So the test was wrong :-)

.gitignore

@@ -51,7 +51,3 @@ Module.symvers
 Mkfile.old
 dkms.conf
 
-shortcodes.c
-shortcodes
-tests
-.cgreen*

Makefile

@@ -1,13 +0,0 @@
-CC=gcc
-all: test
-shortcodes.c: shortcodes.rl
-	ragel -G2 shortcodes.rl -o shortcodes.c
-tests.so: shortcodes.c tests.c
-	$(CC) -fPIC -shared -g -o $@ $^ -lbg -lcgreen
-clean:
-	rm -f shortcodes.c *.o *.so tests
-test: tests.so
-	cgreen-runner $^
-debug:
-	cgreen-debug tests.so
-.PHONY: test debug

README.md

@@ -1,5 +1,7 @@
 # Shortcodes
 
+## What it is
+
 This is a parser for the shortcode spec as explained in the Hugo docs
 and used in Hugo and Nikola. Approximately.
 
@@ -12,11 +14,47 @@ as close as practical.
 * Allocates no memory, because all strings are references to
   pieces of input.
 
-What works:
+## What works
 
 * Detect shortcodes with names
 * Standalone and matched shortcodes
 * Capture data between tags in matched shortcodes
 * Capture arguments with and without names
-* Capture values with and without quotes (with details, see TODO above)
+* Capture values with and without quotes (with details, see [TODO](TODO.md))
 
+## Building
+
+You need [Ragel](http://www.colm.net/open-source/ragel/) and a C compiler.
+
+Ragel is used to generate `shortcodes.c` out of `shortcodes.rl`.
+As a convenience there is a generated `shortcodes.c` in the repo,
+
+Then:
+
+```shell
+  cd src && make
+```
+
+To run tests:
+
+```shell
+  crystal spec
+```
+
+## Installation
+
+1. Add the dependency to your `shard.yml`:
+
+   ```yaml
+   dependencies:
+     cr-discount:
+       github: ralsina/shortcodes
+   ```
+
+2. Run `shards install`
+
+## Usage
+
+```crystal
+require "shortcodes"
+```

shard.yml

@@ -6,4 +6,7 @@ authors:
 
