[rkward-cvs] [rkward] scripts: Provide line number information for messages extracted from xml files.

[Thomas Friedrichsmeier]

Git commit 4174de709b5d3894fe65c01bfd180bd578eeeb50 by Thomas Friedrichsmeier.
Committed on 08/12/2014 at 11:04.
Pushed by tfry into branch 'master'.

Provide line number information for messages extracted from xml files.
Unfortunately, there seems to be no clean way to do this. This hack works by inserting dummy attributes into the document before parsing.

M  +27   -6    scripts/update_plugin_messages.py

http://commits.kde.org/rkward/4174de709b5d3894fe65c01bfd180bd578eeeb50

diff --git a/scripts/update_plugin_messages.py b/scripts/update_plugin_messages.py
index 90b8f63..cfb7e2b 100755
--- a/scripts/update_plugin_messages.py
+++ b/scripts/update_plugin_messages.py
@@ -30,6 +30,7 @@ import subprocess
 from xml.dom import minidom
 import HTMLParser
 import copy
+import re
 
 # You might want to adjust the following values (can also be overridden from environment variable):
 BUGADDR = "http://p.sf.net/rkward/bugs"     # Technically, this is for bugs _in the translation_
@@ -52,6 +53,8 @@ attributes_to_extract_for_tag={
   'about':  { "attributes" : ['name', 'shortinfo', 'category'], "context": ''},
   'author': { "attributes" : ['name', 'given', 'family'], "context": 'Author name'}
 }
+# HACK for preserving line number information in the DOM tree. This string should be unique enough to not clash with the files' contents!
+LINE_DUMMY_ATTR = '_DUMMY_LINE'
 
 def usage ():
   print ("Usage: " + sys.argv[0] + " [--default_po=PO_ID] [--outdir=DIR] files")
@@ -84,10 +87,20 @@ if (len (toplevel_sources) < 1):
 # For crying out loud! So we are not strictly using XML, because we allow the use of (X)HTML entities, esp. inside <text>-elements,
 # without formally declaring these entities. Python seems to make a point of making it real hard to deal with this. So what we do is
 # escaping all entities before parsing, then passing all through HTMLParser.unescape () before writing the output.
+#
+# The second thing we do is hacking line number information into the parsed XML tree
 def parseFile (filename):
   f = codecs.open (filename, 'r', 'utf-8')
-  content = f.read ().replace ("&", "&")
+  content = f.read ().replace ("&", "&").split ("\n")
   f.close ()
+
+  l = 0
+  enriched = list ()
+  for line in content:
+    l += 1
+    enriched.append (re.sub (r'<(\w+)', r'<\1 ' + LINE_DUMMY_ATTR + '="' + str (l) + '"', line))
+  content = "\n".join (enriched)
+
   try:
     return minidom.parseString (content)
   except:
@@ -107,7 +120,7 @@ def getElementShort (element, dot_attribute=""):
 
 # Try to extract helpful file context information
 def getFileContext (element, attribute=""):
-  ret = "i18n: file: " + infile["infile"] + "\n"
+  ret = "i18n: file: " + infile["infile"] + ":" + str (getLineOf (element, 0)) + "\n"
   ret += "i18n: ectx: "
   if (infile["caption"] != ""):
     ret += "(" + infile["caption"] + ") "
@@ -137,6 +150,14 @@ def quote (text):
   text = HTMLParser.HTMLParser ().unescape (text)	# unescape character entities, Qt does so while parsing the xml
   return "\"" + text.replace ("\\", "\\\\").replace ("\"", "\\\"") + "\""
 
+def stripLineDummy (text):
+  return re.sub (r' ' + LINE_DUMMY_ATTR + '="\d+"', '', text)
+
+def getLineOf (element, default=-1):
+  if element.hasAttribute (LINE_DUMMY_ATTR):
+    return element.getAttribute (LINE_DUMMY_ATTR)
+  return default
+
 # Normalizes larger text fragments. TODO: Do we want to protect <pre>-blocks?
 def normalize (text):
   lines = text.split ("\n")
@@ -150,7 +171,7 @@ def getFullText (element):
   rc = []
   for cn in element.childNodes:
     if cn.nodeType != cn.COMMENT_NODE:
-      rc.append(cn.toxml ("utf-8"))
+      rc.append(stripLineDummy (cn.toxml ("utf-8")))
   return ''.join (rc).strip ().replace ("&", "&")
 
 # get the content of all text nodes inside this node (does not include xml tags)
@@ -158,7 +179,7 @@ def getText (node):
   rc = []
   for cn in node.childNodes:
     if cn.nodeType in [cn.TEXT_NODE, cn.CDATA_SECTION_NODE]:
-      rc.append(cn.data)
+      rc.append(stripLineDummy (cn.data))
   return ''.join (rc).strip ()
 
 # Look for an i18n comment in the given node, and add automatically extracted file context information
@@ -168,7 +189,7 @@ def getI18nComment (node, attribute=""):
     if cn.nodeType == cn.COMMENT_NODE:
       comment = normalize (cn.data.strip ())
       if (comment.lower ().startswith ("i18n:") or comment.lower ().startswith ("translators:")):
-        ret += "i18n: " + comment + "\n"
+        ret += "i18n: " + stripLineDummy (comment) + "\n"
   ret += getFileContext (node, attribute) + " */\n"
   return (ret)
 
@@ -202,7 +223,7 @@ def handleNode (node):
       else:
         handleSubFile (filename, node.tagName == "component", node.tagName == "include")
     if (node.tagName == "script"):
-      handleJSChunk (getText (node), infile["infile"], -1, infile["caption"])
+      handleJSChunk (getText (node), infile["infile"], getLineOf (node), infile["caption"])
     elif (node.tagName in text_containers):
       textchunks = getFullText (node).split ("\n\n")
       for chunk in textchunks: