From: Anthony Towns <aj@azure.humbug.org.au>
Date: Tue, 28 Aug 2007 06:47:47 +0000 (+1000)
Subject: 	* examine_package.py: Summarise duplicate copyright file entries
X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=b18bae735f509f17d35a3f80284cd9bd374bd3a6;p=dak

	* examine_package.py: Summarise duplicate copyright file entries
	(same md5sum) with a reference to the previous instance, rather
	than repeating them.
---

diff --git a/ChangeLog b/ChangeLog
index 64b9754e..469a418c 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -4,6 +4,9 @@
 	processing.
 	* config/debian/dak.conf, scripts/debian/byhand-tag: Automatic
 	processing of tag-overrides.
+	* examine_package.py: Summarise duplicate copyright file entries
+	(same md5sum) with a reference to the previous instance, rather
+	than repeating them.
 
 2007-06-19  Anthony Towns  <ajt@debian.org>
 
diff --git a/dak/examine_package.py b/dak/examine_package.py
index e87cc281..ba2855d4 100755
--- a/dak/examine_package.py
+++ b/dak/examine_package.py
@@ -32,7 +32,7 @@
 
 ################################################################################
 
-import errno, os, pg, re, sys
+import errno, os, pg, re, sys, md5
 import apt_pkg, apt_inst
 import daklib.database, daklib.utils
 
@@ -82,6 +82,8 @@ Cnf = daklib.utils.get_conf()
 projectB = pg.connect(Cnf["DB::Name"], Cnf["DB::Host"], int(Cnf["DB::Port"]))
 daklib.database.init(Cnf, projectB)
 
+printed_copyrights = {}
+
 ################################################################################
 
 def usage (exit_code=0):
@@ -302,7 +304,15 @@ def print_copyright (deb_filename):
         return
 
     o = os.popen("ar p %s data.tar.gz | tar xzOf - %s" % (deb_filename, copyright))
-    print o.read()
+    copyright = o.read()
+    copyrightmd5 = md5.md5(copyright).hexdigest()
+
+    if printed_copyrights.has_key(copyrightmd5):
+        print "Copyright is the same as %s.\n" % \
+		(printed_copyrights[copyrightmd5])
+    else:
+    	print copyright
+	printed_copyrights[copyrightmd5] = "%s (%s)" % (package, deb_filename)
 
 def check_dsc (dsc_filename):
     print "---- .dsc file for %s ----" % (dsc_filename)