Contribute to Open Source. Search issue labels to find the right project for you!

DC: containter upgrade procedure


Implement an upgrade procedure for nsdc container.

Changes - Add a new nethserver-dc-upgrade event - The event must update the Samba installation and the whole container - Also upgrade to Samba 4.6.0

Updated 27/03/2017 09:34 1 Comments

testExamineEventAndCompileReport failing


Plaso version:

version 1.5.2_20170325

Operating system Plaso is running on:

Ubuntu 16.04

Installation method:

Github clone. All dependencies installed and tested with check_dependencies

Description of problem:

Error during “” command after “python install” of the new set of plaso files.

FAIL: testExamineEventAndCompileReport (analysis.tagging.TaggingAnalysisPluginTest)
Tests the ExamineEvent and CompileReport functions.
Traceback (most recent call last):
  File "/root/plaso/tests/analysis/", line 105, in testExamineEventAndCompileReport
    self.assertEqual(len(storage_writer.event_tags), 4)
AssertionError: 3 != 4


Please describe the source that is causing the issue.

Updated 27/03/2017 05:10 1 Comments

Move sample_files and combine with tutorial


Right now we have sample configs/submissions spread out in the sample_files folder in this repository and then an additional Tutorial repository that contains some.

We should move this all to one repository (sample_files?) that contains all of this in one standardized location and we use for samples. We should the have a separate repo that contains all the configs and things that we use purely for testing in this repo.

Updated 26/03/2017 23:38 1 Comments

Make image_comparison more pytest-y


I’m a bit tired of image_comparison breaking pytest subtly. It was written in a way that mirrors nose with the test as “setup” and the “test” is just saving and comparing the results. This has a number of disadvantages.

However, switching to a decorated function had a problem; we couldn’t change the signature of the decorated function or pytest would get confused about where to send the fixtures. So we couldn’t, for example, parametrize based on extension. The trick I’ve found is that the test doesn’t need to take a fixture if one of its fixtures does. So I hacked this together by having the fixture send the parameters back to the wrapper via the decorated function. Now, image_comparison is simply another decorator as far a pytest is concerned (because it doesn’t modify the function signature.)

This method has a number of advantages: * Test code is actually run as the test phase, not the setup phase. This makes more sense semantically: * Problems in test code is a failure instead of an error. * Stuff like #8346 will work correctly now because it assumes test code is actually in the test phase and doesn’t try to capture warnings from the setup phase. * No signature changes means we can use any fixture we want on an image test. (Marks already worked, but now we can ask for, e.g., monkeypatch or something else as an argument.) * Image comparisons can be parametrized. At the moment, I’ve only setup baseline_images as a possible fixture, but this could be extended. This can be seen being used in the mathtext tests in the last commit.

There’s only one obvious (to me) disadvantage: * Previously, the test phase was just comparisons, so we could parametrize this by baseline image name. Now with the test phase as figure creation and comparison, we can’t parametrize on image name. Since we already parametrize on extension, I think this is a small price to pay.

Updated 27/03/2017 03:53

Python 3.6 issue, tests segfault


Hoping to get python 3.6 support going today but ran into a curious blocker - Tests immediately segfault.

⭆ mkvirtualenv -p python3.6 rasterstats36
⭆ python --version
Python 3.6.1
⭆ pip install --pre 'rasterio>=1.0a7'  # Also tried with master and src_dist build of 0.36
⭆ pip install -r requirements_dev.txt
⭆ pip install -e .
⭆ py.test -v
====================================================== test session starts ======================================================
platform darwin -- Python 3.6.1, pytest-3.0.7, py-1.4.33, pluggy-0.4.0 -- /Users/mperry/env/rasterstats36/bin/python3.6
cachedir: .cache
rootdir: /Users/mperry/projects/python-rasterstats, inifile: setup.cfg
plugins: cov-2.4.0
collecting 0 itemsSegmentation fault: 11

Same procedure passes 👌 on python 3.5.2.

