// trim trailing space if necessary
|
908
|
+
if (non_space > dest_ptr && dest != non_space)
|
909
|
+
dest_len = non_space - dest_ptr;
|
910
|
+
else
|
911
|
+
dest_len = dest - dest_ptr;
|
912
|
+
str_clear(parser->link_target);
|
913
|
+
str_append(parser->link_target, dest_ptr, dest_len);
|
914
|
+
free(dest_ptr);
|
915
|
+
}
|
916
|
+
|
917
|
+
VALUE Wikitext_parser_encode_link_target(VALUE self, VALUE in)
|
918
|
+
{
|
919
|
+
parser_t parser;
|
920
|
+
parser.space_to_underscore = false;
|
921
|
+
parser.link_target = str_new_from_string(in);
|
922
|
+
GC_WRAP_STR(parser.link_target, link_target_gc);
|
923
|
+
wiki_encode_link_target(&parser);
|
924
|
+
return string_from_str(parser.link_target);
|
925
|
+
}
|
926
|
+
|
927
|
+
// returns 1 (true) if supplied string is blank (nil, empty, or all whitespace)
|
928
|
+
// returns 0 (false) otherwise
|
929
|
+
bool wiki_blank(str_t *str)
|
930
|
+
{
|
931
|
+
if (str->len == 0)
|
932
|
+
return true;
|
933
|
+
for (char *ptr = str->ptr,
|
934
|
+
*end = str->ptr + str->len;
|
935
|
+
ptr < end; ptr++)
|
936
|
+
{
|
937
|
+
if (*ptr != ' ')
|
938
|
+
return false;
|
939
|
+
}
|
940
|
+
return true;
|
941
|
+
}
|
942
|
+
|
943
|
+
void wiki_rollback_failed_internal_link(parser_t *parser)
|
944
|
+
{
|
945
|
+
if (!IN(LINK_START))
|
946
|
+
return; // nothing to do!
|
947
|
+
int scope_includes_separator = IN(SEPARATOR);
|
948
|
+
wiki_pop_from_stack_up_to(parser, NULL, LINK_START, true);
|
949
|
+
str_append(parser->output, link_start, sizeof(link_start) - 1);
|
950
|
+
if (parser->link_target->len > 0)
|
951
|
+
{
|
952
|
+
wiki_append_sanitized_link_target(parser->link_target, parser->output, false);
|
953
|
+
if (scope_includes_separator)
|
954
|
+
{
|
955
|
+
str_append(parser->output, separator, sizeof(separator) - 1);
|
956
|
+
if (parser->link_text->len > 0)
|
957
|
+
str_append_str(parser->output, parser->link_text);
|
958
|
+
}
|
959
|
+
}
|
960
|
+
parser->capture = NULL;
|
961
|
+
str_clear(parser->link_target);
|
962
|
+
str_clear(parser->link_text);
|
963
|
+
}
|
964
|
+
|
965
|
+
void wiki_rollback_failed_external_link(parser_t *parser)
|
966
|
+
{
|
967
|
+
if (!IN(EXT_LINK_START))
|
968
|
+
return; // nothing to do!
|
969
|
+
|
970
|
+
// store a couple of values before popping
|
971
|
+
int scope_includes_space = IN(SPACE);
|
972
|
+
VALUE link_class = IN(PATH) ? Qnil : parser->external_link_class;
|
973
|
+
VALUE link_rel = IN(PATH) ? Qnil : parser->external_link_rel;
|
974
|
+
wiki_pop_from_stack_up_to(parser, NULL, EXT_LINK_START, true);
|
975
|
+
|
976
|
+
str_append(parser->output, ext_link_start, sizeof(ext_link_start) - 1);
|
977
|
+
if (parser->link_target->len > 0)
|
978
|
+
{
|
979
|
+
wiki_append_hyperlink(parser, Qnil, parser->link_target, NULL, link_class, link_rel, true);
|
980
|
+
if (scope_includes_space)
|
981
|
+
{
|
982
|
+
str_append(parser->output, space, sizeof(space) - 1);
|
983
|
+
if (parser->link_text->len > 0)
|
984
|
+
str_append_str(parser->output, parser->link_text);
|
985
|
+
}
|
986
|
+
}
|
987
|
+
parser->capture = NULL;
|
988
|
+
str_clear(parser->link_target);
|
989
|
+
str_clear(parser->link_text);
|
990
|
+
}
|
991
|
+
|
992
|
+
void wiki_rollback_failed_link(parser_t *parser)
|
993
|
+
{
|
994
|
+
wiki_rollback_failed_internal_link(parser);
|
995
|
+
wiki_rollback_failed_external_link(parser);
|
996
|
+
}
|
997
|
+
|
998
|
+
VALUE Wikitext_parser_initialize(int argc, VALUE *argv, VALUE self)
|
999
|
+
{
|
1000
|
+
// process arguments
|
1001
|
+
VALUE options;
|
1002
|
+
if (rb_scan_args(argc, argv, "01", &options) == 0) // 0 mandatory arguments, 1 optional argument
|
1003
|
+
options = Qnil;
|
1004
|
+
|
1005
|
+
// defaults
|
1006
|
+
VALUE autolink = Qtrue;
|
1007
|
+
VALUE line_ending = rb_str_new2("\n");
|
1008
|
+
VALUE external_link_class = rb_str_new2("external");
|
1009
|
+
VALUE external_link_rel = Qnil;
|
1010
|
+
VALUE mailto_class = rb_str_new2("mailto");
|
1011
|
+
VALUE link_proc = Qnil;
|
1012
|
+
VALUE internal_link_prefix = rb_str_new2("/wiki/");
|
1013
|
+
VALUE img_prefix = rb_str_new2("/images/");
|
1014
|
+
VALUE output_style = ID2SYM(rb_intern("html"));
|
1015
|
+
VALUE space_to_underscore = Qtrue;
|
1016
|
+
VALUE minimum_fulltext_token_length = INT2NUM(3);
|
1017
|
+
VALUE base_heading_level = INT2NUM(0);
|
1018
|
+
|
1019
|
+
// process options hash (override defaults)
|
1020
|
+
if (!NIL_P(options) && TYPE(options) == T_HASH)
|
1021
|
+
{
|
1022
|
+
#define OVERRIDE_IF_SET(name) rb_funcall(options, rb_intern("has_key?"), 1, ID2SYM(rb_intern(#name))) == Qtrue ? \
|
1023
|
+
rb_hash_aref(options, ID2SYM(rb_intern(#name))) : name
|
1024
|
+
autolink = OVERRIDE_IF_SET(autolink);
|
1025
|
+
line_ending = OVERRIDE_IF_SET(line_ending);
|
1026
|
+
external_link_class = OVERRIDE_IF_SET(external_link_class);
|
1027
|
+
external_link_rel = OVERRIDE_IF_SET(external_link_rel);
|
1028
|
+
mailto_class = OVERRIDE_IF_SET(mailto_class);
|
1029
|
+
link_proc = OVERRIDE_IF_SET(link_proc);
|
1030
|
+
internal_link_prefix = OVERRIDE_IF_SET(internal_link_prefix);
|
1031
|
+
img_prefix = OVERRIDE_IF_SET(img_prefix);
|
1032
|
+
output_style = OVERRIDE_IF_SET(output_style);
|
1033
|
+
space_to_underscore = OVERRIDE_IF_SET(space_to_underscore);
|
1034
|
+
minimum_fulltext_token_length = OVERRIDE_IF_SET(minimum_fulltext_token_length);
|
1035
|
+
base_heading_level = OVERRIDE_IF_SET(base_heading_level);
|
1036
|
+
}
|
1037
|
+
|
1038
|
+
// no need to call super here; rb_call_super()
|
1039
|
+
rb_iv_set(self, "@autolink", autolink);
|
1040
|
+
rb_iv_set(self, "@line_ending", line_ending);
|
1041
|
+
rb_iv_set(self, "@external_link_class", external_link_class);
|
1042
|
+
rb_iv_set(self, "@external_link_rel", external_link_rel);
|
1043
|
+
rb_iv_set(self, "@mailto_class", mailto_class);
|
1044
|
+
rb_iv_set(self, "@link_proc", link_proc);
|
1045
|
+
rb_iv_set(self, "@internal_link_prefix", internal_link_prefix);
|
1046
|
+
rb_iv_set(self, "@img_prefix", img_prefix);
|
1047
|
+
rb_iv_set(self, "@output_style", output_style);
|
1048
|
+
rb_iv_set(self, "@space_to_underscore", space_to_underscore);
|
1049
|
+
rb_iv_set(self, "@minimum_fulltext_token_length", minimum_fulltext_token_length);
|
1050
|
+
rb_iv_set(self, "@base_heading_level", base_heading_level);
|
1051
|
+
return self;
|
1052
|
+
}
|
1053
|
+
|
1054
|
+
VALUE Wikitext_parser_profiling_parse(VALUE self, VALUE string)
|
1055
|
+
{
|
1056
|
+
for (int i = 0; i < 100000; i++)
|
1057
|
+
Wikitext_parser_parse(1, &string, self);
|
1058
|
+
return Qnil;
|
1059
|
+
}
|
1060
|
+
|
1061
|
+
// convert a Ruby object (:xml, :html etc) into an int output style
|
1062
|
+
int Wikitext_output_style(VALUE output)
|
1063
|
+
{
|
1064
|
+
if (TYPE(output) == T_SYMBOL)
|
1065
|
+
{
|
1066
|
+
if (SYM2ID(output) == rb_intern("xml"))
|
1067
|
+
return XML_OUTPUT;
|
1068
|
+
}
|
1069
|
+
return HTML_OUTPUT; // fall back to default
|
1070
|
+
}
|
1071
|
+
|
1072
|
+
VALUE Wikitext_parser_parse(int argc, VALUE *argv, VALUE self)
|
1073
|
+
{
|
1074
|
+
// process arguments
|
1075
|
+
VALUE string, options;
|
1076
|
+
if (rb_scan_args(argc, argv, "11", &string, &options) == 1) // 1 mandatory argument, 1 optional argument
|
1077
|
+
options = Qnil;
|
1078
|
+
if (NIL_P(string))
|
1079
|
+
return Qnil;
|
1080
|
+
string = StringValue(string);
|
1081
|
+
|
1082
|
+
// access these once per parse
|
1083
|
+
VALUE line_ending = rb_iv_get(self, "@line_ending");
|
1084
|
+
line_ending = StringValue(line_ending);
|
1085
|
+
VALUE link_class = rb_iv_get(self, "@external_link_class");
|
1086
|
+
link_class = NIL_P(link_class) ? Qnil : StringValue(link_class);
|
1087
|
+
VALUE link_rel = rb_iv_get(self, "@external_link_rel");
|
1088
|
+
link_rel = NIL_P(link_rel) ? Qnil : StringValue(link_rel);
|
1089
|
+
VALUE link_proc = rb_iv_get(self, "@link_proc");
|
1090
|
+
VALUE mailto_class = rb_iv_get(self, "@mailto_class");
|
1091
|
+
mailto_class = NIL_P(mailto_class) ? Qnil : StringValue(mailto_class);
|
1092
|
+
VALUE prefix = rb_iv_get(self, "@internal_link_prefix");
|
1093
|
+
int output_style = Wikitext_output_style(rb_iv_get(self, "@output_style"));
|
1094
|
+
|
1095
|
+
// process options hash
|
1096
|
+
int base_indent = 0;
|
1097
|
+
