def test_private_comments_do_not_show(self):
self.settings.param.allow_private_bugs = False
database.make_test_instance(self.settings.bugzilla)
#MARK SOME COMMENTS PRIVATE
with DB(self.settings.bugzilla) as db:
private_comments = db.query("""
SELECT
bug_id,
comment_id
FROM
longdescs
ORDER BY
mod(comment_id, 7),
comment_id
LIMIT
5
""")
for c in private_comments:
database.mark_comment_private(db, c.comment_id, 1)
es = elasticsearch.make_test_instance("candidate", self.settings.real.bugs)
es_c = elasticsearch.make_test_instance("candidate_comments", self.settings.real.comments)
bz_etl.main(self.settings, es, es_c)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
verify_no_private_comments(es, private_comments)
def test_whiteboard_screened(self):
GOOD_BUG_TO_TEST = 1046
database.make_test_instance(self.settings.bugzilla)
with DB(self.settings.bugzilla) as db:
es = elasticsearch.make_test_instance("candidate",
self.settings.candidate)
#MARK BUG AS ONE OF THE SCREENED GROUPS
database.add_bug_group(db, GOOD_BUG_TO_TEST,
SCREENED_WHITEBOARD_BUG_GROUPS[0])
db.flush()
#SETUP RUN PARAMETERS
param = Struct()
param.end_time = CNV.datetime2milli(get_current_time(db))
param.start_time = 0
param.start_time_str = extract_bugzilla.milli2string(db, 0)
param.alias_file = self.settings.param.alias_file
param.bug_list = struct.wrap([
GOOD_BUG_TO_TEST
]) # bug 1046 sees lots of whiteboard, and other field, changes
param.allow_private_bugs = True
with ThreadedQueue(es, size=1000) as output:
etl(db, output, param, please_stop=None)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
versions = get_all_bug_versions(es, GOOD_BUG_TO_TEST)
for v in versions:
if v.status_whiteboard not in (None, "", "[screened]"):
Log.error("Expecting whiteboard to be screened")
def test_incremental_etl_catches_tracking_flags(self):
database.make_test_instance(self.settings.bugzilla)
with DB(self.settings.bugzilla) as db:
es = elasticsearch.make_test_instance("candidate", self.settings.candidate)
#SETUP RUN PARAMETERS
param = Struct()
param.end_time = CNV.datetime2milli(get_current_time(db))
# FLAGS ADDED TO BUG 813650 ON 18/12/2012 2:38:08 AM (PDT), SO START AT SOME LATER TIME
param.start_time = CNV.datetime2milli(CNV.string2datetime("02/01/2013 10:09:15", "%d/%m/%Y %H:%M:%S"))
param.start_time_str = extract_bugzilla.milli2string(db, param.start_time)
param.alias_file = self.settings.param.alias_file
param.bug_list = struct.wrap([813650])
param.allow_private_bugs = self.settings.param.allow_private_bugs
with ThreadedQueue(es, size=1000) as output:
etl(db, output, param, please_stop=None)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
versions = get_all_bug_versions(es, 813650)
flags = ["cf_status_firefox18", "cf_status_firefox19", "cf_status_firefox_esr17", "cf_status_b2g18"]
for v in versions:
if v.modified_ts>param.start_time:
for f in flags:
if v[f] != "fixed":
Log.error("813650 should have {{flag}}=='fixed'", {"flag": f})
def test_ambiguous_whiteboard_screened(self):
GOOD_BUG_TO_TEST=1046
database.make_test_instance(self.settings.bugzilla)
with DB(self.settings.bugzilla) as db:
es = elasticsearch.make_test_instance("candidate", self.settings.candidate)
#MARK BUG AS ONE OF THE SCREENED GROUPS
database.add_bug_group(db, GOOD_BUG_TO_TEST, SCREENED_WHITEBOARD_BUG_GROUPS[0])
#MARK BUG AS ONE OF THE *NOT* SCREENED GROUPS
database.add_bug_group(db, GOOD_BUG_TO_TEST, "not screened")
db.flush()
#SETUP RUN PARAMETERS
param = Struct()
