275 lines
7.7 KiB
Python
275 lines
7.7 KiB
Python
#!/usr/bin/python
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import re
|
|
import sys
|
|
from optparse import OptionParser
|
|
import httplib
|
|
import urllib
|
|
import cgi
|
|
try:
|
|
import json
|
|
except ImportError:
|
|
import simplejson as json
|
|
|
|
|
|
namePattern = re.compile(r' \([0-9]+\)')
|
|
|
|
def clean(str):
|
|
return quoteHtml(re.sub(namePattern, "", str))
|
|
|
|
def formatComponents(str):
|
|
str = re.sub(namePattern, '', str).replace("'", "")
|
|
if str != "":
|
|
ret = "(" + str + ")"
|
|
else:
|
|
ret = ""
|
|
return quoteHtml(ret)
|
|
|
|
def quoteHtml(str):
|
|
return cgi.escape(str).encode('ascii', 'xmlcharrefreplace')
|
|
|
|
def mstr(obj):
|
|
if (obj == None):
|
|
return ""
|
|
return unicode(obj)
|
|
|
|
class Version:
|
|
"""Represents a version number"""
|
|
def __init__(self, data):
|
|
self.mod = False
|
|
self.data = data
|
|
found = re.match('^((\d+)(\.\d+)*).*$', data)
|
|
if (found):
|
|
self.parts = [ int(p) for p in found.group(1).split('.') ]
|
|
else:
|
|
self.parts = []
|
|
# backfill version with zeroes if missing parts
|
|
self.parts.extend((0,) * (3 - len(self.parts)))
|
|
|
|
def decBugFix(self):
|
|
self.mod = True
|
|
self.parts[2] -= 1
|
|
return self
|
|
|
|
def __str__(self):
|
|
if (self.mod):
|
|
return '.'.join([ str(p) for p in self.parts ])
|
|
return self.data
|
|
|
|
def __cmp__(self, other):
|
|
return cmp(self.parts, other.parts)
|
|
|
|
class Jira:
|
|
"""A single JIRA"""
|
|
|
|
def __init__(self, data, parent):
|
|
self.key = data['key']
|
|
self.fields = data['fields']
|
|
self.parent = parent
|
|
self.notes = None
|
|
|
|
def getId(self):
|
|
return mstr(self.key)
|
|
|
|
def getDescription(self):
|
|
return mstr(self.fields['description'])
|
|
|
|
def getReleaseNote(self):
|
|
if (self.notes == None):
|
|
field = self.parent.fieldIdMap['Release Note']
|
|
if (self.fields.has_key(field)):
|
|
self.notes=mstr(self.fields[field])
|
|
else:
|
|
self.notes=self.getDescription()
|
|
return self.notes
|
|
|
|
def getPriority(self):
|
|
ret = ""
|
|
pri = self.fields['priority']
|
|
if(pri != None):
|
|
ret = pri['name']
|
|
return mstr(ret)
|
|
|
|
def getAssignee(self):
|
|
ret = ""
|
|
mid = self.fields['assignee']
|
|
if(mid != None):
|
|
ret = mid['displayName']
|
|
return mstr(ret)
|
|
|
|
def getComponents(self):
|
|
return " , ".join([ comp['name'] for comp in self.fields['components'] ])
|
|
|
|
def getSummary(self):
|
|
return self.fields['summary']
|
|
|
|
def getType(self):
|
|
ret = ""
|
|
mid = self.fields['issuetype']
|
|
if(mid != None):
|
|
ret = mid['name']
|
|
return mstr(ret)
|
|
|
|
def getReporter(self):
|
|
ret = ""
|
|
mid = self.fields['reporter']
|
|
if(mid != None):
|
|
ret = mid['displayName']
|
|
return mstr(ret)
|
|
|
|
def getProject(self):
|
|
ret = ""
|
|
mid = self.fields['project']
|
|
if(mid != None):
|
|
ret = mid['key']
|
|
return mstr(ret)
|
|
|
|
|
|
|
|
class JiraIter:
|
|
"""An Iterator of JIRAs"""
|
|
|
|
def __init__(self, versions):
|
|
self.versions = versions
|
|
|
|
resp = urllib.urlopen("https://issues.apache.org/jira/rest/api/2/field")
|
|
data = json.loads(resp.read())
|
|
|
|
self.fieldIdMap = {}
|
|
for part in data:
|
|
self.fieldIdMap[part['name']] = part['id']
|
|
|
|
self.jiras = []
|
|
at=0
|
|
end=1
|
|
count=100
|
|
while (at < end):
|
|
params = urllib.urlencode({'jql': "project in (HADOOP,HDFS,MAPREDUCE,YARN) and fixVersion in ('"+"' , '".join(versions)+"') and resolution = Fixed", 'startAt':at, 'maxResults':count})
|
|
resp = urllib.urlopen("https://issues.apache.org/jira/rest/api/2/search?%s"%params)
|
|
data = json.loads(resp.read())
|
|
if (data.has_key('errorMessages')):
|
|
raise Exception(data['errorMessages'])
|
|
at = data['startAt'] + data['maxResults']
