fdo#42924 don't ignore empty TableCell Element, also check if it's a header

Checking whether an empty TableCell Element is a header row is now in 2 cases:

* | e | h | h | h | h | ... => The empty cell should be a header
This case occurs when this row is the first row in the table.

* | h | p | p | e | p | ... => The empty cell should not be a header
This case occurs when this row is not the first row in the table,
but the first column is the row header.

where | is a column seperator, e is an empty cell, h is a header cell
(which has role="tablehead" attribute), and p is not a header cell.

Note that parsing occurs left-to-right, so isTableHeader depends on
the last TableCell Element in that row. I assume that if the last element
is a header, that row should be a header row.

Currently this code gives correct behaviour, but checking whether a row
is the first row might be more correct.

1 files changed, 18 insertions, 0 deletions

diff --git a/helpcontent2/to-wiki/wikiconv2.py b/helpcontent2/to-wiki/wikiconv2.py
index acc1b457b7..7772030aca 100755
--- a/helpcontent2/to-wiki/wikiconv2.py
+++ b/helpcontent2/to-wiki/wikiconv2.py
@@ -475,8 +475,10 @@ class Text:
 class TableCell(ElementBase):
     def __init__(self, attrs, parent):
         ElementBase.__init__(self, 'tablecell', parent)
+        self.cellHasChildElement = False
 
     def start_element(self, parser, name, attrs):
+        self.cellHasChildElement = True
         if name == 'bookmark':
             self.parse_child(Bookmark(attrs, self, 'div', parser))
         elif name == 'comment':
@@ -492,6 +494,18 @@ class TableCell(ElementBase):
         else:
             self.unhandled_element(parser, name)
 
+    def get_all(self):
+        text = ''
+        if not self.cellHasChildElement: # an empty element
+            if self.parent.isTableHeader: # get from TableRow Element
+                role = 'tablehead'
+            else:
+                role = 'tablecontent'
+            text = text + replace_paragraph_role['start'][role]
+            text = text + replace_paragraph_role['end'][role]
+        text = text + ElementBase.get_all(self)
+        return text
+
 class TableRow(ElementBase):
     def __init__(self, attrs, parent):
         ElementBase.__init__(self, 'tablerow', parent)
@@ -1089,6 +1103,10 @@ class TableContentParagraph(Paragraph):
                 self.role = 'tablecontentcode'
             else:
                 self.role = 'tablecontent'
+        if self.role == 'tablehead':
+            self.parent.parent.isTableHeader = True # self.parent.parent is TableRow Element
+        else:
+            self.parent.parent.isTableHeader = False
 
 class ParserBase:
     def __init__(self, filename, follow_embed, embedding_app, current_app, wiki_page_name, lang, head_object, buffer):