 crystal: 1.8.2
 
+scripts:
+  postinstall: cd src && make shortcodes.o
+
 license: MIT

spec/shortcodes_spec.cr

@@ -1,9 +1,136 @@
 require "./spec_helper"
+include Shortcodes
 
-describe Shortcodes do
+describe "Shortcodes" do
-  # TODO: Write tests
+  it "should parse empty string" do
-
+    result = parse("")
-  it "works" do
+    result.shortcodes.size.should eq 0
-    false.should eq(true)
+    result.errors.size.should eq 0
   end
+
+  it "should parse simeple shortcode" do
+    result = parse("foobar {{% shortcode %}}blah")
+    result.shortcodes.size.should eq 1
+    result.errors.size.should eq 0
+    result.shortcodes[0].name.should eq "shortcode"
+    result.shortcodes[0].matching.should eq 0
+    result.shortcodes[0].args.size.should eq 0
+    result.shortcodes[0].whole.should eq "{{% shortcode %}}"
+  end
+
+  it "should report mismatched tags" do
+    input = "foobar {{% shortcode %}}blah{{% /foo %}}"
+    result = parse(input)
+    result.shortcodes.size.should eq 1
+    result.errors.size.should eq 1
+    result.errors[0].code.should eq ERR_MISMATCHED_CLOSING_TAG
+    input[result.errors[0].position, 8].should eq "{{% /foo"
+  end
+
+  it "should report mismatched brackets" do
+    input = "foobar {{% shortcode >}}blah"
+    result = parse(input)
+    result.shortcodes.size.should eq 0
+    result.errors.size.should eq 1
+    result.errors[0].code.should eq ERR_MISMATCHED_BRACKET
+    input[result.errors[0].position, 3].should eq ">}}"
+  end
+
+  it "should accept mismatched brackets inside data are ok" do
+    input = "foobar {{% sc %}} >}}blah {{% /sc %}} "
+    result = parse(input)
+    result.shortcodes.size.should eq 1
+    result.errors.size.should eq 0
+    result.shortcodes[0].whole.should eq "{{% sc %}} >}}blah {{% /sc %}}"
+    result.shortcodes[0].data.should eq " >}}blah "
+    result.shortcodes[0].args.size.should eq 0
+  end
+
+  it "should accept mismatched brackets in qvals" do
+    input = "foobar {{% sc  \">}}blah\" %}} {{% /sc %}}"
+    result = parse(input)
+    result.shortcodes.size.should eq 1
+    result.errors.size.should eq 0
+    result.shortcodes[0].whole.should eq "{{% sc  \">}}blah\" %}} {{% /sc %}}"
+    result.shortcodes[0].args.size.should eq 1
+    result.shortcodes[0].args[0].@value.should eq ">}}blah"
+  end
+
+  it "should consider spaces in shortcodes optional" do
+    input = "foobar {{%    shortcode%}}blah"
+    result = parse(input)
+    result.shortcodes.size.should eq 1
+    result.errors.size.should eq 0
+    result.shortcodes[0].name.should eq "shortcode"
+    result.shortcodes[0].matching.should eq 0
+    result.shortcodes[0].args.size.should eq 0
+    result.shortcodes[0].whole.should eq "{{%    shortcode%}}"
+  end
+
+  it "should allow path-like names" do
+    input = "foobar {{% shortcode/foo/bar %}}blah"
+    result = parse(input)
+    result.shortcodes.size.should eq 1
+    result.errors.size.should eq 0
+    result.shortcodes[0].name.should eq "shortcode/foo/bar"
+    result.shortcodes[0].matching.should eq 0
+    result.shortcodes[0].args.size.should eq 0
+    result.shortcodes[0].whole.should eq "{{% shortcode/foo/bar %}}"
+  end
+
+  it "should parse multiple shortcodes" do
+    input = "foobar {{% shortcode %}}blah {{<sc2 >}}blahblah"
+    result = parse(input)
+    result.shortcodes.size.should eq 2
+    result.errors.size.should eq 0
+    result.shortcodes[0].name.should eq "shortcode"
+    result.shortcodes[0].matching.should eq 0
+    result.shortcodes[0].args.size.should eq 0
+    result.shortcodes[0].whole.should eq "{{% shortcode %}}"
+    result.shortcodes[1].name.should eq "sc2"
+    result.shortcodes[1].matching.should eq 0
+    result.shortcodes[1].args.size.should eq 0
+    result.shortcodes[1].whole.should eq "{{<sc2 >}}"
+  end
+
+  it "should parse matching shortcodes" do
+    input = "foobar {{% shortcode %}}blah {{% /shortcode %}} blah"
+    result = parse(input)
+    result.shortcodes.size.should eq 1
+    result.errors.size.should eq 0
+    result.shortcodes[0].name.should eq "shortcode"
+    result.shortcodes[0].matching.should eq 1
+    result.shortcodes[0].args.size.should eq 0
+    result.shortcodes[0].whole.should eq "{{% shortcode %}}blah {{% /shortcode %}}"
+    result.shortcodes[0].data.should eq "blah "
+  end
+
+  it "should parse shortcode args" do
+    input = "foobar {{% shortcode foo \"bar\" 42 bat=v1 baz=\"v2\" %}}blah"
+    result = parse(input)
+    result.shortcodes.size.should eq 1
+    result.errors.size.should eq 0
+    result.shortcodes[0].name.should eq "shortcode"
+    result.shortcodes[0].matching.should eq 0
+    result.shortcodes[0].args.size.should eq 5
+    result.shortcodes[0].args[0].name.should eq ""
+    result.shortcodes[0].args[0].value.should eq "foo"
+    result.shortcodes[0].args[1].name.should eq ""
+    result.shortcodes[0].args[1].value.should eq "bar"
+    result.shortcodes[0].args[2].name.should eq ""
+    result.shortcodes[0].args[2].value.should eq "42"
+    result.shortcodes[0].args[3].name.should eq "bat"
+    result.shortcodes[0].args[3].value.should eq "v1"
+    result.shortcodes[0].args[4].name.should eq "baz"
+    result.shortcodes[0].args[4].value.should eq "v2"
+    result.shortcodes[0].whole.should eq "{{% shortcode foo \"bar\" 42 bat=v1 baz=\"v2\" %}}"
+  end
+
+  # BUG?
+  # it "should ignore escaped shortcodes" do
+  #   input = "foobar \\{{% shortcode %}}blah"
+  #   result = parse(input)
+  #   result.shortcodes.size.should eq 0
+  #   result.errors.size.should eq 0
+  # end
 end

src/Makefile

@@ -0,0 +1,7 @@
+all: shortcodes.o
+shortcodes.c: shortcodes.rl
+	ragel -G2 shortcodes.rl -o shortcodes.c
+clean:
+	rm -f shortcodes.o
+%o: %c
+	$(CC) -g -c -o $@ $^

src/run_tests.cr

@@ -0,0 +1 @@
+require "../spec/**"

src/shortcodes.cr

@@ -1,6 +1,118 @@
-# TODO: Write documentation for `Shortcodes`
+@[Link(ldflags: "#{__DIR__}/shortcodes.o")]
-module Shortcodes
+lib LibShortcodes
-  VERSION = "0.1.0"
+  struct Chunk
+    start : UInt32
+    len : UInt32
+  end
 
