rinku 1.4.1 → 1.5.0

data/README.markdown

@@ -23,53 +23,64 @@ Rinku is a standalone library
 
 It exports a single method called `Rinku.auto_link`.
 
-~~~~ruby
+~~~~~ruby
 require 'rinku'
 
-auto_link(text, mode=:all, link_attr=nil)
-auto_link(text, mode=:all, link_attr=nil) { |link_text| ... }
-~~~~~~~~~
+Rinku.auto_link(text, mode=:all, link_attr=nil, skip_tags=nil)
+Rinku.auto_link(text, mode=:all, link_attr=nil, skip_tags=nil) { |link_text| ... }
+~~~~~~
 
 Parses a block of text looking for "safe" urls or email addresses,
 and turns them into HTML links with the given attributes.
 
+NOTE: The block of text may or may not be HTML; if the text is HTML,
+Rinku will skip the relevant tags to prevent double-linking and linking
+inside `pre` blocks by default.
+
+NOTE: If the input text is HTML, it's expected to be already escaped.
+Rinku will perform no escaping.
+
 NOTE: Currently the follow protocols are considered safe and are the
 only ones that will be autolinked.
 
-	http:// https:// ftp:// mailto://
+    http:// https:// ftp:// mailto://
 
 Email addresses are also autolinked by default. URLs without a protocol
-specifier but starting with `www.` will also be autolinked, defaulting to
-the `http://` protocol.
+specifier but starting with 'www.' will also be autolinked, defaulting to
+the 'http://' protocol.
 
-- `text` is a string in plain text or HTML markup. If the string is formatted in
+-   `text` is a string in plain text or HTML markup. If the string is formatted in
 HTML, Rinku is smart enough to skip the links that are already enclosed in `<a>`
-tags.
+tags.`
 
-- `mode` is a symbol, either :all, :urls or :email_addresses, which specifies which
-kind of links will be auto-linked.
+-   `mode` is a symbol, either `:all`, `:urls` or `:email_addresses`, 
+which specifies which kind of links will be auto-linked. 
 
-- `link_attr` is a string containing the link attributes for each link that
+-   `link_attr` is a string containing the link attributes for each link that
 will be generated. These attributes are not sanitized and will be include as-is
 in each generated link, e.g.
 
-	~~~~ruby
-	auto_link('http://www.pokemon.com', :all, 'target="_blank"')
-	# => '<a href="http://www.pokemon.com" target="_blank">http://www.pokemon.com</a>'
-	~~~~
+     ~~~~~ruby
+     auto_link('http://www.pokemon.com', :all, 'target="_blank"')
+     # => '<a href="http://www.pokemon.com" target="_blank">http://www.pokemon.com</a>'
+     ~~~~~
 
-	This string can be autogenerated from a hash using the Rails `tag_options` helper.
+    This string can be autogenerated from a hash using the Rails `tag_options` helper.
 
-- The method takes an optional block argument. If a block is passed, it will
+-   `skip_tags` is a list of strings with the names of HTML tags that will be skipped
+when autolinking. If `nil`, this defaults to the value of the global `Rinku.skip_tags`,
+which is initially `["a", "pre", "code", "kbd", "script"]`.
+
+-   `&block` is an optional block argument. If a block is passed, it will
 be yielded for each found link in the text, and its return value will be used instead
 of the name of the link. E.g.
- 
-	~~~~ruby
- 	auto_link('Check it out at http://www.pokemon.com') do |url|
- 		"THE POKEMAN WEBSITEZ"
- 	end
- 	# => 'Check it out at <a href="http://www.pokemon.com">THE POKEMAN WEBSITEZ</a>'
-	~~~~
+
+    ~~~~~ruby
+    auto_link('Check it out at http://www.pokemon.com') do |url|
+      "THE POKEMAN WEBSITEZ"
+    end
+    # => 'Check it out at <a href="http://www.pokemon.com">THE POKEMAN WEBSITEZ</a>'
+    ~~~~~~
 
 Rinku is a drop-in replacement for Rails 3.1 `auto_link`
 ----------------------------------------------------

data/Rakefile

@@ -50,3 +50,18 @@ file package('.gem') => %w[pkg/ rinku.gemspec] + $spec.files do |f|
 end
 
