From 29416875e00eb898065b81af3055b1d96c540fc2 Mon Sep 17 00:00:00 2001
From: Christian Berendt <berendt@b1-systems.de>
Date: Thu, 15 Aug 2013 11:38:31 +0200
Subject: [PATCH] added scripts to cleanup spaces in XML files

example for remove_trailing_whitespaces.sh:

"  testing   \n" --> "  testing\n"

examples for remove_unnecessary_spaces.py:

"<testing>    foobar foobar" --> "<testing>foobar foobar"
"foobar foobar   </testing>" --> "foobar foobar</testing>"

Change-Id: Ie15a7a1491bf412ff893c84e76061ef623581145
---
 tools/cleanup/remove_trailing_whitespaces.sh | 13 +++++
 tools/cleanup/remove_unnecessary_spaces.py   | 56 ++++++++++++++++++++
 2 files changed, 69 insertions(+)
 create mode 100755 tools/cleanup/remove_trailing_whitespaces.sh
 create mode 100755 tools/cleanup/remove_unnecessary_spaces.py

diff --git a/tools/cleanup/remove_trailing_whitespaces.sh b/tools/cleanup/remove_trailing_whitespaces.sh
new file mode 100755
index 0000000000..3d41284b16
--- /dev/null
+++ b/tools/cleanup/remove_trailing_whitespaces.sh
@@ -0,0 +1,13 @@
+#!/bin/sh
+
+## copyright: B1 Systems GmbH   <info@b1-systems.de>,    2013.
+##    author: Christian Berendt <berendt@b1-systems.de>, 2013.
+##   license: Apache License, Version 2.0
+
+# Call ./tools/cleanup/remove_trailing_whitespaces.sh in the
+# root of openstack-manuals.
+
+files=$(find doc/src/docbkx -name *.xml -not -name pom.xml)
+for file in $files; do
+    sed -i -e 's/[[:space:]]*$//' $file
+done
diff --git a/tools/cleanup/remove_unnecessary_spaces.py b/tools/cleanup/remove_unnecessary_spaces.py
new file mode 100755
index 0000000000..620dab08b1
--- /dev/null
+++ b/tools/cleanup/remove_unnecessary_spaces.py
@@ -0,0 +1,56 @@
+#!/usr/bin/python
+
+## copyright: B1 Systems GmbH   <info@b1-systems.de>,    2013.
+##    author: Christian Berendt <berendt@b1-systems.de>, 2013.
+##   license: Apache License, Version 2.0
+
+# Call ./tools/cleanup/remove_unnecessary_spaces.py in the
+# root of openstack-manuals.
+
+import os
+import re
+import tempfile
+
+# should be the same like in tools/validate.py
+FILE_EXCEPTIONS = ['ha-guide-docinfo.xml','bk001-ch003-associate-general.xml']
+
+elements = [
+    'listitem',
+    'para',
+    'td',
+    'th',
+    'command',
+    'literal',
+    'title',
+    'caption',
+    'filename',
+    'userinput',
+    'programlisting'
+]
+
+checks = []
+for element in elements:
+    checks.append(re.compile("(.*<%s>)\s+([\w\-().:!?{}\[\]]+.*\n)" % element)),
+    checks.append(re.compile("(.*[\w\-().:!?{}\[\]]+)\s+(<\/%s>.*\n)" % element))
+
+for root, dirs, files in os.walk('doc/src/docbkx/'):
+    for f in files:
+        if (not (f.endswith('.xml') and
+                 f != 'pom.xml' and
+                 f not in FILE_EXCEPTIONS)):
+            continue
+        docfile = os.path.abspath(os.path.join(root, f))
+        tmpfile = tempfile.mkstemp()
+        tmpfd  = os.fdopen(tmpfile[0], "w")
+        match = False
+        for line in open(docfile, 'r'):
+            for check in checks:
+                if check.match(line):
+                    line = check.sub(r"\1\2", line)
+                    match = True
+            tmpfd.write(line)
+        tmpfd.close()
+        if match:
+            os.rename(tmpfile[1], docfile)
+        else:
+            os.unlink(tmpfile[1])