-  # TODO: Put your code here
+  struct ScError
+    position : UInt32
+    code : UInt32
+  end
+
+  struct Shortcode
+    whole : Chunk
+    name : Chunk
+    data : Chunk
+    matching : LibC::Char
+    argnames : Chunk[100]
+    argvals : Chunk[100]
+    argcount : UInt32
+  end
+
+  struct ScResult
+    shortcodes : Shortcode[100]
+    sccount : UInt32
+    errors : ScError[10]
+    errcount : UInt32
+  end
+
+  fun parse(input : Pointer(LibC::Char), len : UInt32) : ScResult
+end
+
+module Shortcodes
+  struct Arg
+    property name : String = ""
+    property value : String = ""
+
+    def initialize(@name, @value)
+    end
+  end
+
+  struct Error
+    property position : UInt32
+    property code : UInt32
+
+    def initialize(@position, @code)
+    end
+  end
+
+  struct Shortcode
+    property name : String = ""
+    property data : String = ""
+    property matching : Int32 = 0
+    property args : Array(Arg) = [] of Arg
+    property whole : String = ""
+
+    def initialize(
+      @name,
+      @data,
+      @matching,
+      @args,
+      @whole
+    )
+    end
+  end
+
+  struct Result
+    property shortcodes : Array(Shortcode) = [] of Shortcode
+    property errors : Array(Error) = [] of Error
+  end
+
+  def extract(c : LibShortcodes::Chunk, s : String)
+    s[c.start, c.len]
+  end
+
+  def parse(input : String)
+    r = LibShortcodes.parse(input.to_unsafe, input.bytesize)
+    result = Result.new
+
+    (0...r.sccount).each do |i|
+      sc = r.shortcodes[i]
+      args = [] of Arg
+      (0...sc.argcount).each do |j|
+        args << Arg.new(
+          extract(sc.argnames[j], input),
+          extract(sc.argvals[j], input),
+        )
+      end
+
+      result.shortcodes << Shortcode.new(
+        extract(sc.name, input),
+        extract(sc.data, input),
+        sc.matching,
+        args,
+        extract(sc.whole, input),
+      )
+    end
+    (0...r.errcount).each do |k|
+      result.errors << Error.new(
+        r.errors[k].position,
+        r.errors[k].code,
+      )
+    end
+    result
+  end
+
+  # ### Error codes
+  #
+  # You are closing the wrong shortcode.
+  # Example:
+  # {{% foo %}}  {{% /bar %}}
+  ERR_MISMATCHED_CLOSING_TAG = 1
+
+  # You are using mismatched brackets.
+  # Example:
+  # {{% foo >}}
+  ERR_MISMATCHED_BRACKET = 2
 end

src/shortcodes.h

@@ -5,7 +5,8 @@
 // and goes on for len characters.
 struct chunk
 {
-    unsigned int start, len;
+    unsigned int start;
+    unsigned int len;
 };
 typedef struct chunk chunk;
 
@@ -56,4 +57,4 @@ Example:
 #define ERR_MISMATCHED_BRACKET 2
 
 
-sc_result parse(char *);
+sc_result parse(char *, unsigned int);

src/shortcodes.rl

@@ -110,17 +110,19 @@
 }%%
 
 
-sc_result parse(char *input) {
+sc_result parse(char *input, unsigned int len) {
 
   %%write data;
-  char *eof, *ts, *te = 0;
+  char *eof = input + len;
+  char *ts, *te = 0;
   int cs, act = 0;
 
   char *start = input;
   char *p = input;
-  char *pe = p + strlen(input);
+  char *pe = p + len;
 
   sc_result result;
+  result.errcount = 0;
   shortcode *sc_list = result.sc;
   int c_sc = 0;
 