|
|
end = data['total']
|
|
self.jiras.extend(data['issues'])
|
|
|
|
self.iter = self.jiras.__iter__()
|
|
|
|
def __iter__(self):
|
|
return self
|
|
|
|
def next(self):
|
|
data = self.iter.next()
|
|
j = Jira(data, self)
|
|
return j
|
|
|
|
class Outputs:
|
|
"""Several different files to output to at the same time"""
|
|
|
|
def __init__(self, base_file_name, file_name_pattern, keys, params={}):
|
|
self.params = params
|
|
self.base = open(base_file_name%params, 'w')
|
|
self.others = {}
|
|
for key in keys:
|
|
both = dict(params)
|
|
both['key'] = key
|
|
self.others[key] = open(file_name_pattern%both, 'w')
|
|
|
|
def writeAll(self, pattern):
|
|
both = dict(self.params)
|
|
both['key'] = ''
|
|
self.base.write(pattern%both)
|
|
for key in self.others.keys():
|
|
both = dict(self.params)
|
|
both['key'] = key
|
|
self.others[key].write(pattern%both)
|
|
|
|
def writeKeyRaw(self, key, str):
|
|
self.base.write(str)
|
|
if (self.others.has_key(key)):
|
|
self.others[key].write(str)
|
|
|
|
def close(self):
|
|
self.base.close()
|
|
for fd in self.others.values():
|
|
fd.close()
|
|
|
|
def main():
|
|
parser = OptionParser(usage="usage: %prog [options] [USER-ignored] [PASSWORD-ignored] [VERSION]")
|
|
parser.add_option("-v", "--version", dest="versions",
|
|
action="append", type="string",
|
|
help="versions in JIRA to include in releasenotes", metavar="VERSION")
|
|
parser.add_option("--previousVer", dest="previousVer",
|
|
action="store", type="string",
|
|
help="previous version to include in releasenotes", metavar="VERSION")
|
|
|
|
(options, args) = parser.parse_args()
|
|
|
|
if (options.versions == None):
|
|
options.versions = []
|
|
|
|
if (len(args) > 2):
|
|
options.versions.append(args[2])
|
|
|
|
if (len(options.versions) <= 0):
|
|
parser.error("At least one version needs to be supplied")
|
|
|
|
versions = [ Version(v) for v in options.versions];
|
|
versions.sort();
|
|
|
|
maxVersion = str(versions[-1])
|
|
if(options.previousVer == None):
|
|
options.previousVer = str(versions[0].decBugFix())
|
|
print >> sys.stderr, "WARNING: no previousVersion given, guessing it is "+options.previousVer
|
|
|
|
list = JiraIter(options.versions)
|
|
version = maxVersion
|
|
outputs = Outputs("releasenotes.%(ver)s.html",
|
|
"releasenotes.%(key)s.%(ver)s.html",
|
|
["HADOOP","HDFS","MAPREDUCE","YARN"], {"ver":maxVersion, "previousVer":options.previousVer})
|
|
|
|
head = '<META http-equiv="Content-Type" content="text/html; charset=UTF-8">\n' \
|
|
'<title>Hadoop %(key)s %(ver)s Release Notes</title>\n' \
|
|
'<STYLE type="text/css">\n' \
|
|
' H1 {font-family: sans-serif}\n' \
|
|
' H2 {font-family: sans-serif; margin-left: 7mm}\n' \
|
|
' TABLE {margin-left: 7mm}\n' \
|
|
'</STYLE>\n' \
|
|
'</head>\n' \
|
|
'<body>\n' \
|
|
'<h1>Hadoop %(key)s %(ver)s Release Notes</h1>\n' \
|
|
'These release notes include new developer and user-facing incompatibilities, features, and major improvements. \n' \
|
|
'<a name="changes"/>\n' \
|
|
'<h2>Changes since Hadoop %(previousVer)s</h2>\n' \
|
|
'<ul>\n'
|
|
|
|
outputs.writeAll(head)
|
|
|
|
for jira in list:
|
|
line = '<li> <a href="https://issues.apache.org/jira/browse/%s">%s</a>.\n' \
|
|
' %s %s reported by %s and fixed by %s %s<br>\n' \
|
|
' <b>%s</b><br>\n' \
|
|
' <blockquote>%s</blockquote></li>\n' \
|
|
% (quoteHtml(jira.getId()), quoteHtml(jira.getId()), clean(jira.getPriority()), clean(jira.getType()).lower(),
|
|
quoteHtml(jira.getReporter()), quoteHtml(jira.getAssignee()), formatComponents(jira.getComponents()),
|
|
quoteHtml(jira.getSummary()), quoteHtml(jira.getReleaseNote()))
|
|
outputs.writeKeyRaw(jira.getProject(), line)
|
|
|
|
outputs.writeAll("</ul>\n</body></html>\n")
|
|
outputs.close()
|
|
|
|
if __name__ == "__main__":
|
|
main()
|
|
|