From bcea54fcaa16a7d5feff0c4cd038fea51d1359ea Mon Sep 17 00:00:00 2001
From: Jay Berkenbilt <ejb@ql.org>
Date: Tue, 27 Oct 2020 10:51:41 -0400
Subject: Revert removal of unreadCh change for performance

Turns out unreadCh is much more efficient than seek(-1, SEEK_CUR).
Update comments and code to reflect this.
---
 manual/qpdf-manual.xml | 22 ----------------------
 1 file changed, 22 deletions(-)

(limited to 'manual/qpdf-manual.xml')

diff --git a/manual/qpdf-manual.xml b/manual/qpdf-manual.xml
index 17b778b9..e3afc704 100644
--- a/manual/qpdf-manual.xml
+++ b/manual/qpdf-manual.xml
@@ -4967,28 +4967,6 @@ print "\n";
         </listitem>
        </itemizedlist>
       </listitem>
-      <listitem>
-       <para>
-        Notice of upcoming API change
-       </para>
-       <itemizedlist>
-        <listitem>
-         <para>
-	  The method <function>InputSource::unreadCh(unsigned
-	  char)</function> is deprecated and will be removed in qpdf
-	  11. It has never worked properly to pass a character to
-	  <function>unreadCh</function> other than the most recently
-	  read character. If you happen to be deriving a class from
-	  <type>InputSource</type>, just implement
-	  <function>unreadCh</function> to seek backward by one
-	  character if the current position is greater than 0. If you
-	  are calling this in your own code, replacing with
-	  <literal>seek(-1, SEEK_CUR)</literal> should work in all
-	  cases.
-         </para>
-        </listitem>
-       </itemizedlist>
-      </listitem>
      </itemizedlist>
     </listitem>
    </varlistentry>
-- 
cgit v1.2.3-54-g00ecf