param.end_time = CNV.datetime2milli(get_current_time(db))
param.start_time = 0
param.start_time_str = extract_bugzilla.milli2string(db, 0)
param.alias_file = self.settings.param.alias_file
param.bug_list = struct.wrap([GOOD_BUG_TO_TEST]) # bug 1046 sees lots of whiteboard, and other field, changes
param.allow_private_bugs = True
with ThreadedQueue(es, size=1000) as output:
etl(db, output, param, please_stop=None)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
versions = get_all_bug_versions(es, GOOD_BUG_TO_TEST)
for v in versions:
if v.status_whiteboard not in (None, "", "[screened]"):
Log.error("Expecting whiteboard to be screened")
def test_specific_bugs(self):
"""
USE A MYSQL DATABASE TO FILL AN ES INSTANCE (USE Fake_ES() INSTANCES TO KEEP
THIS TEST LOCAL) WITH VERSIONS OF BUGS FROM settings.param.bugs. COMPARE
THOSE VERSIONS TO A REFERENCE ES (ALSO CHECKED INTO REPOSITORY)
"""
# settings.param.allow_private_bugs = True
database.make_test_instance(self.settings.bugzilla)
with DB(self.settings.bugzilla) as db:
candidate = elasticsearch.make_test_instance(
"candidate", self.settings.candidate)
reference = elasticsearch.open_test_instance(
"reference", self.settings.private_bugs_reference)
#SETUP RUN PARAMETERS
param = Struct()
param.end_time = CNV.datetime2milli(get_current_time(db))
param.start_time = 0
param.start_time_str = extract_bugzilla.milli2string(db, 0)
param.alias_file = self.settings.param.alias_file
param.bug_list = self.settings.param.bugs
param.allow_private_bugs = self.settings.param.allow_private_bugs
with ThreadedQueue(candidate, size=1000) as output:
etl(db, output, param, please_stop=None)
#COMPARE ALL BUGS
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
compare_both(candidate, reference, self.settings,
self.settings.param.bugs)
def test_specific_bugs(self):
"""
USE A MYSQL DATABASE TO FILL AN ES INSTANCE (USE Fake_ES() INSTANCES TO KEEP
THIS TEST LOCAL) WITH VERSIONS OF BUGS FROM settings.param.bugs. COMPARE
THOSE VERSIONS TO A REFERENCE ES (ALSO CHECKED INTO REPOSITORY)
"""
# settings.param.allow_private_bugs = True
database.make_test_instance(self.settings.bugzilla)
with DB(self.settings.bugzilla) as db:
candidate = elasticsearch.make_test_instance("candidate", self.settings.candidate)
reference = elasticsearch.open_test_instance("reference", self.settings.private_bugs_reference)
#SETUP RUN PARAMETERS
param = Struct()
param.end_time = CNV.datetime2milli(get_current_time(db))
param.start_time = 0
param.start_time_str = extract_bugzilla.milli2string(db, 0)
param.alias_file = self.settings.param.alias_file
param.bug_list = self.settings.param.bugs
param.allow_private_bugs = self.settings.param.allow_private_bugs
with ThreadedQueue(candidate, size=1000) as output:
etl(db, output, param, please_stop=None)
#COMPARE ALL BUGS
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
compare_both(candidate, reference, self.settings, self.settings.param.bugs)
def test_private_comments_do_not_show(self):
self.settings.param.allow_private_bugs = False
database.make_test_instance(self.settings.bugzilla)
#MARK SOME COMMENTS PRIVATE
with DB(self.settings.bugzilla) as db:
private_comments = db.query("""
SELECT
bug_id,
comment_id
FROM
longdescs
ORDER BY
mod(comment_id, 7),
comment_id
LIMIT
5
""")
for c in private_comments:
database.mark_comment_private(db, c.comment_id, 1)
es = elasticsearch.make_test_instance("candidate",
self.settings.real.bugs)
es_c = elasticsearch.make_test_instance("candidate_comments",
self.settings.real.comments)
bz_etl.main(self.settings, es, es_c)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
verify_no_private_comments(es, private_comments)
def random_sample_of_bugs(self):
"""
I USE THIS TO FIND BUGS THAT CAUSE MY CODE PROBLEMS. OF COURSE, IT ONLY WORKS
WHEN I HAVE A REFERENCE TO COMPARE TO
"""
NUM_TO_TEST = 100
MAX_BUG_ID = 900000
with DB(self.settings.bugzilla) as db:
candidate = elasticsearch.make_test_instance(
"candidate", self.settings.candidate)
reference = ElasticSearch(self.settings.private_bugs_reference)
#GO FASTER BY STORING LOCAL FILE
local_cache = File(self.settings.param.temp_dir +
"/private_bugs.json")
if local_cache.exists:
private_bugs = set(CNV.JSON2object(local_cache.read()))
else:
with Timer("get private bugs"):
private_bugs = compare_es.get_private_bugs(reference)
local_cache.write(CNV.object2JSON(private_bugs))
while True:
some_bugs = [
b for b in
[Random.int(MAX_BUG_ID) for i in range(NUM_TO_TEST)]
if b not in private_bugs
]
Log.note("Test with the following bug_ids: {{bugs}}",
{"bugs": some_bugs})
#SETUP RUN PARAMETERS
param = Struct()
param.end_time = CNV.datetime2milli(get_current_time(db))
param.start_time = 0
param.start_time_str = extract_bugzilla.milli2string(db, 0)
param.alias_file = self.settings.param.alias_file
try:
with ThreadedQueue(candidate, 100) as output:
etl(db, output, param, please_stop=None)
#COMPARE ALL BUGS
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
found_errors = compare_both(candidate, reference,
self.settings, some_bugs)
if found_errors:
Log.note("Errors found")
break
else:
pass
except Exception, e:
Log.warning(
"Total failure during compare of bugs {{bugs}}",
{"bugs": some_bugs}, e)
def run_both_etl(db, output_queue, es_comments, param):
comment_thread = Thread.run("etl comments", etl_comments, db, es_comments, param)
process_thread = Thread.run("etl", etl, db, output_queue, param)
result = comment_thread.join()
if result.exception:
Log.error("etl_comments had problems", result.exception)
result = process_thread.join()
if result.exception:
Log.error("etl had problems", result.exception)
def run_both_etl(db, output_queue, es_comments, param):
comment_thread = Thread.run("etl comments", etl_comments, db, es_comments,
param)
process_thread = Thread.run("etl", etl, db, output_queue, param)
result = comment_thread.join()
if result.exception:
Log.error("etl_comments had problems", result.exception)
result = process_thread.join()
if result.exception:
Log.error("etl had problems", result.exception)
def random_sample_of_bugs(self):
"""
I USE THIS TO FIND BUGS THAT CAUSE MY CODE PROBLEMS. OF COURSE, IT ONLY WORKS
WHEN I HAVE A REFERENCE TO COMPARE TO
"""
NUM_TO_TEST = 100
MAX_BUG_ID = 900000
with DB(self.settings.bugzilla) as db:
candidate = elasticsearch.make_test_instance("candidate", self.settings.candidate)
reference = ElasticSearch(self.settings.private_bugs_reference)
#GO FASTER BY STORING LOCAL FILE
local_cache = File(self.settings.param.temp_dir + "/private_bugs.json")
if local_cache.exists:
private_bugs = set(CNV.JSON2object(local_cache.read()))
else:
with Timer("get private bugs"):
private_bugs = compare_es.get_private_bugs(reference)
local_cache.write(CNV.object2JSON(private_bugs))
while True:
some_bugs = [b for b in [Random.int(MAX_BUG_ID) for i in range(NUM_TO_TEST)] if b not in private_bugs]
Log.note("Test with the following bug_ids: {{bugs}}", {"bugs":some_bugs})
#SETUP RUN PARAMETERS