int base_heading_level = NUM2INT(rb_iv_get(self, "@base_heading_level"));
|
1098
|
+
if (!NIL_P(options) && TYPE(options) == T_HASH)
|
1099
|
+
{
|
1100
|
+
// :indent => 0 (or more)
|
1101
|
+
ID has_key = rb_intern("has_key?");
|
1102
|
+
ID id = ID2SYM(rb_intern("indent"));
|
1103
|
+
if (rb_funcall(options, has_key, 1, id) == Qtrue)
|
1104
|
+
{
|
1105
|
+
VALUE indent = rb_hash_aref(options, id);
|
1106
|
+
if (indent == Qfalse)
|
1107
|
+
base_indent = -1; // indentation disabled
|
1108
|
+
else
|
1109
|
+
{
|
1110
|
+
base_indent = NUM2INT(indent);
|
1111
|
+
if (base_indent < 0)
|
1112
|
+
base_indent = 0;
|
1113
|
+
}
|
1114
|
+
}
|
1115
|
+
|
1116
|
+
// :base_heading_level => 0/1/2/3/4/5/6
|
1117
|
+
id = ID2SYM(rb_intern("base_heading_level"));
|
1118
|
+
if (rb_funcall(options, has_key, 1, id) == Qtrue)
|
1119
|
+
base_heading_level = NUM2INT(rb_hash_aref(options, id));
|
1120
|
+
|
1121
|
+
// :external_link_rel => 'nofollow'
|
1122
|
+
id = ID2SYM(rb_intern("external_link_rel"));
|
1123
|
+
if (rb_funcall(options, has_key, 1, id) == Qtrue)
|
1124
|
+
{
|
1125
|
+
link_rel = rb_hash_aref(options, id);
|
1126
|
+
link_rel = NIL_P(link_rel) ? Qnil : StringValue(link_rel);
|
1127
|
+
}
|
1128
|
+
|
1129
|
+
// :output_style => :html/:xml
|
1130
|
+
id = ID2SYM(rb_intern("output_style"));
|
1131
|
+
if (rb_funcall(options, has_key, 1, id) == Qtrue)
|
1132
|
+
output_style = Wikitext_output_style(rb_hash_aref(options, id));
|
1133
|
+
|
1134
|
+
// :link_proc => lambda { |link_target| ... }
|
1135
|
+
id = ID2SYM(rb_intern("link_proc"));
|
1136
|
+
if (rb_funcall(options, has_key, 1, id) == Qtrue)
|
1137
|
+
link_proc = rb_hash_aref(options, id);
|
1138
|
+
}
|
1139
|
+
|
1140
|
+
// normalize, regardless of whether this came from instance variable or override
|
1141
|
+
if (base_heading_level < 0)
|
1142
|
+
base_heading_level = 0;
|
1143
|
+
if (base_heading_level > 6)
|
1144
|
+
base_heading_level = 6;
|
1145
|
+
|
1146
|
+
// set up scanner
|
1147
|
+
char *p = RSTRING_PTR(string);
|
1148
|
+
long len = RSTRING_LEN(string);
|
1149
|
+
char *pe = p + len;
|
1150
|
+
|
1151
|
+
// set up parser struct to make passing parameters a little easier
|
1152
|
+
parser_t *parser = parser_new();
|
1153
|
+
GC_WRAP_PARSER(parser, parser_gc);
|
1154
|
+
parser->external_link_class = link_class;
|
1155
|
+
parser->external_link_rel = link_rel;
|
1156
|
+
parser->mailto_class = mailto_class;
|
1157
|
+
parser->img_prefix = rb_iv_get(self, "@img_prefix");
|
1158
|
+
parser->autolink = rb_iv_get(self, "@autolink") == Qtrue ? true : false;
|
1159
|
+
parser->space_to_underscore = rb_iv_get(self, "@space_to_underscore") == Qtrue ? true : false;
|
1160
|
+
parser->line_ending = str_new_from_string(line_ending);
|
1161
|
+
parser->base_indent = base_indent;
|
1162
|
+
parser->base_heading_level = base_heading_level;
|
1163
|
+
parser->output_style = output_style;
|
1164
|
+
|
1165
|
+
// this simple looping design leads to a single enormous function,
|
1166
|
+
// but it's faster than doing actual recursive descent and also secure in the face of
|
1167
|
+
// malicious input that seeks to overflow the stack
|
1168
|
+
// (with "<blockquote><blockquote><blockquote>..." times by 10,000, for example)
|
1169
|
+
// given that we expect to deal with a lot of malformed input, a recursive descent design is less appropriate
|
1170
|
+
// than a straightforward looping translator like this one anyway
|
1171
|
+
token_t _token;
|
1172
|
+
_token.type = NO_TOKEN;
|
1173
|
+
token_t *token = NULL;
|
1174
|
+
do
|
1175
|
+
{
|
1176
|
+
// note that whenever we grab a token we push it into the line buffer
|
1177
|
+
// this provides us with context-sensitive "memory" of what's been seen so far on this line
|
1178
|
+
#define NEXT_TOKEN() token = &_token, next_token(token, token, NULL, pe), ary_push(parser->line_buffer, token->type)
|
1179
|
+
|
1180
|
+
// check to see if we have a token hanging around from a previous iteration of this loop
|
1181
|
+
if (token == NULL)
|
1182
|
+
{
|
1183
|
+
if (_token.type == NO_TOKEN)
|
1184
|
+
{
|
1185
|
+
// first time here (haven't started scanning yet)
|
1186
|
+
token = &_token;
|
1187
|
+
next_token(token, NULL, p, pe);
|
1188
|
+
ary_push(parser->line_buffer, token->type);
|
1189
|
+
}
|
1190
|
+
else
|
1191
|
+
// already scanning
|
1192
|
+
NEXT_TOKEN();
|
1193
|
+
}
|
1194
|
+
int type = token->type;
|
1195
|
+
|
1196
|
+
// can't declare new variables inside a switch statement, so predeclare them here
|
1197
|
+
long remove_strong = -1;
|
1198
|
+
long remove_em = -1;
|
1199
|
+
|
1200
|
+
// general purpose counters, flags and pointers
|
1201
|
+
long i = 0;
|
1202
|
+
long j = 0;
|
1203
|
+
long k = 0;
|
1204
|
+
str_t *output = NULL;
|
1205
|
+
str_t _token_str;
|
1206
|
+
str_t *token_str = &_token_str;
|
1207
|
+
|
1208
|
+
// The following giant switch statement contains cases for all the possible token types.
|
1209
|
+
// In the most basic sense we are emitting the HTML that corresponds to each token,
|
1210
|
+
// but some tokens require context information in order to decide what to output.
|
1211
|
+
// For example, does the STRONG token (''') translate to <strong> or </strong>?
|
1212
|
+
// So when looking at any given token we have three state-maintaining variables which gives us a notion of "where we are":
|
1213
|
+
//
|
1214
|
+
// - the "scope" stack (indicates what HTML DOM structures we are currently nested inside, similar to a CSS selector)
|
1215
|
+
// - the line buffer (records tokens seen so far on the current line)
|
1216
|
+
// - the line "scope" stack (indicates what the scope should be based only on what is visible on the line so far)
|
1217
|
+
//
|
1218
|
+
// Although this is fairly complicated, there is one key simplifying factor:
|
1219
|
+
// The translator continuously performs auto-correction, and this means that we always have a guarantee that the
|
1220
|
+
// scope stack (up to the current token) is valid; our translator can take this as a given.
|
1221
|
+
// Auto-correction basically consists of inserting missing tokens (preventing subsquent HTML from being messed up),
|
1222
|
+
// or converting illegal (unexpected) tokens to their plain text equivalents (providing visual feedback to Wikitext author).
|
1223
|
+
switch (type)
|
1224
|
+
{
|
1225
|
+
case PRE:
|
1226
|
+
if (IN_EITHER_OF(NO_WIKI_START, PRE_START))
|
1227
|
+
{
|
1228
|
+
str_append(parser->output, space, sizeof(space) - 1);
|
1229
|
+
break;
|
1230
|
+
}
|
1231
|
+
else if (IN(BLOCKQUOTE_START))
|
1232
|
+
{
|
1233
|
+
// this kind of nesting not allowed (to avoid user confusion)
|
1234
|
+
wiki_pop_excess_elements(parser);
|
1235
|
+
wiki_start_para_if_necessary(parser);
|
1236
|
+
output = parser->capture ? parser->capture : parser->output;
|
1237
|
+
str_append(output, space, sizeof(space) - 1);
|
1238
|
+
break;
|
1239
|
+
}
|
1240
|
+
|
1241
|
+
// count number of BLOCKQUOTE tokens in line buffer and in scope stack
|
1242
|
+
ary_push(parser->line, PRE);
|
1243
|
+
i = ary_count(parser->line, BLOCKQUOTE);
|
1244
|
+
j = ary_count(parser->scope, BLOCKQUOTE);
|
1245
|
+
if (i < j)
|
1246
|
+
{
|
1247
|
+
// must pop (reduce nesting level)
|
1248
|
+
for (i = j - i; i > 0; i--)
|
1249
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE, true);
|
1250
|
+
}
|
1251
|
+
|
1252
|
+
if (!IN(PRE))
|
1253
|
+
{
|
1254
|
+
parser->pending_crlf = false;
|
1255
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE, false);
|
1256
|
+
wiki_indent(parser);
|
1257
|
+
str_append(parser->output, pre_start, sizeof(pre_start) - 1);
|
1258
|
+
ary_push(parser->scope, PRE);
|
1259
|
+
}
|
1260
|
+
break;
|
1261
|
+
|
1262
|
+
case PRE_START:
|
1263
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1264
|
+
{
|
1265
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1266
|
+
str_append(parser->output, escaped_pre_start, sizeof(escaped_pre_start) - 1);
|
1267
|
+
}
|
1268
|
+
else if (IN(BLOCKQUOTE_START))
|
1269
|
+
{
|
1270
|
+
wiki_rollback_failed_link(parser); // if any
|
1271
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE_START, false);
|
1272
|
+
wiki_append_pre_start(parser, token);
|
1273
|
+
}
|
1274
|
+
else if (IN(BLOCKQUOTE))
|
1275
|
+
{
|
1276
|
+
if (token->column_start == 1) // only allowed in first column