Updated 26/03/2017 14:30

Feature 8: Select Phrase as Actor


Feature Description

This will involve creating the select_phrase.html page (as stated in website design). most of the code for the functionality has been implemented in the Actor class, but might require minor tweaks to create functional code.


  • Entry tasks:
    • None
  • Design feature:
    • [x] Consult overall model and decide if change is required
    • [x] Create UI diagram
  • Build by feature:
    • [ ] Write tests for feature
    • [ ] Write code to pass tests
    • [ ] Refactor where required
Updated 26/03/2017 12:29

TST/BUG: HDFStore.append_to_multiple with dropna=False buggy


xfaile this test:

in pandas/tests/io/ ``` @pytest.mark.xfail(run=False, reason=“append_to_multiple_dropna_false ” “is not raising as failed”) def test_append_to_multiple_dropna_false(self): df1 = tm.makeTimeDataFrame() df2 = tm.makeTimeDataFrame().rename(columns=lambda x: “%s_2” % x) df1.iloc[1, df1.columns.get_indexer([‘A’, ‘B’])] = np.nan df = concat([df1, df2], axis=1)

    with ensure_clean_store(self.path) as store:

        # dropna=False shouldn't synchronize row indexes
            {'df1a': ['A', 'B'], 'df2a': None}, df, selector='df1a',

        with pytest.raises(ValueError):
            store.select_as_multiple(['df1a', 'df2a'])

        assert not'df1a').index.equals(


This sometimes succeed and sometimes fails. Needs some investigation.

Updated 25/03/2017 17:13

Failing test on merge


Failing test on merge, seems to be a flaky test

FAIL: testEmptyBufferedQueues (engine.zeromq_queue.testZeroMQQueues)
Tests the Empty method for buffered queues.
Traceback (most recent call last):
  File "tests/engine/", line 137, in testEmptyBufferedQueues
    queue.PushItem(u'item', block=False)
AssertionError: QueueFull not raised

Updated 26/03/2017 08:07 1 Comments

cmd/compile: defer+delete argument bug


Please answer these questions before submitting your issue. Thanks!

What version of Go are you using (go version)?

Go 1.8. Also happens on

What operating system and processor architecture are you using (go env)?

GOARCH=“amd64” GOBIN=“” GOEXE=“” GOHOSTARCH=“amd64” GOHOSTOS=“darwin” GOOS=“darwin”

What did you do?

Evaluate this code:

package main

import "fmt"

func main() {
    fmt.Println("should be 0:", len(foobar()))

func foobar() map[int]bool {
    m := map[int]bool{}
    for i := 0; i < 3; i++ {
        m[i] = true
        defer delete(m, i)
    return m

What did you expect to see?

should be 0: 0

What did you see instead?

should be 0: 2

The defer statement is supposed to freeze its arguments the instant it is evaluated, but it does not when used with delete. Compare to this program, which works correctly:

package main

import "fmt"

func main() {
    fmt.Println("should be 0:", len(foobar()))

func foobar() map[int]bool {
    m := map[int]bool{}
    for i := 0; i < 3; i++ {
        m[i] = true
        defer myDelete(m, i)
    return m

func myDelete(m map[int]bool, k int) {
    delete(m, k)
Updated 25/03/2017 18:15 7 Comments

Feature dark theme


Pugin - User changeable theme

This plugin provides the possibility to change forum theme by user or a visitor. Widget with switch buttons should be placed below the categories in the side panel.


Now, we can create a custom themes in the /qa-plugins/user-changable-theme/themes/ directory. All new styles must extend the SnowFlat theme.

The dark theme was created - Add transitions - Add several icons as base64 images - Fix some visual bugs


Moreover, I’ve decided to reduce the length of question description. That’s supposed to prevent text wrapping, which often breaks the layout.


pytanie zadane 9 minuty temu w Android, Swift, Symbian przez użytkownika xyz Nowicjusz (120 punkty)


pytanie zadane 9 minuty temu w Android, Swift, Symbian przez xyz Nowicjusz (120 p.)