 # GEMSPEC HELPERS ==========================================================
+task :gather => 'sundown/src/autolink.h' do |t|
+  files =
+    FileList[
+      'sundown/src/{buffer,autolink}.h',
+      'sundown/src/{buffer,autolink}.c',
+    ]
+  cp files, 'ext/rinku/',
+    :preserve => true,
+    :verbose => true
+end
+
+
+file 'sundown/src/autolink.h' do |t|
+  abort "The Sundown submodule is required."
+end

data/ext/rinku/autolink.c

@@ -24,9 +24,9 @@
 int
 sd_autolink_issafe(const uint8_t *link, size_t link_len)
 {
-	static const size_t valid_uris_count = 4;
+	static const size_t valid_uris_count = 5;
 	static const char *valid_uris[] = {
-		"http://", "https://", "ftp://", "mailto://"
+		"/", "http://", "https://", "ftp://", "mailto:"
 	};
 
 	size_t i;
@@ -140,10 +140,9 @@ check_domain(uint8_t *data, size_t size)
 		else if (!isalnum(data[i]) && data[i] != '-') break;
 	}
 
-	if (!isalnum(data[i - 1]) || np == 0)
-		return 0;
-
-	return i;
+	/* a valid domain needs to have at least a dot.
+	 * that's as far as we get */
+	return np ? i : 0;
 }
 
 size_t

data/ext/rinku/buffer.c

@@ -113,9 +113,35 @@ bufprintf(struct buf *buf, const char *fmt, ...)
 	if (!buf || !buf->unit)
 		return;
 
+	int n;
+
+	if (buf == 0 || (buf->size >= buf->asize && bufgrow(buf, buf->size + 1)) < 0)
+		return;
+	
 	va_start(ap, fmt);
-	vbufprintf(buf, fmt, ap);
+	n = _buf_vsnprintf((char *)buf->data + buf->size, buf->asize - buf->size, fmt, ap);
 	va_end(ap);
+
+	if (n < 0) {
+#ifdef _MSC_VER
+		n = _vscprintf(fmt, ap);
+#else
+		return;
+#endif
+	}
+
+	if ((size_t)n >= buf->asize - buf->size) {
+		if (bufgrow(buf, buf->size + n + 1) < 0)
+			return;
+		va_start(ap, fmt);
+		n = _buf_vsnprintf((char *)buf->data + buf->size, buf->asize - buf->size, fmt, ap);
+		va_end(ap);
+	}
+
+	if (n < 0)
+		return;
+
+	buf->size += n;
 }
 
 /* bufput: appends raw data to a buffer */
@@ -198,31 +224,6 @@ bufslurp(struct buf *buf, size_t len)
 void
 vbufprintf(struct buf *buf, const char *fmt, va_list ap)
 {
-	int n;
 
-	if (buf == 0 || (buf->size >= buf->asize && bufgrow(buf, buf->size + 1)) < 0)
-		return;
-
-	n = _buf_vsnprintf((char *)buf->data + buf->size, buf->asize - buf->size, fmt, ap);
-
-	if (n < 0) {
-#ifdef _MSC_VER
-		n = _vscprintf(fmt, ap);
-#else
-		return;
-#endif
-	}
-
-	if ((size_t)n >= buf->asize - buf->size) {
-		if (bufgrow(buf, buf->size + n + 1) < 0)
-			return;
-
-		n = _buf_vsnprintf((char *)buf->data + buf->size, buf->asize - buf->size, fmt, ap);
-	}
-
-	if (n < 0)
-		return;
-
-	buf->size += n;
 }
 

data/ext/rinku/buffer.h

@@ -15,8 +15,8 @@
  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  */
 
-#ifndef __GEN_BUFFER_H__
-#define __GEN_BUFFER_H__
+#ifndef BUFFER_H__
+#define BUFFER_H__
 
 #include <stddef.h>
 #include <stdarg.h>
@@ -85,7 +85,4 @@ void bufslurp(struct buf *, size_t);
 /* bufprintf: formatted printing to a buffer */
 void bufprintf(struct buf *, const char *, ...) __attribute__ ((format (printf, 2, 3)));
 
-/* vbufprintf: stdarg variant of formatted printing into a buffer */
-void vbufprintf(struct buf *, const char * , va_list);
-
 #endif

data/ext/rinku/rinku.c

@@ -116,20 +116,21 @@ autolink__skip_tag(
 	struct buf *ob,
 	const uint8_t *text,
 	size_t size,
-	const char **skip_tags,
-	size_t skip_tags_count)
+	const char **skip_tags)
 {
-	size_t tag, i = 0;
+	size_t i = 0;
 
 	while (i < size && text[i] != '>')
 		i++;
 