|
1277
|
+
{
|
1278
|
+
wiki_rollback_failed_link(parser); // if any
|
1279
|
+
wiki_pop_all_from_stack(parser);
|
1280
|
+
wiki_append_pre_start(parser, token);
|
1281
|
+
}
|
1282
|
+
else // PRE_START illegal here
|
1283
|
+
{
|
1284
|
+
output = parser->capture ? parser->capture : parser->output;
|
1285
|
+
wiki_pop_excess_elements(parser);
|
1286
|
+
wiki_start_para_if_necessary(parser);
|
1287
|
+
str_append(output, escaped_pre_start, sizeof(escaped_pre_start) - 1);
|
1288
|
+
}
|
1289
|
+
}
|
1290
|
+
else
|
1291
|
+
{
|
1292
|
+
wiki_rollback_failed_link(parser); // if any
|
1293
|
+
wiki_pop_from_stack_up_to(parser, NULL, P, true);
|
1294
|
+
wiki_append_pre_start(parser, token);
|
1295
|
+
}
|
1296
|
+
break;
|
1297
|
+
|
1298
|
+
case PRE_END:
|
1299
|
+
if (IN_EITHER_OF(NO_WIKI_START, PRE))
|
1300
|
+
{
|
1301
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1302
|
+
str_append(parser->output, escaped_pre_end, sizeof(escaped_pre_end) - 1);
|
1303
|
+
}
|
1304
|
+
else
|
1305
|
+
{
|
1306
|
+
if (IN(PRE_START))
|
1307
|
+
wiki_pop_from_stack_up_to(parser, parser->output, PRE_START, true);
|
1308
|
+
else
|
1309
|
+
{
|
1310
|
+
output = parser->capture ? parser->capture : parser->output;
|
1311
|
+
wiki_pop_excess_elements(parser);
|
1312
|
+
wiki_start_para_if_necessary(parser);
|
1313
|
+
str_append(output, escaped_pre_end, sizeof(escaped_pre_end) - 1);
|
1314
|
+
}
|
1315
|
+
}
|
1316
|
+
break;
|
1317
|
+
|
1318
|
+
case BLOCKQUOTE:
|
1319
|
+
if (IN_EITHER_OF(NO_WIKI_START, PRE_START))
|
1320
|
+
// no need to check for <pre>; can never appear inside it
|
1321
|
+
str_append(parser->output, escaped_blockquote, TOKEN_LEN(token) + 3); // will either emit ">" or "> "
|
1322
|
+
else if (IN(BLOCKQUOTE_START))
|
1323
|
+
{
|
1324
|
+
// this kind of nesting not allowed (to avoid user confusion)
|
1325
|
+
wiki_pop_excess_elements(parser);
|
1326
|
+
wiki_start_para_if_necessary(parser);
|
1327
|
+
output = parser->capture ? parser->capture : parser->output;
|
1328
|
+
str_append(output, escaped_blockquote, TOKEN_LEN(token) + 3); // will either emit ">" or "> "
|
1329
|
+
break;
|
1330
|
+
}
|
1331
|
+
else
|
1332
|
+
{
|
1333
|
+
ary_push(parser->line, BLOCKQUOTE);
|
1334
|
+
|
1335
|
+
// count number of BLOCKQUOTE tokens in line buffer and in scope stack
|
1336
|
+
i = ary_count(parser->line, BLOCKQUOTE);
|
1337
|
+
j = ary_count(parser->scope, BLOCKQUOTE);
|
1338
|
+
|
1339
|
+
// given that BLOCKQUOTE tokens can be nested, peek ahead and see if there are any more which might affect the decision to push or pop
|
1340
|
+
while (NEXT_TOKEN(), (token->type == BLOCKQUOTE))
|
1341
|
+
{
|
1342
|
+
ary_push(parser->line, BLOCKQUOTE);
|
1343
|
+
i++;
|
1344
|
+
}
|
1345
|
+
|
1346
|
+
// now decide whether to push, pop or do nothing
|
1347
|
+
if (i > j)
|
1348
|
+
{
|
1349
|
+
// must push (increase nesting level)
|
1350
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE, false);
|
1351
|
+
for (i = i - j; i > 0; i--)
|
1352
|
+
{
|
1353
|
+
wiki_indent(parser);
|
1354
|
+
str_append(parser->output, blockquote_start, sizeof(blockquote_start) - 1);
|
1355
|
+
str_append_str(parser->output, parser->line_ending);
|
1356
|
+
ary_push(parser->scope, BLOCKQUOTE);
|
1357
|
+
}
|
1358
|
+
}
|
1359
|
+
else if (i < j)
|
1360
|
+
{
|
1361
|
+
// must pop (reduce nesting level)
|
1362
|
+
for (i = j - i; i > 0; i--)
|
1363
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE, true);
|
1364
|
+
}
|
1365
|
+
|
1366
|
+
// jump to top of the loop to process token we scanned during lookahead
|
1367
|
+
continue;
|
1368
|
+
}
|
1369
|
+
break;
|
1370
|
+
|
1371
|
+
case BLOCKQUOTE_START:
|
1372
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1373
|
+
{
|
1374
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1375
|
+
str_append(parser->output, escaped_blockquote_start, sizeof(escaped_blockquote_start) - 1);
|
1376
|
+
}
|
1377
|
+
else if (IN(BLOCKQUOTE_START))
|
1378
|
+
{
|
1379
|
+
// nesting is fine here
|
1380
|
+
wiki_rollback_failed_link(parser); // if any
|
1381
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE_START, false);
|
1382
|
+
wiki_indent(parser);
|
1383
|
+
str_append(parser->output, blockquote_start, sizeof(blockquote_start) - 1);
|
1384
|
+
str_append_str(parser->output, parser->line_ending);
|
1385
|
+
ary_push(parser->scope, BLOCKQUOTE_START);
|
1386
|
+
ary_push(parser->line, BLOCKQUOTE_START);
|
1387
|
+
}
|
1388
|
+
else if (IN(BLOCKQUOTE))
|
1389
|
+
{
|
1390
|
+
if (token->column_start == 1) // only allowed in first column
|
1391
|
+
{
|
1392
|
+
wiki_rollback_failed_link(parser); // if any
|
1393
|
+
wiki_pop_all_from_stack(parser);
|
1394
|
+
wiki_indent(parser);
|
1395
|
+
str_append(parser->output, blockquote_start, sizeof(blockquote_start) - 1);
|
1396
|
+
str_append_str(parser->output, parser->line_ending);
|
1397
|
+
ary_push(parser->scope, BLOCKQUOTE_START);
|
1398
|
+
ary_push(parser->line, BLOCKQUOTE_START);
|
1399
|
+
}
|
1400
|
+
else // BLOCKQUOTE_START illegal here
|
1401
|
+
{
|
1402
|
+
output = parser->capture ? parser->capture : parser->output;
|
1403
|
+
wiki_pop_excess_elements(parser);
|
1404
|
+
wiki_start_para_if_necessary(parser);
|
1405
|
+
str_append(output, escaped_blockquote_start, sizeof(escaped_blockquote_start) - 1);
|
1406
|
+
}
|
1407
|
+
}
|
1408
|
+
else
|
1409
|
+
{
|
1410
|
+
// would be nice to eliminate the repetition here but it's probably the clearest way
|
1411
|
+
wiki_rollback_failed_link(parser); // if any
|
1412
|
+
wiki_pop_from_stack_up_to(parser, NULL, P, true);
|
1413
|
+
wiki_indent(parser);
|
1414
|
+
str_append(parser->output, blockquote_start, sizeof(blockquote_start) - 1);
|
1415
|
+
str_append_str(parser->output, parser->line_ending);
|
1416
|
+
ary_push(parser->scope, BLOCKQUOTE_START);
|
1417
|
+
ary_push(parser->line, BLOCKQUOTE_START);
|
1418
|
+
}
|
1419
|
+
break;
|
1420
|
+
|
1421
|
+
case BLOCKQUOTE_END:
|
1422
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1423
|
+
{
|
1424
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1425
|
+
str_append(parser->output, escaped_blockquote_end, sizeof(escaped_blockquote_end) - 1);
|
1426
|
+
}
|
1427
|
+
else
|
1428
|
+
{
|
1429
|
+
if (IN(BLOCKQUOTE_START))
|
1430
|
+
wiki_pop_from_stack_up_to(parser, parser->output, BLOCKQUOTE_START, true);
|
1431
|
+
else
|
1432
|
+
{
|
1433
|
+
output = parser->capture ? parser->capture : parser->output;
|
1434
|
+
wiki_pop_excess_elements(parser);
|
1435
|
+
wiki_start_para_if_necessary(parser);
|
1436
|
+
str_append(output, escaped_blockquote_end, sizeof(escaped_blockquote_end) - 1);
|
1437
|
+
}
|
1438
|
+
}
|
1439
|
+
break;
|
1440
|
+
|
1441
|
+
case NO_WIKI_START:
|
1442
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1443
|
+
{
|
1444
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1445
|
+
str_append(parser->output, escaped_no_wiki_start, sizeof(escaped_no_wiki_start) - 1);
|
1446
|
+
}
|
1447
|
+
else
|
1448
|
+
{
|
1449
|
+
wiki_pop_excess_elements(parser);
|
1450
|
+
wiki_start_para_if_necessary(parser);
|
1451
|
+
ary_push(parser->scope, NO_WIKI_START);
|
1452
|
+
ary_push(parser->line, NO_WIKI_START);
|
1453
|
+
}
|
1454
|
+
break;
|
1455
|
+
|
1456
|
+
case NO_WIKI_END:
|
1457
|
+
if (IN(NO_WIKI_START))
|
1458
|
+
// <nowiki> should always only ever be the last item in the stack, but use the helper routine just in case
|
1459
|
+
wiki_pop_from_stack_up_to(parser, NULL, NO_WIKI_START, true);
|
1460
|
+
else
|
1461
|
+
{
|
1462
|
+
wiki_pop_excess_elements(parser);
|
1463
|
+
wiki_start_para_if_necessary(parser);
|
1464
|
+
str_append(parser->output, escaped_no_wiki_end, sizeof(escaped_no_wiki_end) - 1);
|
1465
|
+
}
|
1466
|
+
break;
|
1467
|
+
|
1468
|
+
case STRONG_EM:
|
1469
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1470
|
+
{
|
1471
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1472
|
+
str_append(parser->output, literal_strong_em, sizeof(literal_strong_em) - 1);
|
1473
|
+
break;
|
1474
|
+
}
|
1475
|
+
|
1476
|
+
output = parser->capture ? parser->capture : parser->output;
|
1477
|
+
wiki_pop_excess_elements(parser);
|
1478
|
+
|
1479
|
+
// if you've seen STRONG/STRONG_START or EM/EM_START, must close them in the reverse order that you saw them!