Updated 26/03/2017 21:55 1 Comments

Remove php 7.1 from allowed failures



  • Type: code quality
  • Link to issue:

In raising this pull request, I confirm the following (please check boxes):

  • [x] I have read and understood the Contributing Guidelines?
  • [x] I have checked that another pull request for this purpose does not exist.
  • [x] I wrote some tests for this PR.

Small description of change: This removes php 7.1 from allowed failures to test php 7.1


Updated 26/03/2017 15:17 3 Comments

As a user, I should be able to receive notifications of recent updates on my home page


Priority: High Risk: High Points: 8

  • [ ] Display a transparent box on the home page, where the user will see their recent changes (Front-end) Hours: 2 hours
  • [ ] Every time a TA/Prof edits a course description, store a notification in the path /notifications/[user UID]/[List of unique keys]/Notification_object on Firebase for EVERY subscriber to the edited section. Each notification object should include, email of person making the edit, course details and new description (and maybe old description). Hours: 4 hours
  • [ ] When a user logs in, download all the recent notification objects and listen for changes using the full Redux pipeline. Hours: 5 hours
  • [ ] Push the notifications to the transparent box as clickable cards (Front-End) Hours: 4 hours
  • [ ] When a notification card is clicked on, delete it locally and on Firebase. Hours: 2 hours

Acceptance Criteria:

Updated 26/03/2017 04:38

AT: Write Tests for Session Authentication


Task Description

Implement code to get a valid return value from the authenticateSession function in order to test the valid and invalid authentication cases.

Task Breakdown

  • [ ] obtain valid return from authenticateSession function
  • [ ] test valid authentication case
  • [ ] test invalid authentication case

Agile Information

Priority: medium Risk: Low Story Points: 3

Updated 24/03/2017 15:18

Tests related to .ackrc fail on Windows



#   Failed test 'don't load the same ackrc file twice'
#   at t/config-finder.t line 226.
#     Structures begin differing at:
#          $got->[3] = HASH(0xe9a1c4)
#     $expected->[3] = Does not exist
# [
#   {
#     'path' => 'C:/Documents and Settings/All Users/Application Data/ackrc'
#   },
#   {
#     'path' => 'C:/Documents and Settings/c/Application Data/ackrc'
#   },
#   {
#     'path' => 'C:/tmp/NSSQ2NYQyP/foo/.ackrc'
#   },
#   {
#     'path' => 'C:/tmp/NSSQ2NYQyP/foo/.ackrc',
#     'project' => 1
#   }
# ]
# Looks like you failed 1 test of 26.
#   Failed test at t/ack-o.t line 128.
# []

#   Failed test at t/ack-o.t line 129.
# [
#   'Options --output, --pager and --match are forbidden in project .ackrc files.',
#   'ack: Invalid option in C:\\tmp\\F9DOfwxuyC\\.ackrc'
# ]

#   Failed test at t/ack-o.t line 144.
# []

#   Failed test at t/ack-o.t line 145.
# [
#   'Options --output, --pager and --match are forbidden in project .ackrc files.',
#   'ack: Invalid option in C:\\tmp\\F9DOfwxuyC\\.ackrc'
# ]
# Looks like you failed 4 tests of 12.
t/ack-o.t .................... 
Dubious, test returned 4 (wstat 1024, 0x400)
#   Failed test at t/ack-match.t line 63.
# []

#   Failed test at t/ack-match.t line 64.
# [
#   'Options --output, --pager and --match are forbidden in project .ackrc files.',
#   'ack: Invalid option in C:\\tmp\\Vgx4THDhNe\\.ackrc'
# ]

#   Failed test at t/ack-match.t line 79.
# []

#   Failed test at t/ack-match.t line 80.
# [
#   'Options --output, --pager and --match are forbidden in project .ackrc files.',
#   'ack: Invalid option in C:\\tmp\\Vgx4THDhNe\\.ackrc'
# ]
# Looks like you failed 4 tests of 14.
Updated 24/03/2017 03:18 1 Comments

HiPE: Fix off-by-one bug in register allocators


The register allocators accept a low-water mark called SpillLimit which indicates which temporaries they must not spill. However, it turns out they do not agree with hipe_regalloc_loop, which provides SpillLimit, on whether it is an inclusive or exclusive bound.

