From 95c3c983e87d5ee1fb02848f1315cb65a7c6fa0c Mon Sep 17 00:00:00 2001
From: Vivien Kraus <vivien@planete-kraus.eu>
Date: Thu, 6 Oct 2022 21:02:35 +0200
Subject: Use a formatter instead of trying to strip space.

---
 mped.xml             | 160 +++++++++++----------------------------------------
 tangle-bootstrap.xsl |  76 +++---------------------
 2 files changed, 42 insertions(+), 194 deletions(-)

diff --git a/mped.xml b/mped.xml
index ecfe859..04dd33d 100644
--- a/mped.xml
+++ b/mped.xml
@@ -140,7 +140,8 @@
   <xsl:value-of select="@mped:tangle-to" />
   <xsl:text>")&#xA;</xsl:text>
 
-  <xsl:text>cat >> </xsl:text>
+  <xsl:apply-templates select="." mode="source-code-formatter" />
+  <xsl:text> > </xsl:text>
   <xsl:value-of select="@mped:tangle-to" />
   <xsl:text> &lt;&lt; "_MPED_EOF"&#xA;</xsl:text>
 
@@ -152,139 +153,45 @@
       </programlisting>
       <para>
         This template starts by creating the directory where the file
-        should go, then fills the file with the source code. For this to
-        work, we need to do two things about the text of the program
-        listing: remove the first empty lines and the last empty
-        lines of the content (but preserve indentation).
+        should go, then fills the file with the source code. XML has a
+        precise behavior when it comes to whitespace preservation, but
+        it’s not always the prettiest when we write it with whitespace
+        output in mind. So, the code output is frequently not indented
+        correctly, and has too many empty lines. To counter this
+        effect, we use a code source <emphasis>formatter</emphasis>, a
+        program that reads source code and indents it correctly. For
+        XSLT, we can use <command>xmllint</command> from
+        <application>libxml2</application>. The important thing about
+        the formatter is that it should take its input from standard
+        input and write the formatted code to standard output.
       </para>
-      <para>
-        Let us start with removing leading or trailing empty
-        lines. Removing leading empty lines seems easier.
-      </para>
-      <programlisting language="xml"
-                      xml:id="mped-private-leading-empty-lines">
-        <![CDATA[
-<xsl:template name="mped-private-leading-empty-lines">
-  <xsl:param name="indentation" select="''" />
-  <xsl:param name="text" select="." />
-  <xsl:choose>
-    <xsl:when test="substring($text, 1, 1) = '&#xA;'">
-      <xsl:call-template name="mped-private-leading-empty-lines">
-        <xsl:with-param name="indentation" select="''" />
-        <xsl:with-param name="text"
-             select="substring($text, 2)" />
-      </xsl:call-template>
-    </xsl:when>
-
-    <xsl:when test="(
-        substring($text, 1, 1) = ' '
-        or substring($text, 1, 1) = '&#9;'
-        or substring($text, 1, 1) = '&#13;')">
-      <xsl:call-template name="mped-private-leading-empty-lines">
-        <xsl:with-param name="indentation" 
-             select="concat($indentation, substring($text, 1, 1))" />
-        <xsl:with-param name="text"
-             select="substring($text, 2)" />
-      </xsl:call-template>
-    </xsl:when>
-
-    <xsl:otherwise>
-      <xsl:value-of select="$indentation" />
-      <xsl:value-of select="$text" />
-    </xsl:otherwise>
-  </xsl:choose>
-</xsl:template>
-        ]]>
-      </programlisting>
-      <para>
-        There are three different cases. If the text starts with a
-        newline, discard the indentation that we carried and the
-        newline. If the text starts with whitespace, carry it and look
-        at the next character. Otherwise, the whitespace that we carried
-        is indentation, so print it, and print the text.
-      </para>
-      <para>
-        To avoid exposing the carried indentation, it is better to mark
-        this template as internal and wrap it in a new template.
-      </para>
-      <programlisting language="xml" xml:id="remove-leading-empty-lines">
+      <programlisting language="xml" xml:id="formatter-for-xml">
         <![CDATA[
-<xsl:template name="remove-leading-empty-lines">
-  <xsl:param name="text" select="." />
-  <xsl:call-template name="mped-private-leading-empty-lines">
-    <xsl:with-param name="indentation" select="''" />
-    <xsl:with-param name="text" select="$text" />
-  </xsl:call-template>
+<xsl:template match="docbook:programlisting[@language = 'xml']" 
+              mode="source-code-formatter">
+  <xsl:text>xmllint --format -</xsl:text>
 </xsl:template>
         ]]>
       </programlisting>
       <para>
-        To remove trailing empty lines, the solution is easier since
-        there is no indentation to keep around: just discard all the
-        trailing whitespace.
