Merge lp:~openerp-dev/openobject-server/6.0-opw-17600-odo into lp:openobject-server/6.0

[Olivier Dony (Odoo)]

(Correct target branch this time, sorry for previous one)

This patch makes the report preprocessing system properly handle separately inner and trailing text for each XML node. This fixes potential issues where the tailing text would be put inside the node if the node is empty, e.g:
   <p></p>bar would become <p>bar</p>

But it also avoids ignoring trailing text when there is an inner text, e.g.:
   <p>[[ o.name ]]</p>[[ removeParentNode('td') ]] would have the removeParentNode() part ignored.

To be further tested with real reports, this is untested so far.

Unmerged revisions

3510. By Olivier Dony (Odoo) on 2011-10-03

[FIX] report.preprocess: properly handle node.text and node.tail [CASE 17600]

=== modified file 'bin/report/preprocess.py'
--- bin/report/preprocess.py	2010-05-18 10:32:42 +0000
+++ bin/report/preprocess.py	2011-10-03 12:07:14 +0000
@@ -71,11 +71,17 @@
                         n = n.getparent()
                     n.set('rml_loop', txt.group(2))
                     return '[['+txt.group(1)+"''"+txt.group(4)+']]'
-                t = _regex1.sub(_sub1, node.text or node.tail)
-                if t == " ":
-                    t = _regex11.sub(_sub1, node.text  or node.tail)
-                t = _regex3.sub(_sub3, t)
-                node.text = _regex2.sub(_sub2, t)
+                def preprocess_string(text):
+                    t = _regex1.sub(_sub1, text)
+                    if t == " ":
+                        t = _regex11.sub(_sub1, text)
+                    t = _regex3.sub(_sub3, t)
+                    t = _regex2.sub(_sub2, t)
+                    return t
+                if node.text:
+                    node.text = preprocess_string(node.text)
+                if node.tail:
+                    node.tail = preprocess_string(node.tail)
             self.preprocess_rml(node,type)
         return root_node