-----
title: "Glyph 0.5.0 Released"
content-type: article
subtitle: "Featuring Calibre integration, macro composition, Turing-completeness, and more"
timestamp: 1314559080
tags: "glyph|ruby|opensource"
-----

<section class="section">
  <p>Too much time passed since the last Glyph release. Way too much. Finally I found the time and will to tidy up the
    last few remaining bugs, update the docs, and release it!</p>
  <p>This new release was mainly focused on extending the features of Glyph as a <em>language</em>. Besides a few
    improvements that make writing Glyph code easier and more readable (e.g. macro composition), Glyph is now
    Turing-complete. It supports iterations, recursion, variable assignments, basic arithmetics&#8230; you can even
    write a program to compute the factorial of an integer, if you wanted to.</p>
  <p>Additionally, it also features enhanced content reuse through fragments and output-independent macros, and a few
    bugfixes.</p>

  <section class="section">
    <header>
      <h1 id="h_1" class="toc">Calibre Integration</h1>
    </header>
    <p><a href="https://github.com/tammycravit">Tammy Cravit</a> proposed (and more or less implemented) an interesting
      new feature: integrate <a href="http://calibre-ebook.com/">Calibre</a> to generate ebooks in <span
        class="caps">EPUB</span> and <span class="caps">MOBI</span> format from Glyph's native standalone <span
        class="caps">HTML</span> output format.</p>
    <p>Although the support is still somewhat rough, you can, as a matter of fact, generate ebooks with Glyph, using
      Calibre.</p>

  </section>
  <section class="section">
    <header>
      <h1 id="h_2" class="toc">Macro Composition</h1>
    </header>
    <p>This release features an update at syntax-level: the possibility of &#8220;composing&#8221; macros, thereby
      eliminating nesting provided that containers take only one parameter and no attributes. What? This:</p>
    <div class="CodeRay">
      <div class="code">
        <pre><span class="line-numbers"><a href="#n1" name="n1">1</a></span>?[
<span class="line-numbers"><a href="#n2" name="n2">2</a></span>  not[output?[pdf]]|
<span class="line-numbers"><a href="#n3" name="n3">3</a></span>  ...
<span class="line-numbers"><a href="#n4" name="n4">4</a></span>]</pre>
      </div>
    </div>


    <p>Can be written like this:</p>

    <div class="CodeRay">
      <div class="code">
        <pre><span class="line-numbers"><a href="#n1" name="n1">1</a></span>?[
<span class="line-numbers"><a href="#n2" name="n2">2</a></span>  not/output?[pdf]|
<span class="line-numbers"><a href="#n3" name="n3">3</a></span>  ...
<span class="line-numbers"><a href="#n4" name="n4">4</a></span>]</pre>
      </div>
    </div>


    <p>In this case, the <code>not</code> macro was composed with the <code>output?</code> macro, thus removing one
      level of nesting.</p>
    <p>Additionally, I used this features to create an <code>xml</code> macro dispatcher that can be used to render raw
      <span class="caps">XML</span> tags, and an <code>s</code> macro dispatcher that basically is able to call nearly
      all the instance methods of the Ruby String class. So you can write things like
      <code>s/sub[This feature makes my life easier|/my/|your]</code> and similar.
    </p>