+        If no formatter applies, then we can resort to
+        <command>cat</command>.
       </para>
-      <programlisting language="xml" xml:id="remove-trailing-whitespace">
+      <programlisting language="xml" xml:id="default-formatter">
         <![CDATA[
-<xsl:template name="remove-trailing-whitespace">
-  <xsl:param name="text" select="." />
-  <xsl:choose>
-    <xsl:when test="$text = ''">
-      <xsl:value-of select="$text" />
-    </xsl:when>
-    <xsl:otherwise>
-      <xsl:variable name="last" select="substring($text, string-length ($text), 1)" />
-      <xsl:variable name="before" select="substring($text, 1, string-length ($text) - 1)" />
-      <xsl:choose>
-        <xsl:when test="$last = ' ' or $last = '&#9;'
-                        or $last = '&#10;' or $last = '&#13;'">
-          <xsl:call-template name="remove-trailing-whitespace">
-            <xsl:with-param name="text" select="$before" />
-          </xsl:call-template>
-        </xsl:when>
-        <xsl:otherwise>
-          <xsl:value-of select="$text" />
-        </xsl:otherwise>
-      </xsl:choose>
-    </xsl:otherwise>
-  </xsl:choose>
+<xsl:template match="docbook:programlisting"
+              mode="source-code-formatter">
+  <xsl:text>cat</xsl:text>
 </xsl:template>
         ]]>
       </programlisting>
       <para>
-        Using these templates, we can process the program listing code
-        in the “copy-source-code” mode: if there is only one text node,
-        then remove the leading emtpy lines and trailing
-        whitespace. Otherwise, remove the leading emtpy lines from the
-        first text node and the trailing whitespace from the last text
-        node. By “first” (respectively, “last”) text node, I mean the
-        text node that has no preceding (respectively, following)
-        siblings. Maybe there are no such text nodes.
+        We also need to specify how the source code is copied.
       </para>
       <programlisting language="xml" xml:id="copy-source-code-text">
         <![CDATA[
-<xsl:template match="text()[position() = 1 and position() = last()]"
-              mode="copy-source-code">
-  <xsl:call-template name="remove-trailing-whitespace">
-    <xsl:with-param name="text">
-      <xsl:call-template name="remove-leading-empty-lines">
-        <xsl:with-param name="text" select="." />
-      </xsl:call-template>
-    </xsl:with-param>
-  </xsl:call-template>
-</xsl:template>
-
-<xsl:template match="text()[position() = 1 and position() != last()]"
-              mode="copy-source-code">
-  <xsl:call-template name="remove-leading-empty-lines">
-    <xsl:with-param name="text" select="." />
-  </xsl:call-template>
-</xsl:template>
-
-<xsl:template match="text()[position() > 1 and position() = last()]"
-              mode="copy-source-code">
-  <xsl:call-template name="remove-trailing-whitespace">
-    <xsl:with-param name="text" select="." />
-  </xsl:call-template>
+<xsl:template match="text()" mode="copy-source-code">
+  <xsl:value-of select="." />
 </xsl:template>
         ]]>
       </programlisting>
@@ -340,10 +247,12 @@
       <para>
         The collection of all these templates gives the following:
       </para>
-      <programlisting language="xml" xml:id="whole-tangling-stylesheet"
-                      mped:tangle-to="tangle.xsl">
-        <![CDATA[
-<?xml version="1.0"?>
+      <programlisting
+          language="xml"
+          xml:id="whole-tangling-stylesheet"
+          mped:tangle-to="tangle.xsl"
+          ><![CDATA[<?xml version="1.0"?>
+
 <xsl:stylesheet version="1.0"
                 xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
                 xmlns:mped="https://labo.planete-kraus.eu/mped.git"
@@ -353,9 +262,8 @@
   <xsl:strip-space elements="*" />
         ]]>
         <mped:copy linkend="tangle-programlisting" />
-        <mped:copy linkend="mped-private-leading-empty-lines" />
-        <mped:copy linkend="remove-leading-empty-lines" />
-        <mped:copy linkend="remove-trailing-whitespace" />
+        <mped:copy linkend="formatter-for-xml" />
+        <mped:copy linkend="default-formatter" />
         <mped:copy linkend="copy-source-code-text" />
         <mped:copy linkend="ignore-text-other-than-source" />