|
1480
|
+
// otherwise, must open them
|
1481
|
+
remove_strong = -1;
|
1482
|
+
remove_em = -1;
|
1483
|
+
j = parser->scope->count;
|
1484
|
+
for (j = j - 1; j >= 0; j--)
|
1485
|
+
{
|
1486
|
+
int val = ary_entry(parser->scope, (int)j);
|
1487
|
+
if (val == STRONG || val == STRONG_START)
|
1488
|
+
{
|
1489
|
+
str_append(output, strong_end, sizeof(strong_end) - 1);
|
1490
|
+
remove_strong = j;
|
1491
|
+
}
|
1492
|
+
else if (val == EM || val == EM_START)
|
1493
|
+
{
|
1494
|
+
str_append(output, em_end, sizeof(em_end) - 1);
|
1495
|
+
remove_em = j;
|
1496
|
+
}
|
1497
|
+
}
|
1498
|
+
|
1499
|
+
if (remove_strong > remove_em) // must remove strong first
|
1500
|
+
{
|
1501
|
+
ary_pop(parser->scope);
|
1502
|
+
if (remove_em > -1)
|
1503
|
+
ary_pop(parser->scope);
|
1504
|
+
else // there was no em to remove!, so consider this an opening em tag
|
1505
|
+
{
|
1506
|
+
str_append(output, em_start, sizeof(em_start) - 1);
|
1507
|
+
ary_push(parser->scope, EM);
|
1508
|
+
ary_push(parser->line, EM);
|
1509
|
+
}
|
1510
|
+
}
|
1511
|
+
else if (remove_em > remove_strong) // must remove em first
|
1512
|
+
{
|
1513
|
+
ary_pop(parser->scope);
|
1514
|
+
if (remove_strong > -1)
|
1515
|
+
ary_pop(parser->scope);
|
1516
|
+
else // there was no strong to remove!, so consider this an opening strong tag
|
1517
|
+
{
|
1518
|
+
str_append(output, strong_start, sizeof(strong_start) - 1);
|
1519
|
+
ary_push(parser->scope, STRONG);
|
1520
|
+
ary_push(parser->line, STRONG);
|
1521
|
+
}
|
1522
|
+
}
|
1523
|
+
else // no strong or em to remove, so this must be a new opening of both
|
1524
|
+
{
|
1525
|
+
wiki_start_para_if_necessary(parser);
|
1526
|
+
str_append(output, strong_em_start, sizeof(strong_em_start) - 1);
|
1527
|
+
ary_push(parser->scope, STRONG);
|
1528
|
+
ary_push(parser->line, STRONG);
|
1529
|
+
ary_push(parser->scope, EM);
|
1530
|
+
ary_push(parser->line, EM);
|
1531
|
+
}
|
1532
|
+
break;
|
1533
|
+
|
1534
|
+
case STRONG:
|
1535
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1536
|
+
{
|
1537
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1538
|
+
str_append(parser->output, literal_strong, sizeof(literal_strong) - 1);
|
1539
|
+
}
|
1540
|
+
else
|
1541
|
+
{
|
1542
|
+
output = parser->capture ? parser->capture : parser->output;
|
1543
|
+
if (IN(STRONG_START))
|
1544
|
+
// already in span started with <strong>, no choice but to emit this literally
|
1545
|
+
str_append(output, literal_strong, sizeof(literal_strong) - 1);
|
1546
|
+
else if (IN(STRONG))
|
1547
|
+
// STRONG already seen, this is a closing tag
|
1548
|
+
wiki_pop_from_stack_up_to(parser, output, STRONG, true);
|
1549
|
+
else
|
1550
|
+
{
|
1551
|
+
// this is a new opening
|
1552
|
+
wiki_pop_excess_elements(parser);
|
1553
|
+
wiki_start_para_if_necessary(parser);
|
1554
|
+
str_append(output, strong_start, sizeof(strong_start) - 1);
|
1555
|
+
ary_push(parser->scope, STRONG);
|
1556
|
+
ary_push(parser->line, STRONG);
|
1557
|
+
}
|
1558
|
+
}
|
1559
|
+
break;
|
1560
|
+
|
1561
|
+
case STRONG_START:
|
1562
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1563
|
+
{
|
1564
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1565
|
+
str_append(parser->output, escaped_strong_start, sizeof(escaped_strong_start) - 1);
|
1566
|
+
}
|
1567
|
+
else
|
1568
|
+
{
|
1569
|
+
output = parser->capture ? parser->capture : parser->output;
|
1570
|
+
if (IN_EITHER_OF(STRONG_START, STRONG))
|
1571
|
+
str_append(output, escaped_strong_start, sizeof(escaped_strong_start) - 1);
|
1572
|
+
else
|
1573
|
+
{
|
1574
|
+
wiki_pop_excess_elements(parser);
|
1575
|
+
wiki_start_para_if_necessary(parser);
|
1576
|
+
str_append(output, strong_start, sizeof(strong_start) - 1);
|
1577
|
+
ary_push(parser->scope, STRONG_START);
|
1578
|
+
ary_push(parser->line, STRONG_START);
|
1579
|
+
}
|
1580
|
+
}
|
1581
|
+
break;
|
1582
|
+
|
1583
|
+
case STRONG_END:
|
1584
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1585
|
+
{
|
1586
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1587
|
+
str_append(parser->output, escaped_strong_end, sizeof(escaped_strong_end) - 1);
|
1588
|
+
}
|
1589
|
+
else
|
1590
|
+
{
|
1591
|
+
output = parser->capture ? parser->capture : parser->output;
|
1592
|
+
if (IN(STRONG_START))
|
1593
|
+
wiki_pop_from_stack_up_to(parser, output, STRONG_START, true);
|
1594
|
+
else
|
1595
|
+
{
|
1596
|
+
// no STRONG_START in scope, so must interpret the STRONG_END without any special meaning
|
1597
|
+
wiki_pop_excess_elements(parser);
|
1598
|
+
wiki_start_para_if_necessary(parser);
|
1599
|
+
str_append(output, escaped_strong_end, sizeof(escaped_strong_end) - 1);
|
1600
|
+
}
|
1601
|
+
}
|
1602
|
+
break;
|
1603
|
+
|
1604
|
+
case EM:
|
1605
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1606
|
+
{
|
1607
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1608
|
+
str_append(parser->output, literal_em, sizeof(literal_em) - 1);
|
1609
|
+
}
|
1610
|
+
else
|
1611
|
+
{
|
1612
|
+
output = parser->capture ? parser->capture : parser->output;
|
1613
|
+
if (IN(EM_START))
|
1614
|
+
// already in span started with <em>, no choice but to emit this literally
|
1615
|
+
str_append(output, literal_em, sizeof(literal_em) - 1);
|
1616
|
+
else if (IN(EM))
|
1617
|
+
// EM already seen, this is a closing tag
|
1618
|
+
wiki_pop_from_stack_up_to(parser, output, EM, true);
|
1619
|
+
else
|
1620
|
+
{
|
1621
|
+
// this is a new opening
|
1622
|
+
wiki_pop_excess_elements(parser);
|
1623
|
+
wiki_start_para_if_necessary(parser);
|
1624
|
+
str_append(output, em_start, sizeof(em_start) - 1);
|
1625
|
+
ary_push(parser->scope, EM);
|
1626
|
+
ary_push(parser->line, EM);
|
1627
|
+
}
|
1628
|
+
}
|
1629
|
+
break;
|
1630
|
+
|
1631
|
+
case EM_START:
|
1632
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1633
|
+
{
|
1634
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1635
|
+
str_append(parser->output, escaped_em_start, sizeof(escaped_em_start) - 1);
|
1636
|
+
}
|
1637
|
+
else
|
1638
|
+
{
|
1639
|
+
output = parser->capture ? parser->capture : parser->output;
|
1640
|
+
if (IN_EITHER_OF(EM_START, EM))
|
1641
|
+
str_append(output, escaped_em_start, sizeof(escaped_em_start) - 1);
|
1642
|
+
else
|
1643
|
+
{
|
1644
|
+
wiki_pop_excess_elements(parser);
|
1645
|
+
wiki_start_para_if_necessary(parser);
|
1646
|
+
str_append(output, em_start, sizeof(em_start) - 1);
|
1647
|
+
ary_push(parser->scope, EM_START);
|
1648
|
+
ary_push(parser->line, EM_START);
|
1649
|
+
}
|
1650
|
+
}
|
1651
|
+
break;
|
1652
|
+
|
1653
|
+
case EM_END:
|
1654
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1655
|
+
{
|
1656
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1657
|
+
str_append(parser->output, escaped_em_end, sizeof(escaped_em_end) - 1);
|
1658
|
+
}
|
1659
|
+
else
|
1660
|
+
{
|
1661
|
+
output = parser->capture ? parser->capture : parser->output;
|
1662
|
+
if (IN(EM_START))
|
1663
|
+
wiki_pop_from_stack_up_to(parser, output, EM_START, true);
|
1664
|
+
else
|
1665
|
+
{
|
1666
|
+
// no EM_START in scope, so must interpret the EM_END without any special meaning
|
1667
|
+
wiki_pop_excess_elements(parser);
|
1668
|
+
wiki_start_para_if_necessary(parser);
|
1669
|
+
str_append(output, escaped_em_end, sizeof(escaped_em_end) - 1);
|
1670
|
+
}
|
1671
|
+
}
|
1672
|
+
break;
|
1673
|
+
|
1674
|
+
case TT:
|
1675
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1676
|
+
{
|
1677
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1678
|
+
str_append(parser->output, backtick, sizeof(backtick) - 1);
|
1679
|
+
}
|
1680
|
+
else
|
1681
|
+
{
|
1682
|
+
output = parser->capture ? parser->capture : parser->output;
|
1683
|
+
if (IN(TT_START))
|
1684
|
+
// already in span started with <tt>, no choice but to emit this literally
|
1685
|
+
str_append(output, backtick, sizeof(backtick) - 1);
|
1686
|
+
else if (IN(TT))
|
1687
|
+
// TT (`) already seen, this is a closing tag
|
1688
|
+
wiki_pop_from_stack_up_to(parser, output, TT, true);
|
1689
|
+
else
|
1690
|
+
{
|
1691
|
+
// this is a new opening
|
1692
|
+
wiki_pop_excess_elements(parser);
|
1693
|
+
wiki_start_para_if_necessary(parser);
|
1694
|
+
str_append(output, code_start, sizeof(code_start) - 1);
|
1695
|
+
ary_push(parser->scope, TT);
|
1696
|
+
ary_push(parser->line, TT);
|
1697
|
+
}
|
1698
|
+
}
|
1699
|
+
break;
|
1700
|
+
|
1701
|
+
case TT_START:
|
1702
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1703
|
+
{
|
1704
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1705
|
+
str_append(parser->output, escaped_tt_start, sizeof(escaped_tt_start) - 1);
|
1706
|
+
}
|
1707
|
+
else
|
1708
|
+
{
|
1709
|
+
output = parser->capture ? parser->capture : parser->output;
|
1710
|
+
if (IN_EITHER_OF(TT_START, TT))
|
1711
|
+
str_append(output, escaped_tt_start, sizeof(escaped_tt_start) - 1);
|
1712
|
+
else
|
1713
|
+
{
|
1714
|
+
wiki_pop_excess_elements(parser);
|
1715
|
+
wiki_start_para_if_necessary(parser);
|
1716
|
+
str_append(output, code_start, sizeof(code_start) - 1);
|
1717
|
+
ary_push(parser->scope, TT_START);
|
1718
|
+
ary_push(parser->line, TT_START);
|
1719
|
+
}
|
1720
|
+
}
|
1721
|
+
break;
|
1722
|
+
|
1723
|
+
case TT_END:
|
1724
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1725
|
+
{
|
1726
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1727
|
+
str_append(parser->output, escaped_tt_end, sizeof(escaped_tt_end) - 1);
|
1728
|
+
}
|
1729
|
+
else
|
1730
|
+
{
|
1731
|
+
output = parser->capture ? parser->capture : parser->output;
|
1732
|
+
if (IN(TT_START))
|
1733
|
+
wiki_pop_from_stack_up_to(parser, output, TT_START, true);
|
1734
|
+
else
|
1735
|
+
{
|
1736
|
+
// no TT_START in scope, so must interpret the TT_END without any special meaning
|
1737
|
+
wiki_pop_excess_elements(parser);
|
1738
|
+
wiki_start_para_if_necessary(parser);
|
1739
|
+
str_append(output, escaped_tt_end, sizeof(escaped_tt_end) - 1);
|
1740
|
+
}
|
1741
|
+
}
|
1742
|
+
break;
|
1743
|
+
|
1744
|
+
case OL:
|
1745
|
+
case UL:
|
1746
|
+
if (IN_EITHER_OF(NO_WIKI_START, PRE_START))
|
1747
|
+
{
|
1748
|
+
// no need to check for PRE; can never appear inside it
|
1749
|
+
str_append(parser->output, token->start, TOKEN_LEN(token));
|
1750
|
+
break;
|
1751
|
+
}
|
1752
|
+
|
1753
|
+
// count number of tokens in line and scope stacks
|
1754
|
+
int bq_count = ary_count(parser->scope, BLOCKQUOTE_START);
|
1755
|
+
i = parser->line->count - ary_count(parser->line, BLOCKQUOTE_START);
|
1756
|
+
j = parser->scope->count - bq_count;
|
1757
|
+
k = i;
|
1758
|
+
|
1759
|
+
// list tokens can be nested so look ahead for any more which might affect the decision to push or pop
|
1760
|
+
for (;;)
|
1761
|
+
{
|
1762
|
+
type = token->type;
|
1763
|
+
if (type == OL || type == UL)
|
1764
|
+
{
|
1765
|
+
token = NULL;
|
1766
|
+
if (i - k >= 2) // already seen at least one OL or UL
|
1767
|
+
{
|
1768
|
+
ary_push(parser->line, NESTED_LIST); // which means this is a nested list
|
1769
|
+
i += 3;
|
1770
|
+
}
|
1771
|
+
else
|
1772
|
+
i += 2;
|
1773
|
+
ary_push(parser->line, type);
|
1774
|
+
ary_push(parser->line, LI);
|
1775
|
+
|
1776
|
+
// want to compare line with scope but can only do so if scope has enough items on it
|
1777
|
+
if (j >= i)
|
1778
|
+
{
|
1779
|
+
if (ary_entry(parser->scope, (int)(i + bq_count - 2)) == type &&
|
1780
|
+
ary_entry(parser->scope, (int)(i + bq_count - 1)) == LI)
|
1781
|
+
{
|
1782
|
+
// line and scope match at this point: do nothing yet
|
1783
|
+
}
|
1784
|
+
else
|
1785
|
+
{
|
1786
|
+
// item just pushed onto line does not match corresponding slot of scope!
