Improve keep-entities stuff even further, add another test case.

8 files changed, 135 insertions, 19 deletions

diff --git a/xml2po/ChangeLog b/xml2po/ChangeLog
index efcf568..fc378fb 100644
--- a/xml2po/ChangeLog
+++ b/xml2po/ChangeLog
@@ -1,3 +1,7 @@
+2006-04-01  Danilo Šegan  <danilo@gnome.org>
+
+	* xml2po.py (replaceNodeContentsWithText): Improve fix for #336765.
+
 2006-03-31  Danilo Šegan  <danilo@gnome.org>
 
 	Fixes bug #317502.
diff --git a/xml2po/tests/ChangeLog b/xml2po/tests/ChangeLog
index ccb2011..25977aa 100644
--- a/xml2po/tests/ChangeLog
+++ b/xml2po/tests/ChangeLog
@@ -1,3 +1,14 @@
+2006-04-01  Danilo Šegan  <danilo@gnome.org>
+
+	* adjacent-ents.xml, adjacent-ents.xml.out, adjacent-ents.pot, 
+	adjacent-ents.po: 
+	* test.py (SIMPLETESTS): 
+	Added adjacent-ents.xml test-case.
+
+2006-04-01  Danilo Šegan  <danilo@gnome.org>
+
+	* test.py (SIMPLETESTS): Updated framework to support options.
+
 2006-03-31  Danilo Šegan  <danilo@gnome.org>
 
 	* keepents.xml, keepents.xml.out, keepents.pot, keepents.po, 
diff --git a/xml2po/tests/adjacent-ents.po b/xml2po/tests/adjacent-ents.po
new file mode 100644
index 0000000..037eccf
--- /dev/null
+++ b/xml2po/tests/adjacent-ents.po
@@ -0,0 +1,28 @@
+msgid ""
+msgstr ""
+"Project-Id-Version: adjacent entities and elements test\n"
+"POT-Creation-Date: 2006-04-01 00:02+0200\n"
+"PO-Revision-Date: 2006-04-01 01:10+0200\n"
+"Last-Translator: Данило\n"
+"Language-Team: српски\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=UTF-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+
+#: adjacent-ents.xml:8(para) 
+msgid "&suselinux; &suse-version; Release Notes"
+msgstr "&suselinux; &suse-version; Објава издања"
+
+#: adjacent-ents.xml:9(para) 
+msgid "<application>SUSE Linux</application> &suse-version; is ..."
+msgstr "<application>SUSE Линукс</application> &suse-version; је ..."
+
+#: adjacent-ents.xml:10(para) 
+msgid "&suselinux; <application>&suse-version;</application> is ..."
+msgstr "&suselinux; <application>&suse-version;</application> је ..."
+
+#. Put one translator per line, in the form of NAME <EMAIL>, YEAR1, YEAR2.
+#: adjacent-ents.xml:0(None) 
+msgid "translator-credits"
+msgstr ""
+
diff --git a/xml2po/tests/adjacent-ents.pot b/xml2po/tests/adjacent-ents.pot
new file mode 100644
index 0000000..71c12ab
--- /dev/null
+++ b/xml2po/tests/adjacent-ents.pot
@@ -0,0 +1,28 @@
+msgid ""
+msgstr ""
+"Project-Id-Version: PACKAGE VERSION\n"
+"POT-Creation-Date: \n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <LL@li.org>\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=UTF-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+
+#: adjacent-ents.xml:8(title) 
+msgid "&suselinux; &suse-version; Release Notes"
+msgstr ""
+
+#: adjacent-ents.xml:9(para) 
+msgid "<application>SUSE Linux</application> &suse-version; is ..."
+msgstr ""
+
+#: adjacent-ents.xml:10(para) 
+msgid "&suselinux; <application>&suse-version;</application> is ..."
+msgstr ""
+
+#. Put one translator per line, in the form of NAME <EMAIL>, YEAR1, YEAR2.
+#: adjacent-ents.xml:0(None) 
+msgid "translator-credits"
+msgstr ""
+
diff --git a/xml2po/tests/adjacent-ents.xml b/xml2po/tests/adjacent-ents.xml
new file mode 100644
index 0000000..42f292a
--- /dev/null
+++ b/xml2po/tests/adjacent-ents.xml
@@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="utf-8"?>
+<!DOCTYPE sect1 PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN" "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd" [
+<!ENTITY suse "SUSE">
+<!ENTITY suselinux "&suse; Linux">
+<!ENTITY suse-version "10.1">
+]>
+<sect1>
+ <title><!-- blah --> &suselinux; &suse-version; Release Notes</title>
+ <para><application>SUSE Linux</application> &suse-version; is ...</para>
+ <para>&suselinux; <application> &suse-version;</application> is ...</para>
+</sect1>
diff --git a/xml2po/tests/adjacent-ents.xml.out b/xml2po/tests/adjacent-ents.xml.out
new file mode 100644
index 0000000..fb01f26
--- /dev/null
+++ b/xml2po/tests/adjacent-ents.xml.out
@@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="utf-8"?>
+<!DOCTYPE sect1 PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN" "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd" [
+<!ENTITY suse "SUSE">
+<!ENTITY suselinux "&suse; Linux">
+<!ENTITY suse-version "10.1">
+]>
+<sect1>
+ <title>&suselinux; &suse-version; Објава издања</title>
+ <para><application>SUSE Линукс</application> &suse-version; је ...</para>
+ <para>&suselinux; <application>&suse-version;</application> је ...</para>
+</sect1>
diff --git a/xml2po/tests/test.py b/xml2po/tests/test.py
index 512eed9..c482b6e 100755
--- a/xml2po/tests/test.py
+++ b/xml2po/tests/test.py
@@ -1,7 +1,15 @@
 #!/usr/bin/env python
 import sys, os
 