tests.c

@@ -1,226 +0,0 @@
-#include <cgreen/cgreen.h>
-#include <bglibs/str.h>
-#include "shortcodes.h"
-
-Describe(parse);
-BeforeEach(parse) {}
-AfterEach(parse) {}
-
-sc_result result;
-str s;
-
-void chunk_s(char *buffer, chunk c)
-{
-    str_copyb(&s, buffer + c.start, c.len);
-}
-
-Ensure(parse, empty_string)
-{
-    char *input = "";
-    result = parse(input);
-    // This means no shortcodes in it
-    assert_that(result.sccount, is_equal_to(0));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-Ensure(parse, simple_shortcode)
-{
-    char *input = "foobar {{% shortcode %}}blah";
-    result = parse(input);
-    // Only 1 shortcode
-    assert_that(result.sccount, is_equal_to(1));
-
-    // It's a simple one called shortcode, no args
-    chunk_s(input, result.sc[0].name);
-    assert_that(s.s, is_equal_to_string("shortcode"));
-    assert_that(result.sc[0].matching, is_equal_to(0));
-    assert_that(result.sc[0].argcount, is_equal_to(0));
-    // The whole shortcode is the whole thing
-    chunk_s(input, result.sc[0].whole);
-    assert_that(s.s, is_equal_to_string("{{% shortcode %}}"));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-Ensure(parse, mismatched_tags)
-{
-    char *input = "foobar {{% shortcode %}}blah{{% /foo %}}";
-    result = parse(input);
-    // One shortcode, one error
-    assert_that(result.sccount, is_equal_to(1));
-    assert_that(result.errcount, is_equal_to(1));
-    assert_that(result.errors[0].code, is_equal_to(ERR_MISMATCHED_CLOSING_TAG));
-    str_copyb(&s, input + result.errors[0].position, 8);
-    assert_that(s.s, is_equal_to_string("{{% /foo"));
-}
-
-Ensure(parse, mismatched_brackets)
-{
-    char *input = "foobar {{% shortcode >}}blah";
-    result = parse(input);
-    // No shortcodes, 1 error
-    assert_that(result.sccount, is_equal_to(0));
-    assert_that(result.errcount, is_equal_to(1));
-    assert_that(result.errors[0].code, is_equal_to(ERR_MISMATCHED_BRACKET));
-    str_copyb(&s, input + result.errors[0].position, 3);
-    assert_that(s.s, is_equal_to_string(">}}"));
-}
-
-Ensure(parse, mismatched_brackets_inside_data_are_ok)
-{
-    char *input = "foobar {{% sc %}} >}}blah {{% /sc %}} ";
-    result = parse(input);
-    // 1 shortcode
-    assert_that(result.sccount, is_equal_to(1));
-    chunk_s(input, result.sc[0].whole);
-    assert_that(s.s, is_equal_to_string("{{% sc %}} >}}blah {{% /sc %}}"));
-    chunk_s(input, result.sc[0].data);
-    assert_that(s.s, is_equal_to_string(" >}}blah "));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-Ensure(parse, mismatched_brackets_in_qval_are_ok)
-{
-    char *input = "foobar {{% sc  \">}}blah\" %}} {{% /sc %}}";
-    result = parse(input);
-    // 1 shortcode
-    assert_that(result.sccount, is_equal_to(1));
-    chunk_s(input, result.sc[0].whole);
-    assert_that(s.s, is_equal_to_string("{{% sc  \">}}blah\" %}} {{% /sc %}}"));
-    chunk_s(input, result.sc[0].argvals[0]);
-    assert_that(s.s, is_equal_to_string(">}}blah"));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-Ensure(parse, inner_spaces_optional)
-{
-    char *input = "foobar {{%    shortcode%}}blah";
-    result = parse(input);
-    // Only 1 shortcode
-    assert_that(result.sccount, is_equal_to(1));
-
-    // It's a simple one called shortcode, no args
-    chunk_s(input, result.sc[0].name);
-    assert_that(s.s, is_equal_to_string("shortcode"));
-    assert_that(result.sc[0].matching, is_equal_to(0));
-    assert_that(result.sc[0].argcount, is_equal_to(0));
-    // The whole shortcode is the whole thing
-    chunk_s(input, result.sc[0].whole);
-    assert_that(s.s, is_equal_to_string("{{%    shortcode%}}"));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-Ensure(parse, name_can_be_path)
-{
-    char *input = "foobar {{% shortcode/foo/bar %}}blah";
-    result = parse(input);
-    // Only 1 shortcode
-    assert_that(result.sccount, is_equal_to(1));
-
-    // It's a simple one called shortcode, no args
-    chunk_s(input, result.sc[0].name);
-    assert_that(s.s, is_equal_to_string("shortcode/foo/bar"));
-    assert_that(result.sc[0].matching, is_equal_to(0));
-    assert_that(result.sc[0].argcount, is_equal_to(0));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-Ensure(parse, multiple_shortcodes)
-{
-    char *input = "foobar {{% shortcode %}}blah {{<sc2 >}}blahblah";
-    result = parse(input);
-    // 2 shortcodes
-    assert_that(result.sccount, is_equal_to(2));
-
-    // It's a simple one called shortcode, no args
-    chunk_s(input, result.sc[0].name);
-    assert_that(s.s, is_equal_to_string("shortcode"));
-    assert_that(result.sc[0].matching, is_equal_to(0));
-    assert_that(result.sc[0].argcount, is_equal_to(0));
-    // The whole shortcode is the whole thing
-    chunk_s(input, result.sc[0].whole);
-    assert_that(s.s, is_equal_to_string("{{% shortcode %}}"));
-    // It's a simple one called sc2, no args
-    chunk_s(input, result.sc[1].name);
-    assert_that(s.s, is_equal_to_string("sc2"));
-    assert_that(result.sc[1].matching, is_equal_to(0));
-    assert_that(result.sc[1].argcount, is_equal_to(0));
-    // The whole shortcode is the whole thing
-    chunk_s(input, result.sc[1].whole);
-    assert_that(s.s, is_equal_to_string("{{<sc2 >}}"));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-Ensure(parse, matching_shortcode)
-{
-    char *input = "blah {{% shortcode %}}foo bar{{% /shortcode %}} blah";
-    result = parse(input);
-
-    // Only 1 shortcode
-    assert_that(result.sccount, is_equal_to(1));
-
-    // It's a matching one called shortcode, no args
-    chunk_s(input, result.sc[0].name);
-    assert_that(s.s, is_equal_to_string("shortcode"));
-    assert_that(result.sc[0].matching, is_equal_to(1));
-    assert_that(result.sc[0].argcount, is_equal_to(0));
-    // data is the stuff between the shortcode tags
-    chunk_s(input, result.sc[0].data);
-    assert_that(s.s, is_equal_to_string("foo bar"));
-    // The whole shortcode is the whole thing
-    chunk_s(input, result.sc[0].whole);
-    assert_that(s.s, is_equal_to_string("{{% shortcode %}}foo bar{{% /shortcode %}}"));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-Ensure(parse, shortcode_args)
-{
-    char *input = "foobar {{% shortcode foo \"bar\" 42 bat=v1 baz=\"v2\" %}}blah";
-    result = parse(input);
-    // Only 1 shortcode
-    assert_that(result.sccount, is_equal_to(1));
-
-    // The whole shortcode is the whole thing
-    chunk_s(input, result.sc[0].whole);
-    assert_that(s.s, is_equal_to_string("{{% shortcode foo \"bar\" 42 bat=v1 baz=\"v2\" %}}"));
-
-    // Name is shortcode
-    chunk_s(input, result.sc[0].name);
-    assert_that(s.s, is_equal_to_string("shortcode"));
-    assert_that(result.sc[0].matching, is_equal_to(0));
-
-    // Has 5 args
-    assert_that(result.sc[0].argcount, is_equal_to(5));
-
-    // Arg1 is foo, no name
-    assert_that(result.sc[0].argnames[0].len, is_equal_to(0));
-    chunk_s(input, result.sc[0].argvals[0]);
-    assert_that(s.s, is_equal_to_string("foo"));
-    // Arg2 is bar, no name
-    assert_that(result.sc[0].argnames[1].len, is_equal_to(0));
-    chunk_s(input, result.sc[0].argvals[1]);
-    assert_that(s.s, is_equal_to_string("bar"));
-    // Arg3 is 42, no name
-    assert_that(result.sc[0].argnames[2].len, is_equal_to(0));
-    chunk_s(input, result.sc[0].argvals[2]);
-    assert_that(s.s, is_equal_to_string("42"));
-    // Arg4 is bat=v1
-    chunk_s(input, result.sc[0].argnames[3]);
-    assert_that(s.s, is_equal_to_string("bat"));
-    chunk_s(input, result.sc[0].argvals[3]);
-    assert_that(s.s, is_equal_to_string("v1"));
-    // Arg5 is baz=v2
-    chunk_s(input, result.sc[0].argnames[4]);
-    assert_that(s.s, is_equal_to_string("baz"));
-    chunk_s(input, result.sc[0].argvals[4]);
-    assert_that(s.s, is_equal_to_string("v2"));
-    assert_that(result.errcount, is_equal_to(0));
-}
-
-// BUG?
-// Ensure(parse, escaped_shortcode)
-// {
-//     char *input = "foobar \\{{% shortcode %}}";
-//     result = parse(input);
-//     // No shortcodes
-//     assert_that(result.sc[0].name.len, is_equal_to(0));
-// }