r3941 - jhalfs/trunk/LFS

pierre at higgs.linuxfromscratch.org pierre at higgs.linuxfromscratch.org
Mon Mar 27 09:14:49 PDT 2017


Author: pierre
Date: Mon Mar 27 09:14:49 2017
New Revision: 3941

Log:
Fix the case of the "symlinks.xml" file, which has revision attributes in sect2, which defeats a filter
based solely on screen and sect1 attributes.

Modified:
   jhalfs/trunk/LFS/lfs.xsl

Modified: jhalfs/trunk/LFS/lfs.xsl
==============================================================================
--- jhalfs/trunk/LFS/lfs.xsl	Mon Mar 27 02:40:41 2017	(r3940)
+++ jhalfs/trunk/LFS/lfs.xsl	Mon Mar 27 09:14:49 2017	(r3941)
@@ -53,15 +53,19 @@
   </xsl:template>
   
   <xsl:template match="sect1">
+<!-- Since this xsl:if tag encloses the whole template, it would
+     be much better to transpose this condition to the select part
+     of the "calling" apply-template. But that would change the numbering,
+     so that it would be difficult to compare to previous versions. So for
+     version 2.4, let us keep this -->
         <xsl:if test="(../@id='chapter-temporary-tools' or
                   ../@id='chapter-building-system' or
                   ../@id='chapter-bootscripts' or
                   ../@id='chapter-bootable') and
-                  count(descendant::screen/userinput) > 0 and
-                  count(descendant::screen/userinput) >
-                      count(descendant::screen[@role='nodump']) and
-                  count(descendant::screen/userinput) >
-                      count(descendant::screen/userinput[starts-with(string(),'chroot')])">
+                  (sect2[not(@revision) or @revision=$revision]//..|.)/
+                      screen[(not(@role) or @role != 'nodump') and
+                             (not(@revision) or @revision=$revision)]/
+                          userinput[not(starts-with(string(),'chroot'))]">
 <!-- The last condition is a hack to allow previous versions of the
      book where the chroot commands did not have role="nodump".
      It only works if the chroot command is the only one on the page -->


More information about the alfs-log mailing list