Generate ePub files of documentation.

[exim.git] / doc / doc-docbook / Pre-xml

diff --git a/doc/doc-docbook/Pre-xml b/doc/doc-docbook/Pre-xml
index 4e606dd27ec1c35ee2b9c844eab8024f0c3390a4..22b343d8db8b234b95ed28bf35b952f478ff6916 100755 (executable)
--- a/doc/doc-docbook/Pre-xml
+++ b/doc/doc-docbook/Pre-xml
@@ -1,7 +1,5 @@
 #! /usr/bin/perl
 
 #! /usr/bin/perl
 

-# $Cambridge: exim/doc/doc-docbook/Pre-xml,v 1.3 2006/02/01 11:01:01 ph10 Exp $
-

 # Script to pre-process XML input before processing it for various purposes.
 # Options specify which transformations are to be done. Monospaced literal
 # layout blocks are never touched.
 # Script to pre-process XML input before processing it for various purposes.
 # Options specify which transformations are to be done. Monospaced literal
 # layout blocks are never touched.


@@ -30,7 +28,7 @@
 #            newline, the space and newline are removed, because otherwise you
 #            get a blank line in the HTML output.
 #
 #            newline, the space and newline are removed, because otherwise you
 #            get a blank line in the HTML output.
 #

-# -noindex   Remove the XML to generate a Concept and an Options index.
+# -noindex   Remove the XML that generates indexes.

 # -oneindex  Ditto, but add XML to generate a single index.
 #
 # -optbreak  Insert an optional line break (zero-width space, ​) after
 # -oneindex  Ditto, but add XML to generate a single index.
 #
 # -optbreak  Insert an optional line break (zero-width space, ​) after


@@ -38,7 +36,9 @@
 #            except when preceded by <entry> (i.e. not in tables). The same is
 #            also done within a word of four or more upper-case letters (for
 #            compile-time options).
 #            except when preceded by <entry> (i.e. not in tables). The same is
 #            also done within a word of four or more upper-case letters (for
 #            compile-time options).

-
+#
+# -epub      Convert date formats to comply with epub specification
+#

 
 
 # The function that processes non-literal, non-monospaced text
 
 
 # The function that processes non-literal, non-monospaced text


@@ -73,11 +73,18 @@ if ($ascii)
 $s;
 }
 
 $s;
 }
 

+# Mapping needed for epub
+
+my %months = (
+  jan => '01', feb => '02', mar => '03', apr => '04', may => '05', jun => '06',
+  jul => '07', aug => '08', sep => '09', oct => '10', nov => '11', dec => '12',
+);

 
 # The main program
 
 $ascii     = 0;
 $bookinfo  = 0;
 
 # The main program
 
 $ascii     = 0;
 $bookinfo  = 0;

+$epub      = 0;

 $html      = 0;
 $inliteral = 0;
 $inliterallayout = 0;
 $html      = 0;
 $inliteral = 0;
 $inliterallayout = 0;


@@ -93,6 +100,7 @@ foreach $arg (@ARGV)
   if    ($arg eq "-fi")       { $ligatures = 1; }
   elsif ($arg eq "-ascii")    { $ascii = 1; }
   elsif ($arg eq "-bookinfo") { $bookinfo = 1; }
   if    ($arg eq "-fi")       { $ligatures = 1; }
   elsif ($arg eq "-ascii")    { $ascii = 1; }
   elsif ($arg eq "-bookinfo") { $bookinfo = 1; }

+  elsif ($arg eq "-epub")     { $epub = 1; }

   elsif ($arg eq "-html")     { $html = 1; }
   elsif ($arg eq "-noindex")  { $noindex = 1; }
   elsif ($arg eq "-oneindex") { $oneindex = 1; }
   elsif ($arg eq "-html")     { $html = 1; }
   elsif ($arg eq "-noindex")  { $noindex = 1; }
   elsif ($arg eq "-oneindex") { $oneindex = 1; }


@@ -143,6 +151,25 @@ while (<STDIN>)
     next;
     }
 
     next;
     }
 

+  # Adjust dates to YYYY-MM-DD
+
+  if ($epub && /^\s*<date[\s>]/)
+    {
+    print;
+    while (<STDIN>)
+      {
+      last if /^\s*<\/date>/;
+
+      if (/^ \s* (\d{1,2}) \s+ ([a-zA-Z]{3}) \s+ (\d{4})/x)
+        {
+        die "Unknown month '$2'\n" unless exists $months{lc $2};
+        my $month = $months{lc $2};
+        s//$3-$month-$1/;
+        }
+      print;
+      }
+    }
+

   # A line that is not in a monospaced literal block; keep track of which
   # parts are in <literal> and which not. The latter get processed by the
   # function above. Items in <literal> get quoted unless they are also in
   # A line that is not in a monospaced literal block; keep track of which
   # parts are in <literal> and which not. The latter get processed by the
   # function above. Items in <literal> get quoted unless they are also in