[yelp-tools] yelp-check: Added tool to check Mallard page licenses

[Shaun McCance <shaunm src gnome org>]

commit 34c6ceca3967fd845b8a42f4fe54309073273d4b
Author: Shaun McCance <shaunm gnome org>
Date:   Wed Jan 29 13:21:52 2014 -0500

    yelp-check: Added tool to check Mallard page licenses

 tools/yelp-check.in  |   82 ++++++++++++++++++++++++++++++++++++++++
 xslt/Makefile.am     |    1 +
 xslt/mal-license.xsl |  102 ++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 185 insertions(+), 0 deletions(-)
---
diff --git a/tools/yelp-check.in b/tools/yelp-check.in
index 1ad0c20..05531a6 100755
--- a/tools/yelp-check.in
+++ b/tools/yelp-check.in
@@ -17,6 +17,7 @@
 # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 
 xsl_mal_link='@XSL_MAL_LINK@'
+xsl_mal_license='@DATADIR@/xslt/mal-license.xsl'
 xsl_mal_status='@DATADIR@/xslt/mal-status.xsl'
 xsl_mal_rng='@DATADIR@/xslt/mal-rng.xsl'
 xsl_comments='@DATADIR@/xslt/comments.xsl'
@@ -108,6 +109,7 @@ yelp_usage () {
         echo "  comments      Print the editorial comments in a document"
         echo "  hrefs         Find broken external links in a document"
         echo "  ids           Find Mallard page IDs that do not match file names"
+       echo "  license       Report the license of Mallard pages"
         echo "  links         Find broken xref or linkend links in a document"
         echo "  orphans       Find orphaned pages in a Mallard document"
         echo "  status        Report the status of Mallard pages"
@@ -162,6 +164,25 @@ yelp_usage_comments () {
         echo "  comment element in Mallard and the remark element in DocBook."
     ) 1>&2
 }
+yelp_usage_license () {
+    (
+        echo "Usage: yelp-check license <FILES>"
+        echo ""
+        echo "  Report the license of the Mallard page files FILES. Each"
+        echo "  matching page is reporting along with its status, reported"
+       echo "  based on the href attribute of the license element. Common"
+       echo "  licenses use a shortened identifier. Pages with multiple"
+       echo "  licenses have the identifiers separated by spaces. Pages"
+       echo "  with no license element report 'none'. Licenses with no"
+       echo "  href attribute are reported as 'unknown'."
+        echo ""
+        echo "Options:"
+        echo "  --only LICENSES     Only show pages whose license is in LICENSES"
+        echo "  --except LICENSES   Exclude pages whose license is in LICENSES"
+        echo "  --totals            Show total counts for each license"
+        echo "LICENSES may be a comma- and/or space-separated list."
+    ) 1>&2
+}
 yelp_usage_status () {
     (
         echo "Usage: yelp-check status <FILES>"
@@ -537,6 +558,64 @@ yelp_comments () {
     done
 }
 