  </section>
  <section class="section">
    <header>
      <h1 id="h_3" class="toc">Turing-Completeness</h1>
    </header>
    <p>As of this version, Glyph can be considered <em>Turing-complete</em>, as it satisfies the following <a
        href="http://c2.com/cgi/wiki?LanguageRequirementsForTuringCompleteness">requirements for
        Turing-completeness</a>:</p>
    <ul>
      <li>A conditional construct, implemented via the <code>condition</code> macro.</li>
      <li>Variable assignment, by setting the value of snippets using the <code>snippet:</code> macro and of attributes
        using the <code>attribute:</code> macro.</li>
      <li>(infinite) iteration implemented through the new <code>while</code> macro or recursion, which is possible
        thanks to the new <code>define:</code> macro.</li>
      <li>A memory model which emulates an infinite store: there are no enforced limits on attribute/snippets
        allocations and number of algorithms or parameters.</li>
    </ul>
    <p>Plus, Glyph now understand basic integer arithmetic:</p>
    <div class="CodeRay">
      <div class="code">
        <pre><span class="line-numbers"><a href="#n1" name="n1">1</a></span>def:[factorial|
<span class="line-numbers"><a href="#n2" name="n2">2</a></span>  ?[
<span class="line-numbers"><a href="#n3" name="n3">3</a></span>    eq[{{0}}|0]|1|
<span class="line-numbers"><a href="#n4" name="n4">4</a></span>    multiply[
<span class="line-numbers"><a href="#n5" name="n5">5</a></span>      {{0}} | factorial[subtract[{{0}}|1]]
<span class="line-numbers"><a href="#n6" name="n6">6</a></span>    ]
<span class="line-numbers"><a href="#n7" name="n7">7</a></span>  ]
<span class="line-numbers"><a href="#n8" name="n8">8</a></span>]</pre>
      </div>
    </div>

    <p>Not that you <em>need</em> to be able to calculate factorials in your documents, but know that now you
      <em>can</em>. An you can also define lexically scoped variables, err&#8230; <em>attributes</em>, like this:
    </p>
    <div class="CodeRay">
      <div class="code">
        <pre><span class="line-numbers"><a href="#n1" name="n1">1</a></span>let[
<span class="line-numbers"><a href="#n2" name="n2">2</a></span>  @:[a|bits]
<span class="line-numbers"><a href="#n3" name="n3">3</a></span>  @:[b|bobs]
<span class="line-numbers"><a href="#n4" name="n4">4</a></span>  section[
<span class="line-numbers"><a href="#n5" name="n5">5</a></span>    @title[Something more about attributes]
<span class="line-numbers"><a href="#n6" name="n6">6</a></span>Attributes are like lexically scoped variables. You can use them to store @[a] and @[b].
<span class="line-numbers"><a href="#n7" name="n7">7</a></span>  ]
<span class="line-numbers"><a href="#n8" name="n8">8</a></span>]</pre>
      </div>
    </div>

    <p>Handy enough.</p>

  </section>
  <section class="section">
    <header>
      <h1 id="h_4" class="toc">Embeddable fragments</h1>
    </header>
    <p>Too lazy to create snippets? Feel the urge to re-use something you already wrote somewhere? Use a
      <em>fragment</em> and embed it, as follows:
    </p>
    <div class="CodeRay">
      <div class="code">
        <pre><span class="line-numbers"><a href="#n1" name="n1">1</a></span>Snippets and fragments ##[good_way|are a good way to reuse] small chunks of content, 
<span class="line-numbers"><a href="#n2" name="n2">2</a></span>while the include and load macros <span class="error">&lt;</span>=[good_way] entire files.</pre>
      </div>
    </div>

    <p>&#8230;And you can also use a new <code>load</code> macro, to embed entire files without performing any
      evaluation (like <code>include</code> does).</p>

  </section>
  <section class="section">
    <header>
      <h1 id="h_5" class="toc">Incompatibilities with previous versions</h1>
    </header>
    <p>To sum up:</p>
    <ul>
      <li><code>snippets.yml</code> is no more, define all your snippets inside your document instead.</li>
      <li>New &#8220;invisible space separator&#8221;: <code>\/</code> instead of <code>\.</code>. Because it's
        slightly prettier, nothing else.</li>
      <li>The <code>rewrite:</code> macro has been replaced by the <code>define:</code> macro, which also allows
        recursion, so be careful!</li>
      <li>If you want to render raw <span class="caps">XML</span> tags, use <code>xml/tag_name</code> instead of
        <code>=tag_name</code>.
      </li>
      <li>No more <code>match</code> macro, use <code>s/match</code> instead.</li>
    </ul>
    <p>For the full list of the issues fixed in this release, see the <a
        href="/glyph/book/changelog.html">Changelog</a>.</p>
    <p>Hope you'll enjoy this new release of Glyph. If you want to contribute, go ahead and <a
        href="https://github.com/h3rald/glyph">fork the repo</a>!</p>

  </section>

</section>