[Python-modules-commits] r13332 - in packages/pdfminer/trunk/debian (4 files)

[jwilk at users.alioth.debian.org]

    Date: Tuesday, June 8, 2010 @ 16:31:27
  Author: jwilk
Revision: 13332

Reuse CMaps provided by the poppler-data package.

Added:
  packages/pdfminer/trunk/debian/patches/poppler-data.diff
Modified:
  packages/pdfminer/trunk/debian/changelog
  packages/pdfminer/trunk/debian/control
  packages/pdfminer/trunk/debian/patches/series

Modified: packages/pdfminer/trunk/debian/changelog
===================================================================
--- packages/pdfminer/trunk/debian/changelog	2010-06-08 15:23:23 UTC (rev 13331)
+++ packages/pdfminer/trunk/debian/changelog	2010-06-08 16:31:27 UTC (rev 13332)
@@ -6,5 +6,7 @@
       [dfsg-testsuite.diff]
   * Disable test suite for psparser.py, as it is currently broken.
     [psparser-testsuite.diff]
+  * Reuse CMaps provided by the poppler-data package.
+    [poppler-data.diff]
 
- -- Jakub Wilk <jwilk at debian.org>  Tue, 08 Jun 2010 16:21:38 +0200
+ -- Jakub Wilk <jwilk at debian.org>  Tue, 08 Jun 2010 18:23:54 +0200

Modified: packages/pdfminer/trunk/debian/control
===================================================================
--- packages/pdfminer/trunk/debian/control	2010-06-08 15:23:23 UTC (rev 13331)
+++ packages/pdfminer/trunk/debian/control	2010-06-08 16:31:27 UTC (rev 13332)
@@ -15,7 +15,7 @@
 Package: python-pdfminer
 Architecture: all
 Depends: ${misc:Depends}, ${python:Depends}
-Suggests: python-pdfminer-cmap
+Suggests: poppler-data
 Description: PDF parser and analyser
  PDFMiner is a tool for extracting information from PDF documents, which
  focuses entirely on getting and analyzing text data. It allows to obtain the
@@ -26,13 +26,3 @@
  .
  This package provides the Python module and the command-line tools: pdf2txt
  and dumppdf.
-
-Package: python-pdfminer-cmap
-Architecture: all
-Depends: ${misc:Depends}, ${python:Depends}, python-pdfminer
-Description: PDF parser and analyser (CMaps)
- PDFMiner is a tool for extracting information from PDF documents, which
- focuses entirely on getting and analyzing text data. 
- .
- This package contains the encoding data needed to read some PDF documents in
- CJK (Chinese, Japanese, Korean) languages.

Added: packages/pdfminer/trunk/debian/patches/poppler-data.diff
===================================================================
--- packages/pdfminer/trunk/debian/patches/poppler-data.diff	                        (rev 0)
+++ packages/pdfminer/trunk/debian/patches/poppler-data.diff	2010-06-08 16:31:27 UTC (rev 13332)
@@ -0,0 +1,61 @@
+Description:
+  Reuse CMaps provided by the poppler-data package. This way we don't need to
+  ship pdfminer.cmap.* modules in the Debian package.
+Author: Jakub Wilk <jwilk at debian.org>
+Bug-Debian: http://bugs.debian.org/584555
+Forwarded: not-needed
+Last-Update: 2010-06-08
+
+--- pdfminer-20100424+dfsg.orig/pdfminer/cmapdb.py
++++ pdfminer-20100424+dfsg/pdfminer/cmapdb.py
+@@ -15,6 +15,7 @@ import sys
+ import re
+ import os
+ import os.path
++import glob
+ from struct import pack, unpack
+ from psparser import PSStackParser
+ from psparser import PSException, PSSyntaxError, PSTypeError, PSEOF
+@@ -225,8 +226,20 @@ class CMapDB(object):
+         try:
+             module = __import__(modname, fromlist=['pdfminer.cmap'])
+         except ImportError:
++            for directory in glob.glob('/usr/share/poppler/cMap/*/'):
++                if not os.path.exists(directory + name):
++                    continue
++                cmap = FileCMap()
++                fp = file(directory + name, 'rb')
++                try:
++                    CMapParser(cmap, fp).run()
++                finally:
++                    fp.close()
++                return cmap
++            raise ValueError(name)
+             raise CMapDB.CMapNotFound(name)
+-        return PyCMap(name, module)
++        else:
++            return PyCMap(name, module)
+ 
+     @classmethod
+     def get_unicode_map(klass, name, vertical=False):
+@@ -236,8 +249,19 @@ class CMapDB(object):
+         try:
+             module = __import__(modname, fromlist=['pdfminer.cmap'])
+         except ImportError:
++            directory = '/usr/share/poppler/cidToUnicode/'
++            if os.path.exists(directory + name):
++                cmap = FileUnicodeMap()
++                fp = file(directory + name, 'rb')
++                try:
++                    CMapParser(cmap, fp).run()
++                finally:
++                    fp.close()
++                return cmap
++            raise ValueError(name)
+             raise CMapDB.CMapNotFound(name)
+-        return PyUnicodeMap(name, module, vertical)
++        else:
++            return PyUnicodeMap(name, module, vertical)
+ 
+ 
+ ##  CMapParser

Modified: packages/pdfminer/trunk/debian/patches/series
===================================================================
--- packages/pdfminer/trunk/debian/patches/series	2010-06-08 15:23:23 UTC (rev 13331)
+++ packages/pdfminer/trunk/debian/patches/series	2010-06-08 16:31:27 UTC (rev 13332)
@@ -1,2 +1,3 @@
 pdf-testsuite.diff
 psparser-testsuite.diff
+poppler-data.diff