-	for (tag = 0; tag < skip_tags_count; ++tag) {
-		if (html_is_tag(text, size, skip_tags[tag]) == HTML_TAG_OPEN)
+	while (*skip_tags != NULL) {
+		if (html_is_tag(text, size, *skip_tags) == HTML_TAG_OPEN)
 			break;
+
+		skip_tags++;
 	}
 
-	if (tag < skip_tags_count) {
+	if (*skip_tags != NULL) {
 		for (;;) {
 			while (i < size && text[i] != '<')
 				i++;
@@ -137,7 +138,7 @@ autolink__skip_tag(
 			if (i == size)
 				break;
 
-			if (html_is_tag(text + i, size - i, skip_tags[tag]) == HTML_TAG_CLOSE)
+			if (html_is_tag(text + i, size - i, *skip_tags) == HTML_TAG_CLOSE)
 				break;
 
 			i++;
@@ -159,7 +160,6 @@ rinku_autolink(
 	unsigned int flags,
 	const char *link_attr,
 	const char **skip_tags,
-	size_t skip_tags_count,
 	void (*link_text_cb)(struct buf *ob, const struct buf *link, void *payload),
 	void *payload)
 {
@@ -212,8 +212,7 @@ rinku_autolink(
 
 		if (action == AUTOLINK_ACTION_SKIP_TAG) {
 			end += autolink__skip_tag(ob,
-				text + end, size - end,
-				skip_tags, skip_tags_count);
+				text + end, size - end, skip_tags);
 
 			continue;
 		}
@@ -266,6 +265,26 @@ autolink_callback(struct buf *link_text, const struct buf *link, void *block)
 	bufput(link_text, RSTRING_PTR(rb_link_text), RSTRING_LEN(rb_link_text));
 }
 
+const char **rinku_load_tags(VALUE rb_skip)
+{
+	const char **skip_tags;
+	size_t i, count;
+
+	Check_Type(rb_skip, T_ARRAY);
+
+	count = RARRAY_LEN(rb_skip);
+	skip_tags = xmalloc(sizeof(void *) * (count + 1));
+
+	for (i = 0; i < count; ++i) {
+		VALUE tag = rb_ary_entry(rb_skip, i);
+		Check_Type(tag, T_STRING);
+		skip_tags[i] = StringValueCStr(tag);
+	}
+
+	skip_tags[count] = NULL;
+	return skip_tags;
+}
+
 /*
  * Document-method: auto_link
  *
@@ -286,52 +305,55 @@ autolink_callback(struct buf *link_text, const struct buf *link, void *block)
  * NOTE: Currently the follow protocols are considered safe and are the
  * only ones that will be autolinked.
  *
- *		http:// https:// ftp:// mailto://
+ *     http:// https:// ftp:// mailto://
  *
  * Email addresses are also autolinked by default. URLs without a protocol
  * specifier but starting with 'www.' will also be autolinked, defaulting to
  * the 'http://' protocol.
  *
- * +text+ is a string in plain text or HTML markup. If the string is formatted in
- * HTML, Rinku is smart enough to skip the links that are already enclosed in <a>
- * tags.
+ * -   `text` is a string in plain text or HTML markup. If the string is formatted in
+ * HTML, Rinku is smart enough to skip the links that are already enclosed in `<a>`
+ * tags.`
  *
- * +mode+ is a symbol, either :all, :in_code, :urls or :email_addresses, 
- * which specifies which kind of links will be auto-linked. The :in_code symbol
- * will autolink e
+ * -   `mode` is a symbol, either `:all`, `:urls` or `:email_addresses`, 
+ * which specifies which kind of links will be auto-linked. 
  *
- * +link_attr+ is a string containing the link attributes for each link that
+ * -   `link_attr` is a string containing the link attributes for each link that
  * will be generated. These attributes are not sanitized and will be include as-is
  * in each generated link, e.g.
  *
- *	auto_link('http://www.pokemon.com', :all, 'target="_blank"')
- *	# => '<a href="http://www.pokemon.com" target="_blank">http://www.pokemon.com</a>'
+ *      ~~~~~ruby
+ *      auto_link('http://www.pokemon.com', :all, 'target="_blank"')
+ *      # => '<a href="http://www.pokemon.com" target="_blank">http://www.pokemon.com</a>'
+ *      ~~~~~
  *
- * This string can be autogenerated from a hash using the Rails `tag_options` helper.
+ *     This string can be autogenerated from a hash using the Rails `tag_options` helper.
  *
- * +skip_tags+ is a list of strings with the names of HTML tags that will be skipped
- * when autolinking. If nil, this defaults to ["a", "pre", "code", "kbd", "script"].
+ * -   `skip_tags` is a list of strings with the names of HTML tags that will be skipped
+ * when autolinking. If `nil`, this defaults to the value of the global `Rinku.skip_tags`,
+ * which is initially `["a", "pre", "code", "kbd", "script"]`.
  *
- * +block+ The method takes an optional block argument. If a block is passed, it will
+ * -   `&block` is an optional block argument. If a block is passed, it will
  * be yielded for each found link in the text, and its return value will be used instead
  * of the name of the link. E.g.
  *
- *	auto_link('Check it out at http://www.pokemon.com') do |url|
- *		"THE POKEMAN WEBSITEZ"
- *	end
- *	# => 'Check it out at <a href="http://www.pokemon.com">THE POKEMAN WEBSITEZ</a>'
+ *     ~~~~~ruby
+ *     auto_link('Check it out at http://www.pokemon.com') do |url|
+ *       "THE POKEMAN WEBSITEZ"
+ *     end
+ *     # => 'Check it out at <a href="http://www.pokemon.com">THE POKEMAN WEBSITEZ</a>'
+ *     ~~~~~~
  */
 static VALUE
 rb_rinku_autolink(int argc, VALUE *argv, VALUE self)
 {
-	static const char *SKIP_TAGS[] = {"a", "pre", "code", "kbd", "script"};
+	static const char *SKIP_TAGS[] = {"a", "pre", "code", "kbd", "script", NULL};
 
