###############################################################################
#
# Diff two licenses.xml files. Splits the result in three sets:
# 1) Removed
#   Packages present in the old file, but not in the new.
#
# 2) Added
#   Packages present in the new file, but not in the old.
#
# 3) Changed
#   Packages present in both files, but with different versions.
#
# Note that packages with the unchanged version is not listed.
#
# The output is presented as XML code printed to stdout.
#
###############################################################################

import sys
import subprocess
import xml.etree.ElementTree as ET

def get_pkgs(file_spec):
    if file_spec.find(":") >= 0:
        cmd = ("git", "show", file_spec)
        root = ET.fromstring(subprocess.check_output(cmd))

    else:
        root = ET.parse(file_spec)

    for node in root.iter("section"):
        if "id" in node.attrib:
            if node.attrib["id"] == "licenses_packages":
                break

    else:
        return None

    for node in node:
        if node.tag == "informaltable":
            break
    else:
        return None

    tab = node[0][-1]

    plist = dict()
    for row in tab:
        pname = row[0].text
        pver  = row[1].text
        if not pname in plist:
            plist[pname] = set()

        plist[pname].add(pver)

    return set(plist), plist

#----------------------------------------

if len(sys.argv) != 3:
    sys.stderr.write("Usage:\n")
    sys.stderr.write("  %s " % sys.argv[0])
    sys.stderr.write("<new license file> <old license file>\n")
    sys.stderr.write("\n")
    sys.stderr.write("  E.g.:\n")
    sys.stderr.write("    %s " % sys.argv[0])
    sys.stderr.write("licenses-5.0-ppc.xml licenses.xml\n")
    exit()

new_file, old_file = sys.argv[1:3]

sys.stderr.write("New license file  : %s\n" % new_file)
sys.stderr.write("Prev license file : %s\n" % old_file)

old_pset, old_pdict = get_pkgs(old_file)
new_pset, new_pdict = get_pkgs(new_file)

added   = new_pset - old_pset # Set subtraction
removed = old_pset - new_pset # Set subtraction
common  = old_pset & new_pset
changed = [ p for p in common if old_pdict[p] != new_pdict[p] ]

print '<?xml version="1.0" encoding="ISO-8859-1"?>'
print '<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"'
print '"http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">'
print '<section id="relinfo-package-set-changes">'
print '  <title>Changes in the Set of Provided Package</title>'
print '  <para>'
print '    This section describes changes in the provided packages.'
print '  </para>'
print '  <section id="relinfo-added-packages">'
print '    <title>Added Packages</title>'
print '    <informaltable>'
print '      <tgroup cols="2">'
print '        <thead>'
print '          <row>'
print '            <entry>Package</entry>'
print '            <entry>Version(s)</entry>'
print '          </row>'
print '        </thead>'
print '        <tbody>'

for p in sorted(list(added)):
    print '          <row>'
    print '            <entry>%s</entry>' % p
    print '            <entry>%s</entry>' % ", ".join(sorted(new_pdict[p]))
    print '          </row>'

print '        </tbody>'
print '      </tgroup>'
print '    </informaltable>'
print '  </section>'

print '  <section id="relinfo-removed-packages">'
print '    <title>Removed Packages</title>'
print '    <informaltable>'
print '      <tgroup cols="2">'
print '        <thead>'
print '          <row>'
print '            <entry>Package</entry>'
print '            <entry>Version(s)</entry>'
print '          </row>'
print '        </thead>'
print '        <tbody>'
for p in sorted(list(removed)):
    print '          <row>'
    print '            <entry>%s</entry>' % p
    print '            <entry>%s</entry>' % ", ".join(sorted(old_pdict[p]))
    print '          </row>'
print '        </tbody>'
print '      </tgroup>'
print '    </informaltable>'
print '  </section>'

print '  <section id="relinfo-changed-packages">'
print '    <title>Changed Package Versions</title>'
print '    <informaltable>'
print '      <tgroup cols="3">'
print '        <thead>'
print '          <row>'
print '            <entry>Package</entry>'
print '            <entry>Previous Version(s)</entry>'
print '            <entry>New Version(s)</entry>'
print '          </row>'
print '        </thead>'
print '        <tbody>'
for p in sorted(list(changed)):
    print '          <row>'
    print '            <entry>%s</entry>' % p
    print '            <entry>%s</entry>' % ", ".join(sorted(old_pdict[p]))
    print '            <entry>%s</entry>' % ", ".join(sorted(new_pdict[p]))
    print '          </row>'

print '        </tbody>'
print '      </tgroup>'
print '    </informaltable>'
print '  </section>'
print '</section>'

sys.stderr.write("Package Summary:\n")
sys.stderr.write("  Prev file : %3d\n" % len(old_pset))
sys.stderr.write("  New file  : %3d\n" % len(new_pset))
sys.stderr.write("  Added     : %3d\n" % len(added))
sys.stderr.write("  Removed   : %3d\n" % len(removed))
sys.stderr.write("  Changed   : %3d\n" % len(changed))
sys.stderr.write("  Unchanged : %3d\n" % (len(common) - len(changed)))
sys.stderr.write("Done\n")