This caused the register allocators to occasionally spill the first “unspillable” temporary. This caused a failure in a newly added assertion, added by #1380, when hipe-compiling dets_v9 on x86.

Updated 24/03/2017 10:51

[Meta] Add nettest tests to all conn implementations


The go semi-standard library has recently added a set of tests to automatically run on anything that implements net.Conn. We should run these over all of our streams and conns to make sure everything is up to snuff. Some initial things I want to test:

  • [ ] yamux
  • [ ] spdy
  • [ ] multiplex
  • [ ] secio
  • [ ] swarm stream
Updated 23/03/2017 17:17

DC: add NTP server support


Windows and Linux clients joined to an Active Directory should use the AD server itself as NTP server.


  • Install ntp RPM inside the container
  • Enable and start ntpd server


See also

Updated 24/03/2017 16:33 1 Comments

DC: allow changing container IP


In certain scenarios the administrator needs a simple way to change the IP address of Samba Active Directory guest container.

After the IP change, the host must be correctly joined to the guest.

See also

Updated 27/03/2017 09:54 1 Comments

add TPR, cleanup support to walkthrough e2e script


1) Now will use a tpr-backed API server if you pass in --with-tpr. 2) --cleanup will get rid of all resources spun up during walkthrough.

Although the output is still quite ugly, this allows for people to test the walkthrough easily on their local machine.

Jenkins will still only run the etcd-based walkthrough; there is an issue outstanding ( #600 ) that prevents the deletion of TPR resources. EDIT: Now Jenkins will run the tests on a TPR-based walkthrough as well up until deletion, and then hackily deletes the resources manually.

Updated 24/03/2017 20:24 4 Comments

Need to create adapter test suites


If we want 3rd parties to create adapters for the mixer (goal for Q2), and for our own use, we must create adapter test suites that validate that the common expected functionality for each type of adapter is working as expected.

So for example, we can have a generic quota test suite to ensure that both memQuota and redisQuota adhere to the same semantics.

Updated 23/03/2017 00:18 2 Comments

Test helpers


It could be useful to have tools for testing email-related stuff. E.g. postmarker fixture for pytest which will not send anything - just to inspect API calls. Probably, something from the test suite could be exposed.

Updated 22/03/2017 12:36

Create data generator class for unit testing

  • Unit tests currently all set up their own data
  • result: redundant code, larger binaries, etc
  • Task: Create a class in unit test that can generate data for all kinds of tests we are using
  • Then, deploy the class in unit tests, i.e. call its methods rather than generating by hand everytime.


Updated 22/03/2017 09:28

Allow capital letters in shared folder name


The validator of shared folder name is too restrictive.

Current validator

The names of shared folders are validated by USERNAME validator that is too restrictive.


Current validator will be replaced by a new validator.

The new validator checks if the name you entered meets the Windows naming conventions.

Naming conventions can be checked here:

See also

Updated 22/03/2017 14:06 1 Comments

Missing Value Error is not being caught when there are missing values in the matrix


If there are any missing values in the matrix, a missing value error should be returned. Right now, a missing value error is not being returned. It also seems like the missing value error is being caught in the wrong spot - almost as if it’s a catch-all for all of the problems that could arise in the matrix. In the spot that it looks like a missing value error should be caught, it seems as though an invalidMatrixData warning is being returned instead. However, in all missing value test files, there are no errors or warnings being returned at all.

@mihirsamdarshi, passing this one over to you. Making this priority 0, as we really should not be allowing any missing values at all, and I’m not fully sure what’s going on with the invalidMatrixData warning.

Updated 22/03/2017 05:27

Fork me on GitHub