Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

DOCX Input: Simplify generated markup

Browse Source 
DOCX Input: Simplify generated markup by avoiding the use of unnecessary
<span> tags for whitespace preservation.
This commit is contained in:
Kovid Goyal 2013-06-12 09:45:59 +05:30
parent 02bf6b9aec
commit a447b0818d
1 changed files with 9 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
src/calibre/ebooks/docx/to_html.py
View File

@ -41,6 +41,8 @@ class Convert(object):
def __init__(self, path_or_stream, dest_dir=None, log=None, notes_text=None): def __init__(self, path_or_stream, dest_dir=None, log=None, notes_text=None):
self.docx = DOCX(path_or_stream, log=log) self.docx = DOCX(path_or_stream, log=log)
self.ms_pat = re.compile(r'\s{2,}')
self.ws_pat = re.compile(r'[\n\r\t]')
self.log = self.docx.log self.log = self.docx.log
self.notes_text = notes_text or _('Notes') self.notes_text = notes_text or _('Notes')
self.dest_dir = dest_dir or os.getcwdu() self.dest_dir = dest_dir or os.getcwdu()
@ -414,7 +416,14 @@ class Convert(object):
if not child.text: if not child.text:
continue continue
space = child.get(XML('space'), None) space = child.get(XML('space'), None)
preserve = False
if space == 'preserve': if space == 'preserve':
# Only use a <span> with white-space:pre-wrap if this element
# actually needs it, i.e. if it has more than one
# consecutive space or it has newlines or tabs.
multi_spaces = self.ms_pat.search(child.text) is not None
preserve = multi_spaces or self.ws_pat.search(child.text) is not None
if preserve:
text.add_elem(SPAN(child.text, style="white-space:pre-wrap")) text.add_elem(SPAN(child.text, style="white-space:pre-wrap"))
ans.append(text.elem) ans.append(text.elem)
else: else: