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

Apps using gluestick should be able to define/override Jest setup


In theory, I should be able to define "jest" in the app package.json and add options like setupFiles.

  • Option 1: Our test command parses package.json from generated app and options defined by user have precedence
  • Option 2: We can provide a file called jestSetup.js and our test command will look for it if exists. There are some limitations with this approach (setupFiles accepts and array) + not handling other options
Updated 27/03/2017 12:11

Add home page

  • [ ] Show current estimated score
  • [ ] Show highest performance category
  • [ ] Show lowest performance category
  • [ ] Show quiz performance line graph
  • [ ] Show median response time
  • [ ] Show ideal response time
Updated 27/03/2017 05:07

Add a side menu

  • [ ] Add menu button to practice skeleton
  • [ ] Build common side menu component
  • [ ] Add session panel complete with currently authed user
  • [ ] Add lite navigation
  • [ ] Add a sign out button
  • [ ] Add report an issue button
Updated 27/03/2017 05:11

Quest (62) A Spirit Guide


Quest ID 9410

The Ancestral Spirit Wolf (npc=17077) is supposed to ignore all combat and keep walking up the hill. Currently, he doesn’t. If you don’t tag the mob that’s attacking before the dog gets hit, he will get one shot. He only has 300 some odd health. Here’s a link to how this quest is supposed to work

Currently, this quest is unable to be completed. This is marked as a high priority due to the fact that this chain unlocks the Mag'har rep.

Updated 27/03/2017 03:28

Memory Leak in autograd


Here is an example as small as I could make it that triggers what I assume is a memory leak.

import torch
import gc
import resource
from torch import nn
from torch.autograd import Variable

# inp = Variable(torch.Tensor(1,3), volatile=True) # This one does not have a problem
inp = Variable(torch.Tensor(1,3)) # This one does
linear_layer = nn.Linear(3, 6)

iter_idx = 0
while True:
    out = linear_layer(inp).view(-1)
    a = Variable(torch.zeros(6))

    for elt_idx, out in enumerate(out):
        a[elt_idx] = out + a[elt_idx]

    if iter_idx % 10000 == 0:
        max_mem_used = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss
        print("{:.2f} MB".format(max_mem_used / 1024))

    iter_idx += 1

The “network” used here is fairly small, a linear layer with 18 parameters. The backward pass is not even called.

If the input to the network is given as volatile, memory usage is stable (81.23 MB) on my machine. If not, it quickly blows out of proportion, (after 1 min, it already uses 4GB on my machine)

This is probably not very important, I figured out how to work around this in my implementation but I figured I’d still report it.

My setup is the following (everything was run from a virtualenv, pytorch was installed today just for this):

python --version is Python 3.6.0 pip freeze returns appdirs==1.4.3 numpy==1.12.1 olefile==0.44 packaging==16.8 Pillow==4.0.0 pyparsing==2.2.0 PyYAML==3.12 six==1.10.0 torch==0.1.10.post2 torchvision==0.1.7

My OS is a relatively up to date ArchLinux install. Same bug (not the minimal example but the original version) was also observed on my workstation which is an Ubuntu 16.04 install.

Let me know if I can provide more help to debug this.

Edit: I tried the same thing with a python 2.7 install and the results are exactly the same.

Updated 27/03/2017 09:55 2 Comments

Cannot add numerical attrribute as first column


Adding a categorical column first and then adding a numerical one works though.

