Merge lp:~henninge/launchpad/devel-update-copyright into lp:launchpad

[Henning Eggers]

New utility scripts: find-changed-files.sh and update-copyright.

Adds a little script that updates the year information in the copyright header of all changed files. This should make it easy to keep them up to date.

I extracted the code that finds changed files using bzr from lint.sh, which is used by 'make lint'.

No test, this is just a little helper script. Use "bzr diff" after a run to see that it worked.

[Curtis Hovey (sinzui)]

Thank you so much for this.

review: Approve (code)

[Robert Collins (lifeless)]

There is a bzr plugin that can do this without it being in the LP
source tree. It would be nicer to use that.

[Brad Crittenden (bac)]

I'm glad to learn of bzr-update-copyright.

Unfortunately when we first put the copyright notice in the files we started with 2009. Using the bzr plugin will change that to use the actual years the file was touched. IANACL so I don't know which is correct but I'd like that to be a decision we make rather than whatever the tool currently writes, so I'd discourage use of the bzr plugin until we can agree.

Note the two tools do not coexist well now as Henning's uses collapsed year ranges, which we prefer, (e.g. "2009-2010") while the plugin produces a comma-separated list. Once the latter is put in the code then this tool is broken.

Perhaps the best effort would be towards improving the plugin.

I'd like for us to get one or the other b/c we spend a lot of time fixing silly copyright notices or leaving them wrong.

[Robert Collins (lifeless)]

I'd like to see the bzr plugin improved if its insufficient. Perhaps a
file in the bzr extensions namespace in the tree, with a config file
to say (use -, and start in 2009).

=== modified file 'buildout-templates/bin/lint.sh.in'
--- buildout-templates/bin/lint.sh.in	2010-07-17 21:42:06 +0000
+++ buildout-templates/bin/lint.sh.in	2010-09-28 07:58:44 +0000
@@ -1,6 +1,6 @@
 #!/bin/bash
 #
-# Copyright 2009 Canonical Ltd.  This software is licensed under the
+# Copyright 2009-2010 Canonical Ltd.  This software is licensed under the
 # GNU Affero General Public License version 3 (see the file LICENSE).
 #
 # Runs pocketlint on files changed from parent branch.
@@ -12,47 +12,9 @@
 [ -z "$utilitiesdir" ] && utilitiesdir=.
 
 
-bzr() {
-    # For pylint to operate properly, PYTHONPATH must point to the ./lib
-    # directory in the launchpad tree. This directory includes a bzrlib. When
-    # this script calls bzr, we want it to use the system bzrlib, not the one
-    # in the launchpad tree.
-    PYTHONPATH='' `which bzr` "$@"
-}
-
-
 if [ -z "$1" ]; then
-    # No command line argument provided, use the default logic.
-    bzr diff > /dev/null
-    diff_status=$?
-    if [ $diff_status -eq 0 ] ; then
-        # No uncommitted changes in the tree.
-        bzr status | grep "^Current thread:" > /dev/null
-        if [ $? -eq 0 ] ; then
-            # This is a loom, lint changes relative to the lower thread.
-            rev_option="-r thread:"
-        else
-            if test "$(bzr pipes | sed -n -e "/^\\*/q;p" | wc -l)" -gt 0; then
-                # This is a pipeline with at least one pipe before the
-                # current, lint changes relative to the previous pipe
-                rev_option="-r ancestor::prev"
-            else
-                # Lint changes relative to the parent.
-                rev=`bzr info | sed \
-                    '/parent branch:/!d; s/ *parent branch: /ancestor:/'`
-                rev_option="-r $rev"
-            fi
-        fi
-    elif [ $diff_status -eq 1 ] ; then
-        # Uncommitted changes in the tree, lint those changes.
-        rev_option=""
-    else
-        # bzr diff failed
-        exit 1
-    fi
-    # Extract filename from status line.  Strip the @ that mark symlinks.
-    files=`bzr st --short $rev_option |
-        sed -e '/^.[MN]/!d; s/.* //' -e 's/@$//'`
+    # No command line argument provided, lint all changed files.
+    files=$($utilitiesdir/find-changed-files.sh)
 else
     # Add newlines so grep filters out pyfiles correctly later.
     files=`echo $* | tr " " "\n"`