 	VALUE result, rb_text, rb_mode, rb_html, rb_skip, rb_block;
 	struct buf *output_buf;
 	int link_mode, count;
 	const char *link_attr = NULL;
 	const char **skip_tags = NULL;
-	size_t skip_tags_count;
 	ID mode_sym;
 
 	rb_scan_args(argc, argv, "13&", &rb_text, &rb_mode, &rb_html, &rb_skip, &rb_block);
@@ -350,25 +372,13 @@ rb_rinku_autolink(int argc, VALUE *argv, VALUE self)
 		link_attr = RSTRING_PTR(rb_html);
 	}
 
-	if (!NIL_P(rb_skip)) {
-		size_t i;
-
-		Check_Type(rb_skip, T_ARRAY);
-
-		skip_tags_count = RARRAY_LEN(rb_skip);
-		skip_tags = malloc(sizeof(void *) * skip_tags_count);
-		if (!skip_tags)
-			rb_raise(rb_eNoMemError, "Out of memory");
-
-		for (i = 0; i < skip_tags_count; ++i) {
-			VALUE tag = rb_ary_entry(rb_skip, i);
-			Check_Type(tag, T_STRING);
+	if (NIL_P(rb_skip))
+		rb_skip = rb_iv_get(self, "@skip_tags");
 
-			skip_tags[i] = StringValueCStr(tag);
-		}
-	} else {
+	if (NIL_P(rb_skip)) {
 		skip_tags = SKIP_TAGS;
-		skip_tags_count = sizeof(SKIP_TAGS) / sizeof(SKIP_TAGS[0]);
+	} else {
+		skip_tags = rinku_load_tags(rb_skip);
 	}
 
 	output_buf = bufnew(32);
@@ -390,7 +400,6 @@ rb_rinku_autolink(int argc, VALUE *argv, VALUE self)
 		link_mode,
 		link_attr,
 		skip_tags,
-		skip_tags_count,
 		RTEST(rb_block) ? &autolink_callback : NULL,
 		(void*)rb_block);
 
@@ -402,7 +411,7 @@ rb_rinku_autolink(int argc, VALUE *argv, VALUE self)
 	}
 
 	if (skip_tags != SKIP_TAGS)
-		free(skip_tags);
+		xfree(skip_tags);
 
 	bufrelease(output_buf);
 	return result;
@@ -414,4 +423,3 @@ void Init_rinku()
 	rb_define_method(rb_mRinku, "auto_link", rb_rinku_autolink, -1);
 }
 
-

data/lib/rails_rinku.rb

@@ -8,11 +8,16 @@ module RailsRinku
     unless args.empty?
       options[:link] = args[0] || :all
       options[:html] = args[1] || {}
+      options[:skip] = args[2]
     end
     options.reverse_merge!(:link => :all, :html => {})
     text = text.html_safe unless text.html_safe?
 