Uncaught (in promise) TypeError: Cannot read property 'map' of undefined
    at NumberColumn.<anonymous> (AVectorColumn.ts:81)
    at step (tslib.es6.js:91)
    at (tslib.es6.js:72)
    at fulfilled (tslib.es6.js:62)
(anonymous) @ AVectorColumn.ts:81
step @ tslib.es6.js:91
(anonymous) @ tslib.es6.js:72
fulfilled @ tslib.es6.js:62
ColumnManager.ts:486 Uncaught (in promise) TypeError: Cannot read property 'forEach' of undefined
    at ColumnManager.multiformsInGroup (http://localhost:8080/app.js:31386:33)
    at http://localhost:8080/app.js:31615:19
    at Array.forEach (native)
    at ColumnManager.updateAggregationLevelForRow (http://localhost:8080/app.js:31614:22)
    at ColumnManager.<anonymous> (http://localhost:8080/app.js:31422:34)
    at step (http://localhost:8080/app.js:108:23)
    at (http://localhost:8080/app.js:89:53)
    at http://localhost:8080/app.js:82:71
    at Object.__awaiter [as a] (http://localhost:8080/app.js:78:12)
    at ColumnManager.calColHeight (http://localhost:8080/app.js:31407:72)
ColumnManager.multiformsInGroup @ ColumnManager.ts:486
(anonymous) @ ColumnManager.ts:677
ColumnManager.updateAggregationLevelForRow @ ColumnManager.ts:676
(anonymous) @ ColumnManager.ts:518
step @ tslib.es6.js:91
(anonymous) @ tslib.es6.js:72
(anonymous) @ tslib.es6.js:65
__awaiter @ tslib.es6.js:61
ColumnManager.calColHeight @ ColumnManager.ts:508
(anonymous) @ ColumnManager.ts:459
step @ tslib.es6.js:91
(anonymous) @ tslib.es6.js:72
fulfilled @ tslib.es6.js:62
ColumnManager.ts:486 Uncaught (in promise) TypeError: Cannot read property 'forEach' of undefined
    at ColumnManager.multiformsInGroup (http://localhost:8080/app.js:31386:33)
    at http://localhost:8080/app.js:31615:19
    at Array.forEach (native)
    at ColumnManager.updateAggregationLevelForRow (http://localhost:8080/app.js:31614:22)
    at ColumnManager.<anonymous> (http://localhost:8080/app.js:31422:34)
    at step (http://localhost:8080/app.js:108:23)
    at (http://localhost:8080/app.js:89:53)
    at http://localhost:8080/app.js:82:71
    at Object.__awaiter [as a] (http://localhost:8080/app.js:78:12)
    at ColumnManager.calColHeight (http://localhost:8080/app.js:31407:72)
Updated 27/03/2017 06:19 1 Comments

Worker Doing Nothing


Getting reports from @Satriani1980 that his workers are idling. He says it’s happening randomly (as in not every worker).


[9:19:24 PM]18280342 Creep no WORK {"creepName":"worker-800-2","pos":{"x":24,"y":10,"roomName":"W73N89"},"Action":"repairing","Creep":"repairNearby"}
[9:19:24 PM]18280342 Creep run {"creepName":"worker-800-2","pos":{"x":24,"y":10,"roomName":"W73N89"},"Behaviour":"worker","Creep":"run"}

Last Action: undefined.
Current Action: bulldozing

More information:

  • is undefined
Updated 27/03/2017 00:10 6 Comments

Switch to Cloudinary for a fully-featured image service / CDN


Blocks: #159


Currently we are serving our assets directly from an Amazon S3 bucket, not through a CDN. We signed up for the Free Tier service on the 10th of April, 2016, and will start being charged one year after (i.e. soon!).

In addition to this, going forward, we need to prioritise serving images quickly, revealing select image URLs securely in an API (signed), as well as being able to manipulate images easily.

Cloudinary is “a cloud-based service that provides an end-to-end image management solution including uploads, storage, manipulations, optimizations and delivery.” In short, it should make our lives easier, and the free package does the job for us + does NOT expire. Thus, we should move our image management implementation over to Cloudinary. And ideally (but not necessarily) do it before the 10th of April.


There are several, detailed resources stepping through how to integrate Cloudinary with Rails, including how to migrate existing images over from another source:

We may have to move over from using the current Paperclip gem to CarrierWave as this is more supported by Cloudinary. But if we want an intermediate step to migrating, there is a limited Paperclip-cloudinary gem we can look into using first.

  • [ ] Set up a DML cloudinary account and record any auth details in our secure passwords spreadsheet (ask the team to add your email if needed). NB: We currently do not enable S3 locally, which makes it hard to test in development mode before staging / production. We might want to change this, so in local development envs we have the same asset solution but pointing to another local Cloudinary instance. So we may need to set up 3 accounts, local / staging / production
  • [ ] Switch from Paperclip to Carrierwave as our image upload tool
  • [ ] Integrate Cloudinary with our solution following the steps in the docs linked above, moving away from using S3. Ensure local, staging and production envs use the appropriate cloudinary account.
  • [ ] Regression test all the things!
Updated 26/03/2017 19:29

Mobs and NPCs at the Stair of Destiny


The Pitlord and his Felguards are all giving XP upon being killed and they should not. Due to their super low respawn time and the amount they give (2.k +, see screens) this can be exploited VERY easily. Also, when you AOE the mobs, the alliance NPCs (or conversely, the horde) will turn hostile towards you. None of the NPC soldiers on the Stair of Destiny should be attackable. These are game breaking bugs. wowscrnshot_032617_144226 wowscrnshot_032617_144231

Updated 26/03/2017 18:54

PCB 2 - SPI1 Bus Bug


O canal da spi1 da placa 2 está apresentando problemas.