=== added file 'utilities/find-changed-files.sh'
--- utilities/find-changed-files.sh	1970-01-01 00:00:00 +0000
+++ utilities/find-changed-files.sh	2010-09-28 07:58:44 +0000
@@ -0,0 +1,45 @@
+#!/bin/bash
+#
+# Copyright 2009-2010 Canonical Ltd.  This software is licensed under the
+# GNU Affero General Public License version 3 (see the file LICENSE).
+#
+# Determine the changed files in Bazaar piplines, looms and plain branches.
+
+bzr() {
+    # PYTHONPATH may point to the ./lib directory in the launchpad tree. This
+    # directory includes a bzrlib. When this script calls bzr, we want it to
+    # use the system bzrlib, not the one in the launchpad tree.
+    PYTHONPATH='' `which bzr` "$@"
+}
+
+bzr diff > /dev/null
+diff_status=$?
+if [ $diff_status -eq 0 ] ; then
+    # No uncommitted changes in the tree.
+    bzr status | grep "^Current thread:" > /dev/null
+    if [ $? -eq 0 ] ; then
+        # This is a loom, lint changes relative to the lower thread.
+        rev_option="-r thread:"
+    elif [ "$(bzr pipes | sed -n -e "/^\\*/q;p" | wc -l)" -gt 0 ]; then
+        # This is a pipeline with at least one pipe before the
+        # current, lint changes relative to the previous pipe
+        rev_option="-r ancestor::prev"
+    else
+        # Lint changes relative to the parent.
+        rev=`bzr info | sed \
+            '/parent branch:/!d; s/ *parent branch: /ancestor:/'`
+        rev_option="-r $rev"
+    fi
+elif [ $diff_status -eq 1 ] ; then
+    # Uncommitted changes in the tree, return those files.
+    rev_option=""
+else
+    # bzr diff failed
+    exit 1
+fi
+# Extract filename from status line.  Strip the @ that mark symlinks.
+files=`bzr st --short $rev_option |
+    sed -e '/^.[MN]/!d; s/.* //' -e 's/@$//'`
+
+echo $files
+

=== added file 'utilities/update-copyright'
--- utilities/update-copyright	1970-01-01 00:00:00 +0000
+++ utilities/update-copyright	2010-09-28 07:58:44 +0000
@@ -0,0 +1,87 @@
+#!/usr/bin/python
+#
+# Copyright 2010 Canonical Ltd.  This software is licensed under the
+# GNU Affero General Public License version 3 (see the file LICENSE).
+
+"""Update the year in copyright notices.
+
+This simple script determines the changed files and updates the copyright
+notice to reflect the current year. Looks for the notice in the first three
+lines of the file and leaves the file unchanged if it finds none.
+"""
+
+from datetime import date
+import os
+import re
+from subprocess import (
+    PIPE,
+    Popen,
+    )
+import sys
+
+
+# This script lives in the 'utilites' directory.
+UTILITIES_DIR = os.path.dirname(__file__)
+CURRENT_YEAR = date.today().year
+copyright_pattern = re.compile(
+    "Copyright (?P<years>(?P<yearfrom>[0-9]{4})(-[0-9]{4})?) Canonical Ltd.")
+
+def years_string(yearfrom):
+    """Build the new years string."""
+    if int(yearfrom) >= CURRENT_YEAR:
+       return yearfrom
+    return "%s-%d" % (yearfrom, CURRENT_YEAR)
+
+def update_copyright(lines):
+    """Update the copyright notice in the given file lines."""
+    for line in range(min(len(lines), 3)):
+        match = copyright_pattern.search(lines[line])
+        if match is not None:
+            old_years = match.group('years')
+            new_years = years_string(match.group('yearfrom'))
+            if old_years != new_years:
+                lines[line] = lines[line].replace(old_years, new_years)
+                return True
+            return False
+    return False
+
+
+def update_files(filenames):
+    """Open the files with the given file names and update them."""
+    for filename in filenames:
+        if not os.path.isfile(filename):
+            print "Skipped: %s does not exist or is not a regular file." %(
+                filename)
+            continue
+        if not os.access(filename, os.W_OK):
+            print "Skipped: %s is not writeable." % filename
+            continue
+        lines = file(filename).readlines()
+        changed = update_copyright(lines)
+        if changed:
+            newfile = open(filename, 'w')
+            newfile.write(''.join(lines))
+            newfile.close()
+            print "Updated: %s" % filename
+        else:
+            print "Unchanged: %s" % filename
+
+def find_changed_files():
+    """Use the find-changed-files.sh script."""
+    find_changed_files_cmd = [
+        os.path.join(UTILITIES_DIR, 'find-changed-files.sh')]
+    filenames = Popen(find_changed_files_cmd, stdout=PIPE).communicate()[0]
+    return filenames.strip()
+
+def find_and_update():
+    """Put it all together."""
+    filenames = find_changed_files()
+    if filenames != '':
+        update_files(filenames.split(' '))
+
+if __name__ == "__main__":
+    if len(sys.argv) < 2:
+        find_and_update()
+    else:
+        update_files(sys.argv[1:])
+