param = Struct()
param.end_time = CNV.datetime2milli(get_current_time(db))
param.start_time = 0
param.start_time_str = extract_bugzilla.milli2string(db, 0)
param.alias_file = self.settings.param.alias_file
try:
with ThreadedQueue(candidate, 100) as output:
etl(db, output, param, please_stop=None)
#COMPARE ALL BUGS
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
found_errors = compare_both(candidate, reference, self.settings, some_bugs)
if found_errors:
Log.note("Errors found")
break
else:
pass
except Exception, e:
Log.warning("Total failure during compare of bugs {{bugs}}", {"bugs": some_bugs}, e)
def test_private_bugs_do_not_show(self):
self.settings.param.allow_private_bugs = False
File(self.settings.param.first_run_time).delete()
File(self.settings.param.last_run_time).delete()
private_bugs = set(Random.sample(self.settings.param.bugs, 3))
Log.note("The private bugs for this test are {{bugs}}", {"bugs": private_bugs})
database.make_test_instance(self.settings.bugzilla)
#MARK SOME BUGS PRIVATE
with DB(self.settings.bugzilla) as db:
for b in private_bugs:
database.add_bug_group(db, b, BUG_GROUP_FOR_TESTING)
es = elasticsearch.make_test_instance("candidate", self.settings.real.bugs)
es_c = elasticsearch.make_test_instance("candidate_comments", self.settings.real.comments)
bz_etl.main(self.settings, es, es_c)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
verify_no_private_bugs(es, private_bugs)
def test_incremental_etl_catches_tracking_flags(self):
database.make_test_instance(self.settings.bugzilla)
with DB(self.settings.bugzilla) as db:
es = elasticsearch.make_test_instance("candidate",
self.settings.candidate)
#SETUP RUN PARAMETERS
param = Struct()
param.end_time = CNV.datetime2milli(get_current_time(db))
# FLAGS ADDED TO BUG 813650 ON 18/12/2012 2:38:08 AM (PDT), SO START AT SOME LATER TIME
param.start_time = CNV.datetime2milli(
CNV.string2datetime("02/01/2013 10:09:15",
"%d/%m/%Y %H:%M:%S"))
param.start_time_str = extract_bugzilla.milli2string(
db, param.start_time)
param.alias_file = self.settings.param.alias_file
param.bug_list = struct.wrap([813650])
param.allow_private_bugs = self.settings.param.allow_private_bugs
with ThreadedQueue(es, size=1000) as output:
etl(db, output, param, please_stop=None)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
versions = get_all_bug_versions(es, 813650)
flags = [
"cf_status_firefox18", "cf_status_firefox19",
"cf_status_firefox_esr17", "cf_status_b2g18"
]
for v in versions:
if v.modified_ts > param.start_time:
for f in flags:
if v[f] != "fixed":
Log.error("813650 should have {{flag}}=='fixed'",
{"flag": f})
def test_private_bugs_do_not_show(self):
self.settings.param.allow_private_bugs = False
File(self.settings.param.first_run_time).delete()
File(self.settings.param.last_run_time).delete()
private_bugs = set(Random.sample(self.settings.param.bugs, 3))
Log.note("The private bugs for this test are {{bugs}}",
{"bugs": private_bugs})
database.make_test_instance(self.settings.bugzilla)
#MARK SOME BUGS PRIVATE
with DB(self.settings.bugzilla) as db:
for b in private_bugs:
database.add_bug_group(db, b, BUG_GROUP_FOR_TESTING)
es = elasticsearch.make_test_instance("candidate",
self.settings.real.bugs)
es_c = elasticsearch.make_test_instance("candidate_comments",
self.settings.real.comments)
bz_etl.main(self.settings, es, es_c)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
verify_no_private_bugs(es, private_bugs)
b.bug_id
FROM
bugs b
LEFT JOIN
bug_group_map m ON m.bug_id=b.bug_id
WHERE
delta_ts >= {{start_time_str}} AND
m.bug_id IS NULL
""", {
"start_time_str": param.start_time_str
}), u"bug_id")
if not bug_list:
return
with Thread.run("alias analysis", alias_analysis.main, settings=settings, bug_list=bug_list):
Log.note("Updating {{num}} bugs:\n{{bug_list|indent}}", {
"num": len(bug_list),
"bug_list": bug_list
})
param.bug_list = bug_list
run_both_etl(**{
"db": db,
"output_queue": output_queue,
"es_comments": es_comments,
"param": param.copy()
})
def full_etl(resume_from_last_run, settings, param, db, es, es_comments, output_queue):
with Thread.run("alias_analysis", alias_analysis.main, settings=settings):
SELECT
b.bug_id
FROM
bugs b
LEFT JOIN
bug_group_map m ON m.bug_id=b.bug_id
WHERE
delta_ts >= {{start_time_str}} AND
m.bug_id IS NULL
""", {"start_time_str": param.start_time_str}), u"bug_id")
if not bug_list:
return
with Thread.run("alias analysis",
alias_analysis.main,
settings=settings,
bug_list=bug_list):
Log.note("Updating {{num}} bugs:\n{{bug_list|indent}}", {
"num": len(bug_list),
"bug_list": bug_list
})
param.bug_list = bug_list
run_both_etl(
**{
"db": db,
"output_queue": output_queue,
"es_comments": es_comments,
"param": param.copy()
})
def test_changes_to_private_bugs_still_have_bug_group(self):
self.settings.param.allow_private_bugs = True
File(self.settings.param.first_run_time).delete()
File(self.settings.param.last_run_time).delete()
private_bugs = set(Random.sample(self.settings.param.bugs, 3))
Log.note("The private bugs for this test are {{bugs}}",
{"bugs": private_bugs})
database.make_test_instance(self.settings.bugzilla)
#MARK SOME BUGS PRIVATE
with DB(self.settings.bugzilla) as db:
for b in private_bugs:
database.add_bug_group(db, b, BUG_GROUP_FOR_TESTING)
es = elasticsearch.make_test_instance("candidate",
self.settings.real.bugs)
es_c = elasticsearch.make_test_instance("candidate_comments",
self.settings.real.comments)
bz_etl.main(self.settings, es, es_c)
# MAKE A CHANGE TO THE PRIVATE BUGS
with DB(self.settings.bugzilla) as db:
for b in private_bugs:
old_bug = db.query(
"SELECT * FROM bugs WHERE bug_id={{bug_id}}",
{"bug_id": b})[0]
new_bug = old_bug.copy()
new_bug.bug_status = "NEW STATUS"
diff(db, "bugs", old_bug, new_bug)
#RUN INCREMENTAL
bz_etl.main(self.settings, es, es_c)
#VERIFY BUG GROUP STILL EXISTS
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
now = datetime.utcnow()
results = es.search({
"query": {
"filtered": {
"query": {
"match_all": {}
},
"filter": {
"and": [{
"terms": {
"bug_id": private_bugs
}
}, {
"range": {
"expires_on": {
"gte": CNV.datetime2milli(now)
}
}
}]
}
}
},
"from": 0,
"size": 200000,
"sort": []
})
latest_bugs = Q.select(results.hits.hits, "_source")
latest_bugs_index = Q.unique_index(
latest_bugs, "bug_id") # IF NOT UNIQUE, THEN ETL IS WRONG
for bug_id in private_bugs:
if latest_bugs_index[bug_id] == None:
Log.error("Expecting to find the private bug {{bug_id}}",
{"bug_id": bug_id})
bug_group = latest_bugs_index[bug_id].bug_group
if not bug_group:
Log.error(
"Expecting private bug ({{bug_id}}) to have a bug group",
{"bug_id": bug_id})
if BUG_GROUP_FOR_TESTING not in bug_group:
Log.error(
"Expecting private bug ({{bug_id}}) to have a \"{{bug_group}}\" bug group",
{
"bug_id": bug_id,
"bug_group": BUG_GROUP_FOR_TESTING
})
def full_etl(resume_from_last_run, settings, param, db, es, es_comments,
output_queue):
with Thread.run("alias_analysis", alias_analysis.main, settings=settings):
end = nvl(settings.param.end,
db.query("SELECT max(bug_id)+1 bug_id FROM bugs")[0].bug_id)
start = nvl(settings.param.start, 0)
if resume_from_last_run:
start = nvl(
settings.param.start,
Math.floor(get_max_bug_id(es), settings.param.increment))
#############################################################
## MAIN ETL LOOP
#############################################################
#TWO WORKERS IS MORE THAN ENOUGH FOR A SINGLE THREAD
# with Multithread([run_both_etl, run_both_etl]) as workers:
for min, max in Q.intervals(start, end, settings.param.increment):
if settings.args.quick and min < end - settings.param.increment and min != 0:
#--quick ONLY DOES FIRST AND LAST BLOCKS
continue
try:
#GET LIST OF CHANGED BUGS
with Timer("time to get {{min}}..{{max}} bug list", {
"min": min,
"max": max
}):
if param.allow_private_bugs:
bug_list = Q.select(
db.query(
"""
SELECT
b.bug_id
FROM
bugs b
WHERE
delta_ts >= {{start_time_str}} AND
({{min}} <= b.bug_id AND b.bug_id < {{max}})
""", {
"min": min,
"max": max,
"start_time_str": param.start_time_str
}), u"bug_id")
else:
bug_list = Q.select(
db.query(
"""
SELECT
b.bug_id
FROM
bugs b
LEFT JOIN
bug_group_map m ON m.bug_id=b.bug_id
WHERE
delta_ts >= {{start_time_str}} AND
({{min}} <= b.bug_id AND b.bug_id < {{max}}) AND
m.bug_id IS NULL
""", {
"min": min,
"max": max,
"start_time_str": param.start_time_str
}), u"bug_id")
if not bug_list:
continue
param.bug_list = bug_list
run_both_etl(
**{
"db": db,
"output_queue": output_queue,
"es_comments": es_comments,
"param": param.copy()
})
except Exception, e:
Log.error(
"Problem with dispatch loop in range [{{min}}, {{max}})", {
"min": min,
"max": max
}, e)
def test_changes_to_private_bugs_still_have_bug_group(self):
self.settings.param.allow_private_bugs = True
File(self.settings.param.first_run_time).delete()
File(self.settings.param.last_run_time).delete()
private_bugs = set(Random.sample(self.settings.param.bugs, 3))
Log.note("The private bugs for this test are {{bugs}}", {"bugs": private_bugs})
database.make_test_instance(self.settings.bugzilla)
#MARK SOME BUGS PRIVATE
with DB(self.settings.bugzilla) as db:
for b in private_bugs:
database.add_bug_group(db, b, BUG_GROUP_FOR_TESTING)
es = elasticsearch.make_test_instance("candidate", self.settings.real.bugs)
es_c = elasticsearch.make_test_instance("candidate_comments", self.settings.real.comments)
bz_etl.main(self.settings, es, es_c)
# MAKE A CHANGE TO THE PRIVATE BUGS
with DB(self.settings.bugzilla) as db:
for b in private_bugs:
old_bug = db.query("SELECT * FROM bugs WHERE bug_id={{bug_id}}", {"bug_id": b})[0]
new_bug = old_bug.copy()
new_bug.bug_status = "NEW STATUS"
diff(db, "bugs", old_bug, new_bug)
#RUN INCREMENTAL
bz_etl.main(self.settings, es, es_c)
#VERIFY BUG GROUP STILL EXISTS
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
now = datetime.utcnow()
results = es.search({
"query": {"filtered": {
"query": {"match_all": {}},
"filter": {"and": [
{"terms": {"bug_id": private_bugs}},
{"range": {"expires_on": {"gte": CNV.datetime2milli(now)}}}
]}
}},
"from": 0,
"size": 200000,
"sort": []
})
latest_bugs = Q.select(results.hits.hits, "_source")
latest_bugs_index = Q.unique_index(latest_bugs, "bug_id") # IF NOT UNIQUE, THEN ETL IS WRONG
for bug_id in private_bugs:
if latest_bugs_index[bug_id] == None:
Log.error("Expecting to find the private bug {{bug_id}}", {"bug_id": bug_id})
bug_group = latest_bugs_index[bug_id].bug_group
if not bug_group:
Log.error("Expecting private bug ({{bug_id}}) to have a bug group", {"bug_id": bug_id})
if BUG_GROUP_FOR_TESTING not in bug_group:
Log.error("Expecting private bug ({{bug_id}}) to have a \"{{bug_group}}\" bug group", {
"bug_id": bug_id,
"bug_group": BUG_GROUP_FOR_TESTING
})
def test_recent_private_stuff_does_not_show(self):
self.settings.param.allow_private_bugs = False
File(self.settings.param.first_run_time).delete()
File(self.settings.param.last_run_time).delete()
database.make_test_instance(self.settings.bugzilla)
es = elasticsearch.make_test_instance("candidate",
self.settings.real.bugs)
es_c = elasticsearch.make_test_instance("candidate_comments",
self.settings.real.comments)
bz_etl.main(self.settings, es, es_c)
#MARK SOME STUFF PRIVATE
with DB(self.settings.bugzilla) as db:
#BUGS
private_bugs = set(Random.sample(self.settings.param.bugs, 3))
Log.note("The private bugs are {{bugs}}", {"bugs": private_bugs})
for b in private_bugs:
database.add_bug_group(db, b, BUG_GROUP_FOR_TESTING)
#COMMENTS
comments = db.query("SELECT comment_id FROM longdescs").comment_id
marked_private_comments = Random.sample(comments, 5)
for c in marked_private_comments:
database.mark_comment_private(db, c, isprivate=1)
#INCLUDE COMMENTS OF THE PRIVATE BUGS
implied_private_comments = db.query(
"""
SELECT comment_id FROM longdescs WHERE {{where}}
""", {
"where":
esfilter2sqlwhere(db, {"terms": {
"bug_id": private_bugs
}})
}).comment_id
private_comments = marked_private_comments + implied_private_comments
Log.note("The private comments are {{comments}}",
{"comments": private_comments})
#ATTACHMENTS
attachments = db.query("SELECT bug_id, attach_id FROM attachments")
private_attachments = Random.sample(attachments, 5)
Log.note("The private attachments are {{attachments}}",
{"attachments": private_attachments})
for a in private_attachments:
database.mark_attachment_private(db, a.attach_id, isprivate=1)
if not File(self.settings.param.last_run_time).exists:
Log.error("last_run_time should exist")
bz_etl.main(self.settings, es, es_c)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
verify_no_private_bugs(es, private_bugs)
verify_no_private_attachments(es, private_attachments)
verify_no_private_comments(es_c, private_comments)
#MARK SOME STUFF PUBLIC
with DB(self.settings.bugzilla) as db:
for b in private_bugs:
database.remove_bug_group(db, b, BUG_GROUP_FOR_TESTING)
bz_etl.main(self.settings, es, es_c)
#VERIFY BUG IS PUBLIC, BUT PRIVATE ATTACHMENTS AND COMMENTS STILL NOT
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
verify_public_bugs(es, private_bugs)
verify_no_private_attachments(es, private_attachments)
verify_no_private_comments(es_c, marked_private_comments)
def test_recent_private_stuff_does_not_show(self):
self.settings.param.allow_private_bugs = False
File(self.settings.param.first_run_time).delete()
File(self.settings.param.last_run_time).delete()
database.make_test_instance(self.settings.bugzilla)
es = elasticsearch.make_test_instance("candidate", self.settings.real.bugs)
es_c = elasticsearch.make_test_instance("candidate_comments", self.settings.real.comments)
bz_etl.main(self.settings, es, es_c)
#MARK SOME STUFF PRIVATE
with DB(self.settings.bugzilla) as db:
#BUGS
private_bugs = set(Random.sample(self.settings.param.bugs, 3))
Log.note("The private bugs are {{bugs}}", {"bugs": private_bugs})
for b in private_bugs:
database.add_bug_group(db, b, BUG_GROUP_FOR_TESTING)
#COMMENTS
comments = db.query("SELECT comment_id FROM longdescs").comment_id