Primeiramente havia um curto entre as linhas do barramento. Foi descoberto que este estava nos pinos do msp (A). Foi resoldado algumas vezes e por fim trocado por um msp(B).

O barramento não apresenta mais curto, porém a leitura com o cartão sd não funciona ainda.

Este barramento é critico pois alem do cartão sd, as memorias e os painéis estão conectados à ele.

Portanto preciso que seja verificado a integridade deste barramento na placa 2.

André e Arthur devem fazer testes usando uma launchpad para tentar se comunicar com o msp via spi. Devem observar as formas de onda no analizador lógico ou osciloscópio.

O André fará o código na launchpad e o Arthur(que ja esta fazendo um driver simples de spi) o ajudará no processo.

Se o problema for na pcb (trilhas do barramento) perderemos varias funções desta placa ou ela por inteiro.

PS: o msp(A) foi soldado na terceira placa junto com os circuitos do cartão sd, e não apresentaram problema nenhum. Portanto o problema não é interno ao chip ou ao cartão sd. E sim à placa 2.

Updated 26/03/2017 14:55

Manufacturing Order UI


Is this a bug or feature request?

Feature Request

What is the current behavior?

Currently, there is no Manufacturing Order UI Window in metasfresh.

What is the expected or desired behavior?

Create a generic window w/ Grid View that lists all nonprocessed Manufacturing Orders. This Window is the starting point for

Updated 26/03/2017 11:25

API for Manufacturing UI Planning/ Doing


Is this a bug or feature request?

Feature Request.

What is the current behavior?

Currently, there is no API to provide the action issue/ receipt for manufacturing in metasfresh WebUI. Important idea: when issuing stuff from HU to planned manufacturing, then the split off/ used HU shall somehow only be blocked (not directly destroyed). This shall allow the user to manipulate until the 4-eyes check has been done. This will improve the flexibility for the user dramatically.

What is the expected or desired behavior?

Concept and implement the provided API. See for further UI details.

Updated 26/03/2017 11:22

Manufacturing UI Planning/ Doing


Type of issue

Feature request

Current behavior

Currently, there is no Manufacturing UI in metasfresh.

Expected behavior

We need a Manufacturing UI that provides a nice and comprehensive way for the user at the machine/ manufacturing resource to do the work in a transparent and adjustable way.

Workflow description

  • The starting point is a generic grid view that shows all the Manufacturing Orders for filtered criteria (Warehouse, Date, Manufacturing Ressource, BOM Product …)
  • The user selects a line and presses the main “manufacturing” action button
  • A modal overlay opens. The user sees the BOM explosion of the to-be-produced-bom-product selected in first window. The first line is the BOM product itself. all other lines in the indented hierarchy are the bom lines components. Co-Products/ By-Products are not indented but on same level as BOM-Product.
  • If the user selects one of those component lines, then on the right side of the screen a HU Editor (similar to the one used in material receipt) is visible that shows all the Handling Units in the same Warehouse that can be used to issue for that selected component. The user can drag the Handling Unit and pull onto the left side (selected component line). A configuration window pops open (similar to stuff in Material Receipt) where the user can define how many units shall be issued.
  • If the user selects a BOM-Product/ By-Product/ Co-Product line, drags that one on the Handling Unit editor on the right side, then automatically a default HU/TU/CU Configuration shall be created for the user (similar to receive step in material receipt candidates.

@cadavre Just first ideas for now. Can we talk about the further ideas/ steps to be taken (user experiance, design)?

Updated 26/03/2017 11:17

Remove flickering for relayout actions


Every time we calculate the layout, the whole table flickers. Is there a way we can handle this? Maybe delay the rendering until the layout calculation is done? Or other ideas?

The perfect solution would of course be to have animated transitions explaining the changes. But I know that his will be difficult to achieve with the current implementation (also see #81).

Updated 26/03/2017 15:43 1 Comments

Spawn error


Invalid creeps can get stuck in your spawnQueues

 Offspring failed: ERR_INVALID_ARGS
 - body: {"attack":9,"heal":8,"move":18,"ranged_attack":8,"tough":8}
 - name: guard-E85S26-guard-1
 - behaviour: warrior
 - destiny: [object Object]
        let result = this.create(,, params.behaviour || params.setup, params.destiny);
        if( !result ){
   = stumb;
        return result;

I think we need to add queue.pop() after the unshift to ensure they don’t stay at the top of the queue.