-SIMPLETESTS = ['deep-finals.xml', 'deep-nonfinals.xml', 'attribute-entities.xml', 'docbook.xml', 'utf8-original.xml', 'footnotes.xml', 'keepents.xml' ]
+SIMPLETESTS = { 'deep-finals.xml' : {},
+                'deep-nonfinals.xml': {},
+                'attribute-entities.xml': {},
+                'docbook.xml' : {},
+                'utf8-original.xml': {},
+                'footnotes.xml': {},
+                'keepents.xml': { "options" : "-k" },
+                'adjacent-ents.xml': { "options" : "-k" },
+                }
 
 OTHERTESTS = [ ('relnotes', 'test.sh') ]
 
@@ -9,16 +17,23 @@ if len(sys.argv) > 1:
     input = sys.argv[1]
     pot = input.replace(".xml", ".pot")
     po = input.replace(".xml", ".po")
+    myopts = ""
+    if len(sys.argv) > 2:
+        for opt in sys.argv[2:]:
+            myopts += " " + opt
     output = input.replace(".xml", ".xml.out")
-    ret = os.system("../xml2po %s | sed 's/\"POT-Creation-Date: .*$/\"POT-Creation-Date: \\\\n\"/' | diff -u %s -" % (input, pot))
+    ret = os.system("../xml2po %s %s | sed 's/\"POT-Creation-Date: .*$/\"POT-Creation-Date: \\\\n\"/' | diff -u %s -" % (myopts, input, pot))
     if ret:
         print "Problem: extraction from '%s'" % (input)
-    ret = os.system("../xml2po -p %s %s | diff -u %s -" % (po, input, output))
+    ret = os.system("../xml2po -p %s %s %s | diff -u %s -" % (po, myopts, input, output))
     if ret:
         print "Problem: merging translation into '%s'" % (input)
 else:
     for t in SIMPLETESTS:
-        if os.system("%s %s" % (sys.argv[0], t)):
+        if SIMPLETESTS[t].has_key("options"):
+            myopts = SIMPLETESTS[t]["options"]
+        else: myopts = ""
+        if os.system("%s %s %s" % (sys.argv[0], t, myopts)):
             print "WARNING: Test %s failed." % (t)
     
     for t in OTHERTESTS:
diff --git a/xml2po/xml2po.py b/xml2po/xml2po.py
index 61835cc..ff9f684 100755
--- a/xml2po/xml2po.py
+++ b/xml2po/xml2po.py
@@ -116,13 +116,17 @@ msgstr ""
 
 
 def normalizeNode(node):
+    #print >>sys.stderr, "<%s> (%s) [%s]" % (node.name, node.type, node.serialize('utf-8'))
     if not node:
         return
     elif isSpacePreserveNode(node):
         return
     elif node.isText():
-        if node.isBlankNode() and (not expand_entities and (not node.next or node.next.type!='entity_ref')):
-            node.setContent('')
+        if node.isBlankNode():
+            if expand_entities or ( not (node.prev and not node.prev.isBlankNode()
+                                         and node.next and not node.next.isBlankNode()) ):
+                #print >>sys.stderr, "BLANK"
+                node.setContent('')
         else:
             node.setContent(re.sub('\s+',' ', node.content))
 
@@ -213,6 +217,7 @@ def getTranslation(text, spacepreserve = 0):
     text should be a string to look for, spacepreserve set to 1
     when spaces should be preserved.
     """
+    #print >>sys.stderr,"getTranslation('%s')" % (text.encode('utf-8'))
     text = normalizeString(text, not spacepreserve)
     if (text.strip() == ''):
         return text
@@ -302,28 +307,26 @@ def getCommentForNode(node):
     else:
         return None
 
-
 def replaceNodeContentsWithText(node,text):
     """Replaces all subnodes of a node with contents of text treated as XML."""
 
     if node.children:
         starttag = node.name #startTagForNode(node)
         endtag = endTagForNode(node)
+
+        # Lets add document DTD so entities are resolved
+        tmp = '<?xml version="1.0" encoding="utf-8" ?>'
         try:
-            # Lets add document DTD so entities are resolved
             dtd = doc.intSubset()
-            tmp = ''
-            # NOTE: We used to get a "Segmentation fault" in libxml2.parseMemory() when we included DTD
-            #       when not expand_entities!
-            #       libxml 2.6.21 is known to work, need to see what's the minimal version.
-            #if expand_entities:
-            tmp = dtd.serialize('utf-8')
-            tmp = tmp + '<%s>%s</%s>' % (starttag, text, endtag)
-        except:
-            tmp = '<%s>%s</%s>' % (starttag, text, endtag)
+            tmp = tmp + dtd.serialize('utf-8')
+        except libxml2.treeError:
+            pass
+            
+        content = '<%s>%s</%s>' % (starttag, text, endtag)
+        tmp = tmp + content.encode('utf-8')
 
         try:
-            ctxt = libxml2.createDocParserCtxt(tmp.encode('utf-8'))
+            ctxt = libxml2.createDocParserCtxt(tmp)
             ctxt.replaceEntities(0)
             ctxt.parseDocument()
             newnode = ctxt.doc()
@@ -339,7 +342,12 @@ def replaceNodeContentsWithText(node,text):
                 free.unlinkNode()
                 free = next
 
-            node.addChildList(newelem.children)
+            add = newelem.children
+            if node:
+                copy = newelem.copyNodeList()
+                next = node.next
+                node.replaceNode(newelem.copyNodeList())
+                node.next = next
         else:
             # In practice, this happens with tags such as "<para>    </para>" (only whitespace in between)
             pass