Only remove list markers up to two characters

List markers that are longer than two characters are most likely not list markers, so leave them alone.
sbsdev · Mar 29, 2022 · d6b84c7 · d6b84c7
1 parent 6eb103e
commit d6b84c7
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 10 deletions.
diff --git a/modules/sbs/epub3-dtbook-migrator/src/main/resources/xml/xslt/dtbook-to-epub3.xsl b/modules/sbs/epub3-dtbook-migrator/src/main/resources/xml/xslt/dtbook-to-epub3.xsl
@@ -1486,22 +1486,22 @@
                         <xsl:apply-templates select="."/>
                     </xsl:when>
                     <xsl:when
-                        test="self::dtbook:* and not(preceding-sibling::node()[self::* or self::text()[normalize-space()]]) and not(text()[1]/preceding-sibling::*) and matches(text()[1],'^(\w+\.|•)')">
-                        <xsl:if test="* or normalize-space(replace(text()[1],'^(\w+\.|•) ','')) != ''">
+                        test="self::dtbook:* and not(preceding-sibling::node()[self::* or self::text()[normalize-space()]]) and not(text()[1]/preceding-sibling::*) and matches(text()[1],'^([0-9a-zA-Z]{1,2}\.|•)')">
+                        <xsl:if test="* or normalize-space(replace(text()[1],'^([0-9a-zA-Z]{1,2}\.|•) ','')) != ''">
                             <xsl:variable name="element">
                                 <xsl:apply-templates select="."/>
                             </xsl:variable>
                             <xsl:for-each select="$element/*">
                                 <xsl:copy exclude-result-prefixes="#all">
                                     <xsl:copy-of select="@*" exclude-result-prefixes="#all"/>
-                                    <xsl:value-of select="replace(text()[1],'^(\w+\.|•) ','')"/>
+                                    <xsl:value-of select="replace(text()[1],'^([0-9a-zA-Z]\.|•) ','')"/>
                                     <xsl:copy-of select="node() except text()[1]" exclude-result-prefixes="#all"/>
                                 </xsl:copy>
                             </xsl:for-each>
                         </xsl:if>
                     </xsl:when>
-                    <xsl:when test="self::text() and not(preceding-sibling::node()[self::* or self::text()[normalize-space()]]) and matches(.,'^(\w+\.|•)')">
-                        <xsl:value-of select="replace(.,'^(\w+\.|•) ','')"/>
+                    <xsl:when test="self::text() and not(preceding-sibling::node()[self::* or self::text()[normalize-space()]]) and matches(.,'^([0-9a-zA-Z]{1,2}\.|•)')">
+                        <xsl:value-of select="replace(.,'^([0-9a-zA-Z]{1,2}\.|•) ','')"/>
                     </xsl:when>
                     <xsl:otherwise>
                         <xsl:apply-templates select="."/>

diff --git a/modules/sbs/epub3-dtbook-migrator/src/test/xspec/dtbook-to-epub3.xspec b/modules/sbs/epub3-dtbook-migrator/src/test/xspec/dtbook-to-epub3.xspec
@@ -2363,7 +2363,7 @@
             <x:expect label="the resulting list should be wrapped in a section element containing the hd">
                 <html:section id="...">
 		    <html:h1 id="...">TEXT</html:h1>
-		    <html:ol id="..." class="list-style-type-none">...</html:ol>
+		    <html:ol id="..." class="list-style-type-none preformatted">...</html:ol>
 	        </html:section>
             </x:expect>
         </x:scenario>
@@ -2607,7 +2607,7 @@
                 </dtbook:list>
             </x:context>
             <x:expect label="the type of list should be detected, the list markers removed and the value should be correct">
-                <html:ol id="...">
+                <html:ol id="..." class="preformatted">
                     <html:li id="..." value="2">Halten Sie
 		  Sie es tunlichst, der Öffentlichkeit jene Hautfladen auch bekannt als Bingofladen, die unter den Oberarmen hängen, zuzumuten. Dasselbe gilt für die Röllchen unter der Achsel, zwischen Oberarm und Oberkörper.</html:li>
                 </html:ol>
@@ -3514,17 +3514,17 @@
         </x:expect>
     </x:scenario>
 
-    <x:scenario label="when processing a li element enumerated with letters of with a string length &gt; 1">
+    <x:scenario label="when processing a li element enumerated with letters with a string length &gt; 2">
         <x:context>
             <dtbook:list type="pl" depth="1">
                 <dtbook:li>aa. ONE</dtbook:li>
                 <dtbook:li>bbb. TWO</dtbook:li>
             </dtbook:list>
         </x:context>
-        <x:expect label="the letters should be correctly parsed">
+        <x:expect label="the list markers should not be removed">
             <html:ol id="..." type="a" class="preformatted">
                 <html:li id="..." value="27">ONE</html:li>
-                <html:li id="..." value="1406">TWO</html:li>
+                <html:li id="..." value="1406">bbb. TWO</html:li>
             </html:ol>
         </x:expect>
     </x:scenario>