Add option to control line length when preprocessing PDF input

src/calibre/ebooks/conversion/plumber.py

@@ -694,7 +694,7 @@ def create_oebbook(log, path_or_stream, opts, input_plugin, reader=None,
     '''
     from calibre.ebooks.oeb.base import OEBBook
     html_preprocessor = HTMLPreProcessor(input_plugin.preprocess_html,
-            opts.preprocess_html)
+            opts.preprocess_html, getattr(opts, 'pdf_line_length', 0.5))
     oeb = OEBBook(log, html_preprocessor,
             pretty_print=opts.pretty_print, input_encoding=encoding)
     if not populate:

src/calibre/ebooks/conversion/preprocess.py

@@ -159,9 +159,11 @@ class HTMLPreProcessor(object):
                      (re.compile('<span[^><]*?id=subtitle[^><]*?>(.*?)</span>', re.IGNORECASE|re.DOTALL),
                       lambda match : '<h3 class="subtitle">%s</h3>'%(match.group(1),)),
                      ]
-    def __init__(self, input_plugin_preprocess, plugin_preprocess):
+    def __init__(self, input_plugin_preprocess, plugin_preprocess,
+            pdf_line_length):
         self.input_plugin_preprocess = input_plugin_preprocess
         self.plugin_preprocess = plugin_preprocess
+        self.pdf_line_length = pdf_line_length
 
     def is_baen(self, src):
         return re.compile(r'<meta\s+name="Publisher"\s+content=".*?Baen.*?"',
@@ -182,7 +184,7 @@ class HTMLPreProcessor(object):
         elif self.is_book_designer(html):
             rules = self.BOOK_DESIGNER
         elif self.is_pdftohtml(html):
-            length = line_length(html, .3)
+            length = line_length(html, self.pdf_line_length)
             line_length_rules = []
             if length:
                 line_length_rules = [

src/calibre/ebooks/html/input.py

@@ -261,6 +261,11 @@ class HTMLInput(InputFormatPlugin):
                 'nasty side effects in the rest of of the conversion pipeline.'
                 )
         ),
+
+        OptionRecommendation(name='pdf_line_length', recommended_value=0.5,
+            help=_('Average line length for line breaking if the HTML is from a '
+                'previous partial conversion of a PDF file.')),
+
     ])
 
     def convert(self, stream, opts, file_ext, log,

src/calibre/ebooks/pdf/input.py

@@ -20,6 +20,8 @@ class PDFInput(InputFormatPlugin):
     options = set([
         OptionRecommendation(name='no_images', recommended_value=False,
             help=_('Do not extract images from the document')),
+        OptionRecommendation(name='pdf_line_length', recommended_value=0.5,
+            help=_('Average line length for line breaking')),
     ])
 
     def convert(self, stream, options, file_ext, log,

src/calibre/trac/plugins/templates/linux.html

@@ -140,6 +140,11 @@ sudo calibre_postinstall
         </form>
       </div>
       <hr/>
+      <h3>Note</h3>
+      <p>
+        If your kernel is compiled with CONFIG_SYSFS_DEPRECATED device detection may not work.
+      </p>
+      <hr/>
       <h3>Dependencies</h3>
       ${app} has the following dependencies (the listed version is the minimum version)
       <br/><br/>