|
1787
|
+
for (; j >= i - 2; j--)
|
1788
|
+
// must pop back before emitting
|
1789
|
+
wiki_pop_from_stack(parser, NULL);
|
1790
|
+
|
1791
|
+
// will emit UL or OL, then LI
|
1792
|
+
break;
|
1793
|
+
}
|
1794
|
+
}
|
1795
|
+
else // line stack size now exceeds scope stack size: must increase nesting level
|
1796
|
+
break; // will emit UL or OL, then LI
|
1797
|
+
}
|
1798
|
+
else
|
1799
|
+
{
|
1800
|
+
// not a OL or UL token!
|
1801
|
+
if (j == i)
|
1802
|
+
// must close existing LI and re-open new one
|
1803
|
+
wiki_pop_from_stack(parser, NULL);
|
1804
|
+
else if (j > i)
|
1805
|
+
{
|
1806
|
+
// item just pushed onto line does not match corresponding slot of scope!
|
1807
|
+
for (; j >= i; j--)
|
1808
|
+
// must pop back before emitting
|
1809
|
+
wiki_pop_from_stack(parser, NULL);
|
1810
|
+
}
|
1811
|
+
break;
|
1812
|
+
}
|
1813
|
+
NEXT_TOKEN();
|
1814
|
+
}
|
1815
|
+
|
1816
|
+
// will emit
|
1817
|
+
if (type == OL || type == UL)
|
1818
|
+
{
|
1819
|
+
// if LI is at the top of a stack this is the start of a nested list
|
1820
|
+
if (j > 0 && ary_entry(parser->scope, -1) == LI)
|
1821
|
+
{
|
1822
|
+
// so we should precede it with a CRLF, and indicate that it's a nested list
|
1823
|
+
str_append(parser->output, parser->line_ending->ptr, parser->line_ending->len);
|
1824
|
+
ary_push(parser->scope, NESTED_LIST);
|
1825
|
+
}
|
1826
|
+
else
|
1827
|
+
{
|
1828
|
+
// this is a new list
|
1829
|
+
if (IN(BLOCKQUOTE_START))
|
1830
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE_START, false);
|
1831
|
+
else
|
1832
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE, false);
|
1833
|
+
}
|
1834
|
+
|
1835
|
+
// emit
|
1836
|
+
wiki_indent(parser);
|
1837
|
+
if (type == OL)
|
1838
|
+
str_append(parser->output, ol_start, sizeof(ol_start) - 1);
|
1839
|
+
else if (type == UL)
|
1840
|
+
str_append(parser->output, ul_start, sizeof(ul_start) - 1);
|
1841
|
+
ary_push(parser->scope, type);
|
1842
|
+
str_append(parser->output, parser->line_ending->ptr, parser->line_ending->len);
|
1843
|
+
}
|
1844
|
+
else if (type == SPACE)
|
1845
|
+
// silently throw away the optional SPACE token after final list marker
|
1846
|
+
token = NULL;
|
1847
|
+
|
1848
|
+
wiki_indent(parser);
|
1849
|
+
str_append(parser->output, li_start, sizeof(li_start) - 1);
|
1850
|
+
ary_push(parser->scope, LI);
|
1851
|
+
|
1852
|
+
// any subsequent UL or OL tokens on this line are syntax errors and must be emitted literally
|
1853
|
+
if (type == OL || type == UL)
|
1854
|
+
{
|
1855
|
+
k = 0;
|
1856
|
+
while (k++, NEXT_TOKEN(), (type = token->type))
|
1857
|
+
{
|
1858
|
+
if (type == OL || type == UL)
|
1859
|
+
str_append(parser->output, token->start, TOKEN_LEN(token));
|
1860
|
+
else if (type == SPACE && k == 1)
|
1861
|
+
{
|
1862
|
+
// silently throw away the optional SPACE token after final list marker
|
1863
|
+
token = NULL;
|
1864
|
+
break;
|
1865
|
+
}
|
1866
|
+
else
|
1867
|
+
break;
|
1868
|
+
}
|
1869
|
+
}
|
1870
|
+
|
1871
|
+
// jump to top of the loop to process token we scanned during lookahead
|
1872
|
+
continue;
|
1873
|
+
|
1874
|
+
case H6_START:
|
1875
|
+
case H5_START:
|
1876
|
+
case H4_START:
|
1877
|
+
case H3_START:
|
1878
|
+
case H2_START:
|
1879
|
+
case H1_START:
|
1880
|
+
if (IN_EITHER_OF(NO_WIKI_START, PRE_START))
|
1881
|
+
{
|
1882
|
+
// no need to check for PRE; can never appear inside it
|
1883
|
+
str_append(parser->output, token->start, TOKEN_LEN(token));
|
1884
|
+
break;
|
1885
|
+
}
|
1886
|
+
|
1887
|
+
// pop up to but not including the last BLOCKQUOTE on the scope stack
|
1888
|
+
if (IN(BLOCKQUOTE_START))
|
1889
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE_START, false);
|
1890
|
+
else
|
1891
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE, false);
|
1892
|
+
|
1893
|
+
// count number of BLOCKQUOTE tokens in line buffer and in scope stack
|
1894
|
+
ary_push(parser->line, type);
|
1895
|
+
i = ary_count(parser->line, BLOCKQUOTE);
|
1896
|
+
j = ary_count(parser->scope, BLOCKQUOTE);
|
1897
|
+
|
1898
|
+
// decide whether we need to pop off excess BLOCKQUOTE tokens (will never need to push; that is handled above in the BLOCKQUOTE case itself)
|
1899
|
+
if (i < j)
|
1900
|
+
{
|
1901
|
+
// must pop (reduce nesting level)
|
1902
|
+
for (i = j - i; i > 0; i--)
|
1903
|
+
wiki_pop_from_stack_up_to(parser, NULL, BLOCKQUOTE, true);
|
1904
|
+
}
|
1905
|
+
|
1906
|
+
// discard any whitespace here (so that "== foo ==" will be translated to "<h2>foo</h2>" rather than "<h2> foo </h2")
|
1907
|
+
while (NEXT_TOKEN(), (token->type == SPACE))
|
1908
|
+
; // discard
|
1909
|
+
|
1910
|
+
ary_push(parser->scope, type);
|
1911
|
+
wiki_indent(parser);
|
1912
|
+
|
1913
|
+
// take base_heading_level into account
|
1914
|
+
type += base_heading_level;
|
1915
|
+
if (type > H6_START) // no need to check for underflow (base_heading_level never negative)
|
1916
|
+
type = H6_START;
|
1917
|
+
|
1918
|
+
// rather than repeat all that code for each kind of heading, share it and use a conditional here
|
1919
|
+
if (type == H6_START)
|
1920
|
+
str_append(parser->output, h6_start, sizeof(h6_start) - 1);
|
1921
|
+
else if (type == H5_START)
|
1922
|
+
str_append(parser->output, h5_start, sizeof(h5_start) - 1);
|
1923
|
+
else if (type == H4_START)
|
1924
|
+
str_append(parser->output, h4_start, sizeof(h4_start) - 1);
|
1925
|
+
else if (type == H3_START)
|
1926
|
+
str_append(parser->output, h3_start, sizeof(h3_start) - 1);
|
1927
|
+
else if (type == H2_START)
|
1928
|
+
str_append(parser->output, h2_start, sizeof(h2_start) - 1);
|
1929
|
+
else if (type == H1_START)
|
1930
|
+
str_append(parser->output, h1_start, sizeof(h1_start) - 1);
|
1931
|
+
|
1932
|
+
// jump to top of the loop to process token we scanned during lookahead
|
1933
|
+
continue;
|
1934
|
+
|
1935
|
+
case H6_END:
|
1936
|
+
case H5_END:
|
1937
|
+
case H4_END:
|
1938
|
+
case H3_END:
|
1939
|
+
case H2_END:
|
1940
|
+
case H1_END:
|
1941
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1942
|
+
{
|
1943
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1944
|
+
str_append(parser->output, token->start, TOKEN_LEN(token));
|
1945
|
+
}
|
1946
|
+
else
|
1947
|
+
{
|
1948
|
+
wiki_rollback_failed_external_link(parser); // if any
|
1949
|
+
if ((type == H6_END && !IN(H6_START)) ||
|
1950
|
+
(type == H5_END && !IN(H5_START)) ||
|
1951
|
+
(type == H4_END && !IN(H4_START)) ||
|
1952
|
+
(type == H3_END && !IN(H3_START)) ||
|
1953
|
+
(type == H2_END && !IN(H2_START)) ||
|
1954
|
+
(type == H1_END && !IN(H1_START)))
|
1955
|
+
{
|
1956
|
+
// literal output only if not in appropriate scope (we stay silent in that case)
|
1957
|
+
wiki_start_para_if_necessary(parser);
|
1958
|
+
str_append(parser->output, token->start, TOKEN_LEN(token));
|
1959
|
+
}
|
1960
|
+
}
|
1961
|
+
break;
|
1962
|
+
|
1963
|
+
case MAIL:
|
1964
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
1965
|
+
{
|
1966
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
1967
|
+
str_append(parser->output, token->start, TOKEN_LEN(token));
|
1968
|
+
}
|
1969
|
+
else if (IN(EXT_LINK_START))
|
1970
|
+
// must be capturing and this must be part of the link text
|
1971
|
+
str_append(parser->capture, token->start, TOKEN_LEN(token));
|
1972
|
+
else
|
1973
|
+
{
|
1974
|
+
wiki_pop_excess_elements(parser);
|
1975
|
+
wiki_start_para_if_necessary(parser);
|
1976
|
+
token_str->ptr = token->start;
|
1977
|
+
token_str->len = TOKEN_LEN(token);
|
1978
|
+
wiki_append_hyperlink(parser, rb_str_new2("mailto:"), token_str, NULL, mailto_class, Qnil, true);
|
1979
|
+
}
|
1980
|
+
break;
|
1981
|
+
|
1982
|
+
case URI:
|
1983
|
+
if (IN(NO_WIKI_START))
|
1984
|
+
{
|
1985
|
+
// user can temporarily suppress autolinking by using <nowiki></nowiki>
|
1986
|
+
// note that unlike MediaWiki, we do allow autolinking inside PRE blocks
|
1987
|
+
token_str->ptr = token->start;
|
1988
|
+
token_str->len = TOKEN_LEN(token);
|
1989
|
+
wiki_append_sanitized_link_target(token_str, parser->output, false);
|
1990
|
+
}
|
1991
|
+
else if (IN(LINK_START))
|
1992
|
+
{
|
1993
|
+
// if the URI were allowed it would have been handled already in LINK_START
|
1994
|
+
wiki_rollback_failed_internal_link(parser);
|
1995
|
+
token_str->ptr = token->start;
|
1996
|
+
token_str->len = TOKEN_LEN(token);
|
1997
|
+
wiki_append_hyperlink(parser, Qnil, token_str, NULL, parser->external_link_class, parser->external_link_rel, true);
|
1998
|
+
}
|
1999
|
+
else if (IN(EXT_LINK_START))
|
2000
|
+
{
|
2001
|
+
if (parser->link_target->len == 0)
|
2002
|
+
{
|
2003
|
+
// this must be our link target: look ahead to make sure we see the space we're expecting to see
|
2004
|
+
token_str->ptr = token->start;
|
2005
|
+
token_str->len = TOKEN_LEN(token);
|
2006
|
+
NEXT_TOKEN();
|
2007
|
+
if (token->type == SPACE)
|
2008
|
+
{
|
2009
|
+
ary_push(parser->scope, SPACE);
|
2010
|
+
str_append_str(parser->link_target, token_str);
|
2011
|
+
str_clear(parser->link_text);
|
2012
|
+
parser->capture = parser->link_text;
|
2013
|
+
token = NULL; // silently consume space
|
2014
|
+
}
|
2015
|
+
else
|
2016
|
+
{
|
2017
|
+
// didn't see the space! this must be an error
|
2018
|
+
wiki_pop_from_stack(parser, NULL);
|
2019
|
+
wiki_pop_excess_elements(parser);
|
2020
|
+
wiki_start_para_if_necessary(parser);
|
2021
|
+
str_append(parser->output, ext_link_start, sizeof(ext_link_start) - 1);
|
2022
|
+
wiki_append_hyperlink(parser, Qnil, token_str, NULL, parser->external_link_class, parser->external_link_rel, true);
|
2023
|
+
continue;
|
2024
|
+
}
|
2025
|
+
}
|
2026
|
+
else
|
2027
|
+
{
|
2028
|
+
token_str->ptr = token->start;
|
2029
|
+
token_str->len = TOKEN_LEN(token);
|
2030
|
+
wiki_append_sanitized_link_target(token_str, parser->link_text, false);
|
2031
|
+
}
|
2032
|
+
}
|
2033
|
+
else
|
2034
|
+
{
|
2035
|
+
wiki_pop_excess_elements(parser);
|
2036
|
+
wiki_start_para_if_necessary(parser);
|
2037
|
+
token_str->ptr = token->start;
|
2038
|
+
token_str->len = TOKEN_LEN(token);
|
2039
|
+
wiki_append_hyperlink(parser, Qnil, token_str, NULL, parser->external_link_class, parser->external_link_rel, true);
|
2040
|
+
}
|
2041
|
+
break;
|
2042
|
+
|
2043
|
+
case PATH:
|
2044
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
2045
|
+
{
|
2046
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
2047
|
+
str_append(parser->output, token->start, TOKEN_LEN(token));
|
2048
|
+
}
|
2049
|
+
else if (IN(EXT_LINK_START))
|
2050
|
+
{
|
2051
|
+
if (parser->link_target->len == 0)
|
2052
|
+
{
|
2053
|
+
// this must be our link target: look ahead to make sure we see the space we're expecting to see
|
2054
|
+
token_str->ptr = token->start;
|
2055
|
+
token_str->len = TOKEN_LEN(token);
|
2056
|
+
NEXT_TOKEN();
|
2057
|
+
if (token->type == SPACE)
|
2058
|
+
{
|
2059
|
+
ary_push(parser->scope, PATH);
|
2060
|
+
ary_push(parser->scope, SPACE);
|
2061
|
+
str_append_str(parser->link_target, token_str);
|
2062
|
+
str_clear(parser->link_text);
|
2063
|
+
parser->capture = parser->link_text;
|
2064
|
+
token = NULL; // silently consume space
|
2065
|
+
}
|
2066
|
+
else
|
2067
|
+
{
|
2068
|
+
// didn't see the space! this must be an error
|
2069
|
+
wiki_pop_from_stack(parser, NULL);
|
2070
|
+
wiki_pop_excess_elements(parser);
|
2071
|
+
wiki_start_para_if_necessary(parser);
|
2072
|
+
str_append(parser->output, ext_link_start, sizeof(ext_link_start) - 1);
|
2073
|
+
str_append_str(parser->output, token_str);
|
2074
|
+
continue;
|
2075
|
+
}
|
2076
|
+
}
|
2077
|
+
else
|
2078
|
+
str_append(parser->link_text, token->start, TOKEN_LEN(token));
|
2079
|
+
}
|
2080
|
+
else
|
2081
|
+
{
|
2082
|
+
output = parser->capture ? parser->capture : parser->output;
|
2083
|
+
wiki_pop_excess_elements(parser);
|
2084
|
+
wiki_start_para_if_necessary(parser);
|
2085
|
+
str_append(output, token->start, TOKEN_LEN(token));
|
2086
|
+
}
|
2087
|
+
break;
|
2088
|
+
|
2089
|
+
// internal links (links to other wiki articles) look like this:
|
2090
|
+
// [[another article]] (would point at, for example, "/wiki/another_article")
|
2091
|
+
// [[the other article|the link text we'll use for it]]
|
2092
|
+
// [[the other article | the link text we'll use for it]]
|
2093
|
+
// MediaWiki has strict requirements about what it will accept as a link target:
|
2094
|
+
// all wikitext markup is disallowed:
|
2095
|
+
// example [[foo ''bar'' baz]]
|
2096
|
+
// renders [[foo <em>bar</em> baz]] (ie. not a link)
|
2097
|
+
// example [[foo <em>bar</em> baz]]
|
2098
|
+
// renders [[foo <em>bar</em> baz]] (ie. not a link)
|
2099
|
+
// example [[foo <nowiki>''</nowiki> baz]]
|
2100
|
+
// renders [[foo '' baz]] (ie. not a link)
|
2101
|
+
// example [[foo <bar> baz]]
|
2102
|
+
// renders [[foo <bar> baz]] (ie. not a link)
|
2103
|
+
// HTML entities and non-ASCII, however, make it through:
|
2104
|
+
// example [[foo €]]
|
2105
|
+
// renders <a href="/wiki/Foo_%E2%82%AC">foo €</a>
|
2106
|
+
// example [[foo €]]
|
2107
|
+
// renders <a href="/wiki/Foo_%E2%82%AC">foo €</a>
|
2108
|
+
// we'll impose similar restrictions here for the link target; allowed tokens will be:
|
2109
|
+
// SPACE, SPECIAL_URI_CHARS, PRINTABLE, PATH, ALNUM, DEFAULT, QUOT and AMP
|
2110
|
+
// everything else will be rejected
|
2111
|
+
case LINK_START:
|
2112
|
+
output = parser->capture ? parser->capture : parser->output;
|
2113
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
2114
|
+
{
|
2115
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
2116
|
+
str_append(output, link_start, sizeof(link_start) - 1);
|
2117
|
+
}
|
2118
|
+
else if (IN(EXT_LINK_START))
|
2119
|
+
// already in external link scope! (and in fact, must be capturing link_text right now)
|
2120
|
+
str_append(output, link_start, sizeof(link_start) - 1);
|
2121
|
+
else if (IN(LINK_START))
|
2122
|
+
{
|
2123
|
+
// already in internal link scope! this is a syntax error
|
2124
|
+
wiki_rollback_failed_internal_link(parser);
|
2125
|
+
str_append(parser->output, link_start, sizeof(link_start) - 1);
|
2126
|
+
}
|
2127
|
+
else if (IN(SEPARATOR))
|
2128
|
+
{
|
2129
|
+
// scanning internal link text
|
2130
|
+
}
|
2131
|
+
else // not in internal link scope yet
|
2132
|
+
{
|
2133
|
+
// will either emit a link, or the rollback of a failed link, so start the para now
|
2134
|
+
wiki_pop_excess_elements(parser);
|
2135
|
+
wiki_start_para_if_necessary(parser);
|
2136
|
+
ary_push(parser->scope, LINK_START);
|
2137
|
+
|
2138
|
+
// look ahead and try to gobble up link target
|
2139
|
+
while (NEXT_TOKEN(), (type = token->type))
|
2140
|
+
{
|
2141
|
+
if (type == SPACE ||
|
2142
|
+
type == SPECIAL_URI_CHARS ||
|
2143
|
+
type == PATH ||
|
2144
|
+
type == PRINTABLE ||
|
2145
|
+
type == ALNUM ||
|
2146
|
+
type == DEFAULT ||
|
2147
|
+
type == QUOT ||
|
2148
|
+
type == QUOT_ENTITY ||
|
2149
|
+
type == AMP ||
|
2150
|
+
type == AMP_ENTITY ||
|
2151
|
+
type == IMG_START ||
|
2152
|
+
type == IMG_END ||
|
2153
|
+
type == LEFT_CURLY ||
|
2154
|
+
type == RIGHT_CURLY)
|
2155
|
+
{
|
2156
|
+
// accumulate these tokens into link_target
|
2157
|
+
if (parser->link_target->len == 0)
|
2158
|
+
{
|
2159
|
+
str_clear(parser->link_target);
|
2160
|
+
parser->capture = parser->link_target;
|
2161
|
+
}
|
2162
|
+
if (type == QUOT_ENTITY)
|
2163
|
+
// don't insert the entity, insert the literal quote
|
2164
|
+
str_append(parser->link_target, quote, sizeof(quote) - 1);
|
2165
|
+
else if (type == AMP_ENTITY)
|
2166
|
+
// don't insert the entity, insert the literal ampersand
|
2167
|
+
str_append(parser->link_target, ampersand, sizeof(ampersand) - 1);
|
2168
|
+
else
|
2169
|
+
str_append(parser->link_target, token->start, TOKEN_LEN(token));
|
2170
|
+
}
|
2171
|
+
else if (type == LINK_END)
|
2172
|
+
{
|
2173
|
+
if (parser->link_target->len == 0) // bail for inputs like "[[]]"
|
2174
|
+
wiki_rollback_failed_internal_link(parser);
|
2175
|
+
break; // jump back to top of loop (will handle this in LINK_END case below)
|
2176
|
+
}
|
2177
|
+
else if (type == SEPARATOR)
|
2178
|
+
{
|
2179
|
+
if (parser->link_target->len == 0) // bail for inputs like "[[|"
|
2180
|
+
wiki_rollback_failed_internal_link(parser);
|
2181
|
+
else
|
2182
|
+
{
|
2183
|
+
ary_push(parser->scope, SEPARATOR);
|
2184
|
+
str_clear(parser->link_text);
|
2185
|
+
parser->capture = parser->link_text;
|
2186
|
+
token = NULL;
|
2187
|
+
}
|
2188
|
+
break;
|
2189
|
+
}
|
2190
|
+
else // unexpected token (syntax error)
|
2191
|
+
{
|
2192
|
+
wiki_rollback_failed_internal_link(parser);
|
2193
|
+
break; // jump back to top of loop to handle unexpected token
|
2194
|
+
}
|
2195
|
+
}
|
2196
|
+
|
2197
|
+
// jump to top of the loop to process token we scanned during lookahead (if any)
|
2198
|
+
continue;
|
2199
|
+
}
|
2200
|
+
break;
|
2201
|
+
|
2202
|
+
case LINK_END:
|
2203
|
+
output = parser->capture ? parser->capture : parser->output;
|
2204
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
2205
|
+
{
|
2206
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
2207
|
+
str_append(output, link_end, sizeof(link_end) - 1);
|
2208
|
+
}
|
2209
|
+
else if (IN(EXT_LINK_START))
|
2210
|
+
// already in external link scope! (and in fact, must be capturing link_text right now)
|
2211
|
+
str_append(output, link_end, sizeof(link_end) - 1);
|
2212
|
+
else if (IN(LINK_START)) // in internal link scope!
|
2213
|
+
{
|
2214
|
+
if (wiki_blank(parser->link_target))
|
2215
|
+
{
|
2216
|
+
// special case for inputs like "[[ ]]"
|
2217
|
+
wiki_rollback_failed_internal_link(parser);
|
2218
|
+
str_append(parser->output, link_end, sizeof(link_end) - 1);
|
2219
|
+
break;
|
2220
|
+
}
|
2221
|
+
if (parser->link_text->len == 0 ||
|
2222
|
+
wiki_blank(parser->link_text))
|
2223
|
+
{
|
2224
|
+
// use link target as link text
|
2225
|
+
str_clear(parser->link_text);
|
2226
|
+
wiki_append_sanitized_link_target(parser->link_target, parser->link_text, true);
|
2227
|
+
}
|
2228
|
+
else
|
2229
|
+
wiki_trim_link_text(parser);
|
2230
|
+
|
2231
|
+
// perform "redlink" check before manipulating link_target
|
2232
|
+
if (NIL_P(link_proc))
|
2233
|
+
j = Qnil;
|
2234
|
+
else
|
2235
|
+
{
|
2236
|
+
j = rb_funcall(link_proc, rb_intern("call"), 1, string_from_str(parser->link_target));
|
2237
|
+
if (!NIL_P(j))
|
2238
|
+
{
|
2239
|
+
VALUE l = j; // can't cast inside StringValue macro
|
2240
|
+
j = StringValue(l);
|
2241
|
+
}
|
2242
|
+
}
|
2243
|
+
wiki_encode_link_target(parser);
|
2244
|
+
wiki_pop_from_stack_up_to(parser, output, LINK_START, true);
|
2245
|
+
parser->capture = NULL;
|
2246
|
+
wiki_append_hyperlink(parser, prefix, parser->link_target, parser->link_text, j, Qnil, false);
|
2247
|
+
str_clear(parser->link_target);
|
2248
|
+
str_clear(parser->link_text);
|
2249
|
+
}
|
2250
|
+
else // wasn't in internal link scope
|
2251
|
+
{
|
2252
|
+
wiki_pop_excess_elements(parser);
|
2253
|
+
wiki_start_para_if_necessary(parser);
|
2254
|
+
str_append(output, link_end, sizeof(link_end) - 1);
|
2255
|
+
}
|
2256
|
+
break;
|
2257
|
+
|
2258
|
+
// external links look like this:
|
2259
|
+
// [http://google.com/ the link text]
|
2260
|
+
// [/other/page/on/site see this page]
|
2261
|
+
// strings in square brackets which don't match this syntax get passed through literally; eg:
|
2262
|
+
// he was very angery [sic] about the turn of events
|
2263
|
+
case EXT_LINK_START:
|
2264
|
+
output = parser->capture ? parser->capture : parser->output;
|
2265
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
2266
|
+
{
|
2267
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
2268
|
+
str_append(output, ext_link_start, sizeof(ext_link_start) - 1);
|
2269
|
+
}
|
2270
|
+
else if (IN(EXT_LINK_START))
|
2271
|
+
// already in external link scope! (and in fact, must be capturing link_text right now)
|
2272
|
+
str_append(output, ext_link_start, sizeof(ext_link_start) - 1);
|
2273
|
+
else if (IN(LINK_START))
|
2274
|
+
{
|
2275
|
+
// already in internal link scope!
|
2276
|
+
if (parser->link_target->len == 0 || !IN(SPACE))
|
2277
|
+
str_append(parser->link_target, ext_link_start, sizeof(ext_link_start) - 1);
|
2278
|
+
else // link target has already been scanned
|
2279
|
+
str_append(parser->link_text, ext_link_start, sizeof(ext_link_start) - 1);
|
2280
|
+
}
|
2281
|
+
else // not in external link scope yet
|
2282
|
+
{
|
2283
|
+
// will either emit a link, or the rollback of a failed link, so start the para now
|
2284
|
+
wiki_pop_excess_elements(parser);
|
2285
|
+
wiki_start_para_if_necessary(parser);
|
2286
|
+
|
2287
|
+
// look ahead: expect an absolute URI (with protocol) or "relative" (path) URI
|
2288
|
+
NEXT_TOKEN();
|
2289
|
+
if (token->type == URI || token->type == PATH)
|
2290
|
+
ary_push(parser->scope, EXT_LINK_START); // so far so good, jump back to the top of the loop
|
2291
|
+
else
|
2292
|
+
// only get here if there was a syntax error (missing URI)
|
2293
|
+
str_append(parser->output, ext_link_start, sizeof(ext_link_start) - 1);
|
2294
|
+
continue; // jump back to top of loop to handle token (either URI or whatever it is)
|
2295
|
+
}
|
2296
|
+
break;
|
2297
|
+
|
2298
|
+
case EXT_LINK_END:
|
2299
|
+
output = parser->capture ? parser->capture : parser->output;
|
2300
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
2301
|
+
{
|
2302
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
2303
|
+
str_append(output, ext_link_end, sizeof(ext_link_end) - 1);
|
2304
|
+
}
|
2305
|
+
else if (IN(EXT_LINK_START))
|
2306
|
+
{
|
2307
|
+
if (parser->link_text->len == 0)
|
2308
|
+
// syntax error: external link with no link text
|
2309
|
+
wiki_rollback_failed_external_link(parser);
|
2310
|
+
else
|
2311
|
+
{
|
2312
|
+
// success!
|
2313
|
+
j = IN(PATH) ? Qnil : parser->external_link_class;
|
2314
|
+
k = IN(PATH) ? Qnil : parser->external_link_rel;
|
2315
|
+
wiki_pop_from_stack_up_to(parser, output, EXT_LINK_START, true);
|
2316
|
+
parser->capture = NULL;
|
2317
|
+
wiki_append_hyperlink(parser, Qnil, parser->link_target, parser->link_text, j, k, false);
|
2318
|
+
}
|
2319
|
+
str_clear(parser->link_target);
|
2320
|
+
str_clear(parser->link_text);
|
2321
|
+
}
|
2322
|
+
else
|
2323
|
+
{
|
2324
|
+
wiki_pop_excess_elements(parser);
|
2325
|
+
wiki_start_para_if_necessary(parser);
|
2326
|
+
str_append(parser->output, ext_link_end, sizeof(ext_link_end) - 1);
|
2327
|
+
}
|
2328
|
+
break;
|
2329
|
+
|
2330
|
+
case SEPARATOR:
|
2331
|
+
output = parser->capture ? parser->capture : parser->output;
|
2332
|
+
wiki_pop_excess_elements(parser);
|
2333
|
+
wiki_start_para_if_necessary(parser);
|
2334
|
+
str_append(output, separator, sizeof(separator) - 1);
|
2335
|
+
break;
|
2336
|
+
|
2337
|
+
case SPACE:
|
2338
|
+
output = parser->capture ? parser->capture : parser->output;
|
2339
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
2340
|
+
{
|
2341
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
2342
|
+
str_append(output, token->start, TOKEN_LEN(token));
|
2343
|
+
}
|
2344
|
+
else
|
2345
|
+
{
|
2346
|
+
// peek ahead to see next token
|
2347
|
+
char *token_ptr = token->start;
|
2348
|
+
long token_len = TOKEN_LEN(token);
|
2349
|
+
NEXT_TOKEN();
|
2350
|
+
type = token->type;
|
2351
|
+
if ((type == H6_END && IN(H6_START)) ||
|
2352
|
+
(type == H5_END && IN(H5_START)) ||
|
2353
|
+
(type == H4_END && IN(H4_START)) ||
|
2354
|
+
(type == H3_END && IN(H3_START)) ||
|
2355
|
+
(type == H2_END && IN(H2_START)) ||
|
2356
|
+
(type == H1_END && IN(H1_START)))
|
2357
|
+
{
|
2358
|
+
// will suppress emission of space (discard) if next token is a H6_END, H5_END etc and we are in the corresponding scope
|
2359
|
+
}
|
2360
|
+
else
|
2361
|
+
{
|
2362
|
+
// emit the space
|
2363
|
+
wiki_pop_excess_elements(parser);
|
2364
|
+
wiki_start_para_if_necessary(parser);
|
2365
|
+
str_append(output, token_ptr, token_len);
|
2366
|
+
}
|
2367
|
+
|
2368
|
+
// jump to top of the loop to process token we scanned during lookahead
|
2369
|
+
continue;
|
2370
|
+
}
|
2371
|
+
break;
|
2372
|
+
|
2373
|
+
case QUOT_ENTITY:
|
2374
|
+
case AMP_ENTITY:
|
2375
|
+
case NAMED_ENTITY:
|
2376
|
+
case DECIMAL_ENTITY:
|
2377
|
+
// pass these through unaltered as they are case sensitive
|
2378
|
+
output = parser->capture ? parser->capture : parser->output;
|
2379
|
+
wiki_pop_excess_elements(parser);
|
2380
|
+
wiki_start_para_if_necessary(parser);
|
2381
|
+
str_append(output, token->start, TOKEN_LEN(token));
|
2382
|
+
break;
|
2383
|
+
|
2384
|
+
case HEX_ENTITY:
|
2385
|
+
// normalize hex entities (downcase them)
|
2386
|
+
output = parser->capture ? parser->capture : parser->output;
|
2387
|
+
wiki_pop_excess_elements(parser);
|
2388
|
+
wiki_start_para_if_necessary(parser);
|
2389
|
+
str_append(output, token->start, TOKEN_LEN(token));
|
2390
|
+
wiki_downcase_bang(output->ptr + output->len - TOKEN_LEN(token), TOKEN_LEN(token));
|
2391
|
+
break;
|
2392
|
+
|
2393
|
+
case QUOT:
|
2394
|
+
output = parser->capture ? parser->capture : parser->output;
|
2395
|
+
wiki_pop_excess_elements(parser);
|
2396
|
+
wiki_start_para_if_necessary(parser);
|
2397
|
+
str_append(output, quot_entity, sizeof(quot_entity) - 1);
|
2398
|
+
break;
|
2399
|
+
|
2400
|
+
case AMP:
|
2401
|
+
output = parser->capture ? parser->capture : parser->output;
|
2402
|
+
wiki_pop_excess_elements(parser);
|
2403
|
+
wiki_start_para_if_necessary(parser);
|
2404
|
+