-    Rinku.auto_link(text.html_safe, options[:link], tag_options(options[:html]), &block)
+    Rinku.auto_link text,
+      options[:link],
+      tag_options(options[:html]),
+      options[:skip],
+      &block
   end
 end
 

data/lib/rinku.rb

@@ -1,5 +1,6 @@
 module Rinku
   VERSION = "1.2.2"
+  attr_accessor :skip_tags
   extend self
 end
 

data/rinku.gemspec

@@ -2,7 +2,7 @@
 
 Gem::Specification.new do |s|
   s.name = 'rinku'
-  s.version = '1.4.1'
+  s.version = '1.5.0'
   s.summary = "Mostly autolinking"
   s.description = <<-EOF
     A fast and very smart autolinking library that

data/test/autolink_test.rb

@@ -15,6 +15,26 @@ class RedcarpetAutolinkTest < Test::Unit::TestCase
     assert_equal expected, Rinku.auto_link(url)
   end
 
+  def test_global_skip_tags
+    assert_equal Rinku.skip_tags, nil
+    Rinku.skip_tags = ['pre']
+    assert_equal Rinku.skip_tags, ['pre']
+
+    Rinku.skip_tags = ['pa']
+    url = 'This is just a <pa>http://www.pokemon.com</pa> test'
+    assert_equal Rinku.auto_link(url), url
+
+    Rinku.skip_tags = nil
+    assert_not_equal Rinku.auto_link(url), url
+  end
+  
+  def test_auto_link_with_single_trailing_punctuation_and_space
+    url = "http://www.youtube.com"
+    url_result = generate_result(url)
+    assert_equal url_result, Rinku.auto_link(url)
+    assert_equal "link: #{url_result}. foo?", Rinku.auto_link("link: #{url}. foo?")
+  end
+
   def test_does_not_segfault
     assert_linked "< this is just a test", "< this is just a test"
   end

metadata

@@ -1,25 +1,32 @@
---- !ruby/object:Gem::Specification
+--- !ruby/object:Gem::Specification 
 name: rinku
-version: !ruby/object:Gem::Version
-  version: 1.4.1
+version: !ruby/object:Gem::Version 
+  hash: 3
   prerelease: 
+  segments: 
+  - 1
+  - 5
+  - 0
+  version: 1.5.0
 platform: ruby
-authors:
-- Vicent Martí
+authors: 
+- "Vicent Mart\xC3\xAD"
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-11-10 00:00:00.000000000 Z
+
+date: 2011-12-04 00:00:00 Z
 dependencies: []
-description: ! "    A fast and very smart autolinking library that\n    acts as a
-  drop-in replacement for Rails `auto_link`\n"
+
+description: "    A fast and very smart autolinking library that\n    acts as a drop-in replacement for Rails `auto_link`\n"
 email: vicent@github.com
 executables: []
-extensions:
+
+extensions: 
 - ext/rinku/extconf.rb
-extra_rdoc_files:
+extra_rdoc_files: 
 - COPYING
-files:
+files: 
 - COPYING
 - README.markdown
 - Rakefile
@@ -35,27 +42,36 @@ files:
 - test/autolink_test.rb
 homepage: http://github.com/tanoku/rinku
 licenses: []
+
 post_install_message: 
 rdoc_options: []
-require_paths:
+
+require_paths: 
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement
+required_ruby_version: !ruby/object:Gem::Requirement 
   none: false
-  requirements:
-  - - ! '>='
-    - !ruby/object:Gem::Version
-      version: '0'
-required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
   none: false
-  requirements:
-  - - ! '>='
-    - !ruby/object:Gem::Version
-      version: '0'
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
 requirements: []
+
 rubyforge_project: 
 rubygems_version: 1.8.6
 signing_key: 
 specification_version: 3
 summary: Mostly autolinking
-test_files:
+test_files: 
 - test/autolink_test.rb