marked_private_comments = Random.sample(comments, 5)
for c in marked_private_comments:
database.mark_comment_private(db, c, isprivate=1)
#INCLUDE COMMENTS OF THE PRIVATE BUGS
implied_private_comments = db.query("""
SELECT comment_id FROM longdescs WHERE {{where}}
""", {
"where": esfilter2sqlwhere(db, {"terms":{"bug_id":private_bugs}})
}).comment_id
private_comments = marked_private_comments + implied_private_comments
Log.note("The private comments are {{comments}}", {"comments": private_comments})
#ATTACHMENTS
attachments = db.query("SELECT bug_id, attach_id FROM attachments")
private_attachments = Random.sample(attachments, 5)
Log.note("The private attachments are {{attachments}}", {"attachments": private_attachments})
for a in private_attachments:
database.mark_attachment_private(db, a.attach_id, isprivate=1)
if not File(self.settings.param.last_run_time).exists:
Log.error("last_run_time should exist")
bz_etl.main(self.settings, es, es_c)
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
verify_no_private_bugs(es, private_bugs)
verify_no_private_attachments(es, private_attachments)
verify_no_private_comments(es_c, private_comments)
#MARK SOME STUFF PUBLIC
with DB(self.settings.bugzilla) as db:
for b in private_bugs:
database.remove_bug_group(db, b, BUG_GROUP_FOR_TESTING)
bz_etl.main(self.settings, es, es_c)
#VERIFY BUG IS PUBLIC, BUT PRIVATE ATTACHMENTS AND COMMENTS STILL NOT
Thread.sleep(2) # MUST SLEEP WHILE ES DOES ITS INDEXING
verify_public_bugs(es, private_bugs)
verify_no_private_attachments(es, private_attachments)
verify_no_private_comments(es_c, marked_private_comments)
def full_etl(resume_from_last_run, settings, param, db, es, es_comments, output_queue):
with Thread.run("alias_analysis", alias_analysis.main, settings=settings):
end = nvl(settings.param.end, db.query("SELECT max(bug_id)+1 bug_id FROM bugs")[0].bug_id)
start = nvl(settings.param.start, 0)
if resume_from_last_run:
start = nvl(settings.param.start, Math.floor(get_max_bug_id(es), settings.param.increment))
#############################################################
## MAIN ETL LOOP
#############################################################
#TWO WORKERS IS MORE THAN ENOUGH FOR A SINGLE THREAD
# with Multithread([run_both_etl, run_both_etl]) as workers:
for min, max in Q.intervals(start, end, settings.param.increment):
if settings.args.quick and min < end - settings.param.increment and min != 0:
#--quick ONLY DOES FIRST AND LAST BLOCKS
continue
try:
#GET LIST OF CHANGED BUGS
with Timer("time to get {{min}}..{{max}} bug list", {"min":min, "max":max}):
if param.allow_private_bugs:
bug_list = Q.select(db.query("""
SELECT
b.bug_id
FROM
bugs b
WHERE
delta_ts >= {{start_time_str}} AND
({{min}} <= b.bug_id AND b.bug_id < {{max}})
""", {
"min": min,
"max": max,
"start_time_str": param.start_time_str
}), u"bug_id")
else:
bug_list = Q.select(db.query("""
SELECT
b.bug_id
FROM
bugs b
LEFT JOIN
bug_group_map m ON m.bug_id=b.bug_id
WHERE
delta_ts >= {{start_time_str}} AND
({{min}} <= b.bug_id AND b.bug_id < {{max}}) AND
m.bug_id IS NULL
""", {
"min": min,
"max": max,
"start_time_str": param.start_time_str
}), u"bug_id")
if not bug_list:
continue
param.bug_list = bug_list
run_both_etl(**{
"db": db,
"output_queue": output_queue,
"es_comments": es_comments,
"param": param.copy()
})
except Exception, e:
Log.error("Problem with dispatch loop in range [{{min}}, {{max}})", {
"min": min,
"max": max
}, e)