str_append(output, amp_entity, sizeof(amp_entity) - 1);
|
2405
|
+
break;
|
2406
|
+
|
2407
|
+
case LESS:
|
2408
|
+
output = parser->capture ? parser->capture : parser->output;
|
2409
|
+
wiki_pop_excess_elements(parser);
|
2410
|
+
wiki_start_para_if_necessary(parser);
|
2411
|
+
str_append(output, lt_entity, sizeof(lt_entity) - 1);
|
2412
|
+
break;
|
2413
|
+
|
2414
|
+
case GREATER:
|
2415
|
+
output = parser->capture ? parser->capture : parser->output;
|
2416
|
+
wiki_pop_excess_elements(parser);
|
2417
|
+
wiki_start_para_if_necessary(parser);
|
2418
|
+
str_append(output, gt_entity, sizeof(gt_entity) - 1);
|
2419
|
+
break;
|
2420
|
+
|
2421
|
+
case IMG_START:
|
2422
|
+
if (IN_ANY_OF(NO_WIKI_START, PRE, PRE_START))
|
2423
|
+
{
|
2424
|
+
wiki_emit_pending_crlf_if_necessary(parser);
|
2425
|
+
str_append(parser->output, token->start, TOKEN_LEN(token));
|
2426
|
+
}
|
2427
|
+
else if (parser->capture)
|
2428
|
+
str_append(parser->capture, token->start, TOKEN_LEN(token));
|
2429
|
+
else
|
2430
|
+
{
|
2431
|
+
// not currently capturing: will be emitting something on success or failure, so get ready
|
2432
|
+
wiki_pop_excess_elements(parser);
|
2433
|
+
wiki_start_para_if_necessary(parser);
|
2434
|
+
|
2435
|
+
// scan ahead consuming PATH, PRINTABLE, ALNUM and SPECIAL_URI_CHARS tokens
|
2436
|
+
// will cheat here and abuse the link_target capture buffer to accumulate text
|
2437
|
+
while (NEXT_TOKEN(), (type = token->type))
|
2438
|
+
{
|
2439
|
+
if (type == PATH || type == PRINTABLE || type == ALNUM || type == SPECIAL_URI_CHARS)
|
2440
|
+
str_append(parser->link_target, token->start, TOKEN_LEN(token));
|
2441
|
+
else if (type == IMG_END && parser->link_target->len > 0)
|
2442
|
+
{
|
2443
|
+
// success
|
2444
|
+
wiki_append_img(parser, parser->link_target->ptr, parser->link_target->len);
|
2445
|
+
token = NULL;
|
2446
|
+
break;
|
2447
|
+
}
|
2448
|
+
else // unexpected token or zero-length target (syntax error)
|
2449
|
+
{
|
2450
|
+
// rollback
|
2451
|
+
str_append(parser->output, literal_img_start, sizeof(literal_img_start) - 1);
|
2452
|
+
if (parser->link_target->len > 0)
|
2453
|
+
str_append(parser->output, parser->link_target->ptr, parser->link_target->len);
|
2454
|
+
break;
|
2455
|
+
}
|
2456
|
+
}
|
2457
|
+
|
2458
|
+
// jump to top of the loop to process token we scanned during lookahead
|
2459
|
+
str_clear(parser->link_target);
|
2460
|
+
continue;
|
2461
|
+
}
|
2462
|
+
break;
|
2463
|
+
|
2464
|
+
case CRLF:
|
2465
|
+
i = parser->pending_crlf;
|
2466
|
+
parser->pending_crlf = false;
|
2467
|
+
wiki_rollback_failed_link(parser); // if any
|
2468
|
+
if (IN_EITHER_OF(NO_WIKI_START, PRE_START))
|
2469
|
+
{
|
2470
|
+
ary_clear(parser->line_buffer);
|
2471
|
+
str_append_str(parser->output, parser->line_ending);
|
2472
|
+
break;
|
2473
|
+
}
|
2474
|
+
else if (IN(PRE))
|
2475
|
+
{
|
2476
|
+
// beware when BLOCKQUOTE on line buffer (not line stack!) prior to CRLF, that must be end of PRE block
|
2477
|
+
if (ary_entry(parser->line_buffer, -2) == BLOCKQUOTE)
|
2478
|
+
// don't emit in this case
|
2479
|
+
wiki_pop_from_stack_up_to(parser, parser->output, PRE, true);
|
2480
|
+
else
|
2481
|
+
{
|
2482
|
+
if (ary_entry(parser->line_buffer, -2) == PRE)
|
2483
|
+
{
|
2484
|
+
// only thing on line is the PRE: emit pending line ending (if we had one)
|
2485
|
+
if (i)
|
2486
|
+
str_append_str(parser->output, parser->line_ending);
|
2487
|
+
}
|
2488
|
+
|
2489
|
+
// clear these _before_ calling NEXT_TOKEN (NEXT_TOKEN adds to the line_buffer)
|
2490
|
+
ary_clear(parser->line);
|
2491
|
+
ary_clear(parser->line_buffer);
|
2492
|
+
|
2493
|
+
// peek ahead to see if this is definitely the end of the PRE block
|
2494
|
+
NEXT_TOKEN();
|
2495
|
+
type = token->type;
|
2496
|
+
if (type != BLOCKQUOTE && type != PRE)
|
2497
|
+
// this is definitely the end of the block, so don't emit
|
2498
|
+
wiki_pop_from_stack_up_to(parser, parser->output, PRE, true);
|
2499
|
+
else
|
2500
|
+
// potentially will emit
|
2501
|
+
parser->pending_crlf = true;
|
2502
|
+
|
2503
|
+
continue; // jump back to top of loop to handle token grabbed via lookahead
|
2504
|
+
}
|
2505
|
+
}
|
2506
|
+
else
|
2507
|
+
{
|
2508
|
+
parser->pending_crlf = true;
|
2509
|
+
|
2510
|
+
// count number of BLOCKQUOTE tokens in line buffer (can be zero) and pop back to that level
|
2511
|
+
// as a side effect, this handles any open span-level elements and unclosed blocks
|
2512
|
+
// (with special handling for P blocks and LI elements)
|
2513
|
+
i = ary_count(parser->line, BLOCKQUOTE) + ary_count(parser->scope, BLOCKQUOTE_START);
|
2514
|
+
for (j = parser->scope->count; j > i; j--)
|
2515
|
+
{
|
2516
|
+
if (parser->scope->count > 0 && ary_entry(parser->scope, -1) == LI)
|
2517
|
+
{
|
2518
|
+
parser->pending_crlf = false;
|
2519
|
+
break;
|
2520
|
+
}
|
2521
|
+
|
2522
|
+
// special handling on last iteration through the loop if the top item on the scope is a P block
|
2523
|
+
if ((j - i == 1) && ary_entry(parser->scope, -1) == P)
|
2524
|
+
{
|
2525
|
+
// if nothing or BLOCKQUOTE on line buffer (not line stack!) prior to CRLF, this must be a paragraph break
|
2526
|
+
// (note that we have to make sure we're not inside a BLOCKQUOTE_START block
|
2527
|
+
// because in those blocks BLOCKQUOTE tokens have no special meaning)
|
2528
|
+
if (NO_ITEM(ary_entry(parser->line_buffer, -2)) ||
|
2529
|
+
(ary_entry(parser->line_buffer, -2) == BLOCKQUOTE && !IN(BLOCKQUOTE_START)))
|
2530
|
+
// paragraph break
|
2531
|
+
parser->pending_crlf = false;
|
2532
|
+
else
|
2533
|
+
// not a paragraph break!
|
2534
|
+
continue;
|
2535
|
+
}
|
2536
|
+
wiki_pop_from_stack(parser, NULL);
|
2537
|
+
}
|
2538
|
+
}
|
2539
|
+
|
2540
|
+
// delete the entire contents of the line scope stack and buffer
|
2541
|
+
ary_clear(parser->line);
|
2542
|
+
ary_clear(parser->line_buffer);
|
2543
|
+
break;
|
2544
|
+
|
2545
|
+
case SPECIAL_URI_CHARS:
|
2546
|
+
case PRINTABLE:
|
2547
|
+
case ALNUM:
|
2548
|
+
case IMG_END:
|
2549
|
+
case LEFT_CURLY:
|
2550
|
+
case RIGHT_CURLY:
|
2551
|
+
output = parser->capture ? parser->capture : parser->output;
|
2552
|
+
wiki_pop_excess_elements(parser);
|
2553
|
+
wiki_start_para_if_necessary(parser);
|
2554
|
+
str_append(output, token->start, TOKEN_LEN(token));
|
2555
|
+
break;
|
2556
|
+
|
2557
|
+
case DEFAULT:
|
2558
|
+
output = parser->capture ? parser->capture : parser->output;
|
2559
|
+
wiki_pop_excess_elements(parser);
|
2560
|
+
wiki_start_para_if_necessary(parser);
|
2561
|
+
wiki_append_entity_from_utf32_char(output, token->code_point);
|
2562
|
+
break;
|
2563
|
+
|
2564
|
+
case END_OF_FILE:
|
2565
|
+
// special case for input like " foo\n " (see pre_spec.rb)
|
2566
|
+
if (IN(PRE) &&
|
2567
|
+
ary_entry(parser->line_buffer, -2) == PRE &&
|
2568
|
+
parser->pending_crlf)
|
2569
|
+
str_append(parser->output, parser->line_ending->ptr, parser->line_ending->len);
|
2570
|
+
|
2571
|
+
// close any open scopes on hitting EOF
|
2572
|
+
wiki_rollback_failed_link(parser); // if any
|
2573
|
+
wiki_pop_all_from_stack(parser);
|
2574
|
+
goto return_output; // break not enough here (want to break out of outer while loop, not inner switch statement)
|
2575
|
+
|
2576
|
+
default:
|
2577
|
+
break;
|
2578
|
+
}
|
2579
|
+
|
2580
|
+
// reset current token; forcing lexer to return another token at the top of the loop
|
2581
|
+
token = NULL;
|
2582
|
+
} while (1);
|
2583
|
+
return_output:
|
2584
|
+
// nasty hack to avoid re-allocating our return value
|
2585
|
+
str_append(parser->output, null_str, 1); // null-terminate
|
2586
|
+
len = parser->output->len - 1; // don't count null termination
|
2587
|
+
|
2588
|
+
VALUE out = rb_str_buf_new(RSTRING_EMBED_LEN_MAX + 1);
|
2589
|
+
free(RSTRING_PTR(out));
|
2590
|
+
RSTRING(out)->as.heap.aux.capa = len;
|
2591
|
+
RSTRING(out)->as.heap.ptr = parser->output->ptr;
|
2592
|
+
RSTRING(out)->as.heap.len = len;
|
2593
|
+
parser->output->ptr = NULL; // don't double-free
|
2594
|
+
return out;
|
2595
|
+
}
|