+yelp_license_page () {
+    xsltproc --xinclude \
+        --stringparam only "$check_only" \
+        --stringparam except "$check_except" \
+        --stringparam totals "$check_totals" \
+        "$xsl_mal_license" "$1"
+}
+
+yelp_license () {
+    if [ "$#" = "0" -o "x$1" = "x--help" ]; then
+        yelp_usage_license
+        exit 1
+    fi
+    while [ "$#" != "0" ]; do
+        case "$1" in
+            "--only")
+                shift
+                check_only="$1"
+                shift
+                ;;
+            "--except")
+                shift
+                check_except="$1"
+                shift
+                ;;
+            "--totals")
+                check_totals="1"
+                shift
+                ;;
+            *)
+                break
+                ;;
+        esac
+    done
+    if [ "$#" = "0" -o "x$1" = "x--help" ]; then
+        yelp_usage_license
+        exit 1
+    fi
+    for xml in "$@"; do
+        ext=`echo "$1" | sed -e 's/.*\.//'`
+        if [ -d "$xml" ]; then
+            for page in "$xml"/*.page; do
+                yelp_license_page "$page" || gret="$?"
+            done
+        elif [ "x$ext" = "xpage" -o "x$ext" = "xcache" ]; then
+            yelp_license_page "$xml" || gret="$?"
+        else
+           echo "Unrecognized page " $xml 2>&1
+           exit 1
+       fi
+    done | if [ "x$check_totals" = "x1" ]; then
+       sort | uniq -c | sed -e 's/^ *//' | awk '{print $2 ": " $1}'
+    else
+       sort
+    fi
+    return 0
+}
+
 yelp_status () {
     if [ "$#" = "0" -o "x$1" = "x--help" ]; then
         yelp_usage_status
@@ -730,6 +809,9 @@ case "x$cmd" in
     "xorphans")
         yelp_orphans "$@"
         ;;
+    "xlicense")
+        yelp_license "$@"
+        ;;
     "xstatus")
         yelp_status "$@"
         ;;
diff --git a/xslt/Makefile.am b/xslt/Makefile.am
index 5dd0ef4..d2dc1e3 100644
--- a/xslt/Makefile.am
+++ b/xslt/Makefile.am
@@ -5,6 +5,7 @@ xsl_DATA =              \
        mal-epub.xsl    \
        mal-opf.xsl     \
        mal-files.xsl   \
+       mal-license.xsl \
        mal-ncx.xsl     \
        mal-status.xsl  \
        mal-rng.xsl
diff --git a/xslt/mal-license.xsl b/xslt/mal-license.xsl
new file mode 100644
index 0000000..cf3a98d
--- /dev/null
+++ b/xslt/mal-license.xsl
@@ -0,0 +1,102 @@
+<xsl:stylesheet
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
+    xmlns:mal="http://projectmallard.org/1.0/";
+    xmlns:str="http://exslt.org/strings";
+    xmlns:exsl="http://exslt.org/common";
+    version="1.0">
+
+<xsl:output method="text"/>
+
+<xsl:param name="only" select="''"/>
+<xsl:variable name="only_" select="concat(' ', translate($only, ',', ' '), ' ')"/>
+<xsl:param name="except" select="''"/>
+<xsl:variable name="except_" select="concat(' ', translate($except, ',', ' '), ' ')"/>
+<xsl:param name="totals" select="''"/>
+
+<xsl:template match="/">
+  <xsl:variable name="idents">
+    <xsl:if test="not(/mal:page/mal:info/mal:license)">
+      <ident>none</ident>
+    </xsl:if>
+    <xsl:for-each select="/mal:page/mal:info/mal:license">
+      <ident>
+       <xsl:call-template name="license.identifier"/>
+      </ident>
+    </xsl:for-each>
+  </xsl:variable>
+  <xsl:variable name="display">
+    <xsl:for-each select="exsl:node-set($idents)/*">
+      <xsl:choose>
+       <xsl:when test="$only != ''">
+         <xsl:if test="contains($only_, concat(' ', ., ' '))">
+           <xsl:text>1</xsl:text>
+         </xsl:if>
+       </xsl:when>
+       <xsl:when test="$except != ''">
+         <xsl:if test="not(contains($except_, concat(' ', ., ' ')))">
+           <xsl:text>1</xsl:text>
+         </xsl:if>
+       </xsl:when>
+       <xsl:otherwise>
+         <xsl:text>1</xsl:text>
+       </xsl:otherwise>
+      </xsl:choose>
+    </xsl:for-each>
+  </xsl:variable>
+  <xsl:if test="$display != ''">
+    <xsl:choose>
+      <xsl:when test="$totals = ''">
+       <xsl:value-of select="concat(/mal:page/@id, ': ')"/>
+       <xsl:for-each select="exsl:node-set($idents)/*">
+         <xsl:if test="position() != 1">
+           <xsl:text>, </xsl:text>
+         </xsl:if>
+         <xsl:value-of select="."/>
+       </xsl:for-each>
+       <xsl:text>&#x0A;</xsl:text>
+      </xsl:when>
+      <xsl:otherwise>
+       <xsl:for-each select="exsl:node-set($idents)/*">
+         <xsl:value-of select="."/>
+         <xsl:text>&#x0A;</xsl:text>
+       </xsl:for-each>
+      </xsl:otherwise>
+    </xsl:choose>
+  </xsl:if>
+</xsl:template>
+
+<xsl:template name="license.identifier">
+  <xsl:param name="href" select="@href"/>
+  <xsl:choose>
+    <xsl:when test="$href = ''">
+      <xsl:text>unknown</xsl:text>
+    </xsl:when>
+    <xsl:when test="starts-with($href, 'http://creativecommons.org/licenses/')">
+      <xsl:text>cc-</xsl:text>
+      <xsl:for-each select="str:split(substring-after($href, 'http://creativecommons.org/licenses/'), '/')">
+       <xsl:if test="position() != 1">
+         <xsl:text>-</xsl:text>
+       </xsl:if>
+       <xsl:value-of select="."/>
+      </xsl:for-each>
+    </xsl:when>
+    <xsl:when test="starts-with($href, 'http://www.gnu.org/licenses/')">
+      <xsl:variable name="basename">
+       <xsl:value-of select="substring-after($href, 'http://www.gnu.org/licenses/')"/>
+      </xsl:variable>
+      <xsl:choose>
+       <xsl:when test="substring($basename, string-length($basename) - 4) = '.html'">
+         <xsl:value-of select="substring($basename, 1, string-length($basename) - 5)"/>
+       </xsl:when>
+       <xsl:otherwise>
+         <xsl:value-of select="$basename"/>
+       </xsl:otherwise>
+      </xsl:choose>
+    </xsl:when>
+    <xsl:otherwise>
+      <xsl:text>unknown</xsl:text>
+    </xsl:otherwise>
+  </xsl:choose>
+</xsl:template>
+
+</xsl:stylesheet>