         <mped:copy linkend="tangle-mped-copy" />
diff --git a/tangle-bootstrap.xsl b/tangle-bootstrap.xsl
index ffe19f5..baa9669 100644
--- a/tangle-bootstrap.xsl
+++ b/tangle-bootstrap.xsl
@@ -7,7 +7,8 @@
     <xsl:value-of select="@mped:tangle-to"/>
     <xsl:text>")
 </xsl:text>
-    <xsl:text>cat &gt;&gt; </xsl:text>
+    <xsl:apply-templates select="." mode="source-code-formatter"/>
+    <xsl:text> &gt; </xsl:text>
     <xsl:value-of select="@mped:tangle-to"/>
     <xsl:text> &lt;&lt; "_MPED_EOF"
 </xsl:text>
@@ -16,75 +17,14 @@
 _MPED_EOF
 </xsl:text>
   </xsl:template>
-  <xsl:template name="mped-private-leading-empty-lines">
-    <xsl:param name="indentation" select="''"/>
-    <xsl:param name="text" select="."/>
-    <xsl:choose>
-      <xsl:when test="substring($text, 1, 1) = '&#10;'">
-        <xsl:call-template name="mped-private-leading-empty-lines">
-          <xsl:with-param name="indentation" select="''"/>
-          <xsl:with-param name="text" select="substring($text, 2)"/>
-        </xsl:call-template>
-      </xsl:when>
-      <xsl:when test="(         substring($text, 1, 1) = ' '         or substring($text, 1, 1) = '&#9;'         or substring($text, 1, 1) = '&#13;')">
-        <xsl:call-template name="mped-private-leading-empty-lines">
-          <xsl:with-param name="indentation" select="concat($indentation, substring($text, 1, 1))"/>
-          <xsl:with-param name="text" select="substring($text, 2)"/>
-        </xsl:call-template>
-      </xsl:when>
-      <xsl:otherwise>
-        <xsl:value-of select="$indentation"/>
-        <xsl:value-of select="$text"/>
-      </xsl:otherwise>
-    </xsl:choose>
+  <xsl:template match="docbook:programlisting[@language = 'xml']" mode="source-code-formatter">
+    <xsl:text>xmllint --format -</xsl:text>
   </xsl:template>
-  <xsl:template name="remove-leading-empty-lines">
-    <xsl:param name="text" select="."/>
-    <xsl:call-template name="mped-private-leading-empty-lines">
-      <xsl:with-param name="indentation" select="''"/>
-      <xsl:with-param name="text" select="$text"/>
-    </xsl:call-template>
+  <xsl:template match="docbook:programlisting" mode="source-code-formatter">
+    <xsl:text>cat</xsl:text>
   </xsl:template>
-  <xsl:template name="remove-trailing-whitespace">
-    <xsl:param name="text" select="."/>
-    <xsl:choose>
-      <xsl:when test="$text = ''">
-        <xsl:value-of select="$text"/>
-      </xsl:when>
-      <xsl:otherwise>
-        <xsl:variable name="last" select="substring($text, string-length ($text), 1)"/>
-        <xsl:variable name="before" select="substring($text, 1, string-length ($text) - 1)"/>
-        <xsl:choose>
-          <xsl:when test="$last = ' ' or $last = '&#9;'                         or $last = '&#10;' or $last = '&#13;'">
-            <xsl:call-template name="remove-trailing-whitespace">
-              <xsl:with-param name="text" select="$before"/>
-            </xsl:call-template>
-          </xsl:when>
-          <xsl:otherwise>
-            <xsl:value-of select="$text"/>
-          </xsl:otherwise>
-        </xsl:choose>
-      </xsl:otherwise>
-    </xsl:choose>
-  </xsl:template>
-  <xsl:template match="text()[position() = 1 and position() = last()]" mode="copy-source-code">
-    <xsl:call-template name="remove-trailing-whitespace">
-      <xsl:with-param name="text">
-        <xsl:call-template name="remove-leading-empty-lines">
-          <xsl:with-param name="text" select="."/>
-        </xsl:call-template>
-      </xsl:with-param>
-    </xsl:call-template>
-  </xsl:template>
-  <xsl:template match="text()[position() = 1 and position() != last()]" mode="copy-source-code">
-    <xsl:call-template name="remove-leading-empty-lines">
-      <xsl:with-param name="text" select="."/>
-    </xsl:call-template>
-  </xsl:template>
-  <xsl:template match="text()[position() &gt; 1 and position() = last()]" mode="copy-source-code">
-    <xsl:call-template name="remove-trailing-whitespace">
-      <xsl:with-param name="text" select="."/>
-    </xsl:call-template>
+  <xsl:template match="text()" mode="copy-source-code">
+    <xsl:value-of select="."/>
   </xsl:template>
   <xsl:template match="text()"/>
   <xsl:template match="mped:copy" mode="copy-source-code">
-- 
cgit v1.2.3