Updated 26/03/2017 03:36

Use a Database

  • [x] Design the DB
  • [ ] Learn how to test a DB?
  • [ ] Write DB tests?
  • [ ] Implement SQL commands
  • [ ] Implement Java code
  • [ ] Important Feature: Group by Deadline, Sort by PercentComplete; Sort individually (deadline, % complete)
  • [ ] Ensure that IDs for Nodes do not overflow?
Updated 25/03/2017 19:15

Allow .tbi and .bai index files to be specified in a separate URL.


Chase: I’ve added a section to the bam url input to explicitly add the bai url. this is for when the bai url is not just the bam url + .bai. i have an updated version of od_samtools on nv-dev-new that works with the DNANexus data

so basically if the bai url is explicitly set then use od_samtools and add the bai url at the end of the command. we need to do the same thing for tabix_od and tbi url’s. i’ve updatd the tabix_od service on nv-dev-new to work as well, but I haven’t had a chance to test it yet

Updated 25/03/2017 21:59

Adding second account breaks Google Drive


Software and hardware versions

Collect v1.4.16, Android 6.0, Moto X Pure

Problem description

When I add a second gmail account, the collect break with NPE.

Steps to reproduce the problem

  • Ensure your google account is not connected to ODK yet: Google Account Permissions (If you have not visited this in a while, now is a good time to go through it).
  • In general settings, change the platform to “Google Drive, Google Sheets”
  • Select your primary gmail account for the google account.
  • Go to get blank form section and press “My Drive”
  • It will ask you to allow Collect to access your drive.
  • Add a second gmail account to your phone: Add Account.
  • Change your google account in the general settings to the new account.
  • Go to the get blank form section and press “My Drive” and Collect will crash.
  • Adding a third gmail account strangely make Collect work again.

Expected behavior

Collect should allow adding the second account without crashing.

Other information

This might not even be issue on the Collect side.

Updated 27/03/2017 11:14 9 Comments

Wrong total number of groups on GCconnex login page


The number of GCconnex groups that appears on the GCconnex login page is incorrect. It currently shows 3143 groups, but there is in fact 7248 groups on GCConnex (confirmed by Chris Lavoie). Please correct the number that appears on the login page as this provides wrong data to users and does not accurately reflect the content/activity on the tool.

image of GCconnex login page with wrong number of GCconnex groups. total number groups gcconnex login page

Updated 24/03/2017 17:22 1 Comments

Create Integration: Tooltip Should Remain Open First Time


This currently requires that the user hover over the plus icon, but should remain open by default for the first time.

<img width=“1434” alt=“screenshot 2017-03-24 11 05 58” src=“”>


save or add step


Updated 24/03/2017 15:10

Find new way to link to teams en-masse quickly


Currently, to link to a team’s page, we use this format in the templates:

<a href="{% url 'team_view' %}">{{ team }}</a>

This is incredibly slow on pages with lots of links, such as /e/, the events list page, where a tags are generated as so:

{% with event.1.teams.all as a %}
        <a href="{% url 'team_view' %}">{{ a.0 }}</a><br>
        <a href="{% url 'team_view' %}">{{ a.1 }}</a><br>
        <a href="{% url 'team_view' %}">{{ a.2 }}</a>
{% endwith %}

where event.1 is simply accessing a tuple of the form ( ... , event.winning_alliance, ... ... ).

This bogs down page load time by a large margin and we should optimize pages like this.

Updated 25/03/2017 22:43 2 Comments

Different date information in admin and editor screens



In the Admin screen, I can see Last Updated dates of November and December, but in an editor’s screen, and in the Editor view, the oldest Last Updated dates are 28 February 2017, and just a few seconds apart. This suggests some kind of automated update.

It is not helpful for the information in the system to differ between screens. Do you know what happened?



Updated 27/03/2017 06:40 2 Comments

Page des livraisons (admin)


L'administrateur disposera d'une nouvelle page afin de pouvoir: * Saisir les informations d'une nouvelle livraison à savoir: * Le verger de provenance * La date de réception * Le type de la noix reçu

  • Avoir la vue sur la liste de toutes les livraisons effectuées.
    • + Bouton pour corriger
    • + Bouton pour supprimer l'information
Updated 24/03/2017 10:08

spark.executor.extraJavaOptions only support one option when submitting spark job with


when I use dist/bin/ -- spark-submit \ --conf 'spark.executor.extraJavaOptions=-XX:+PrintGCTimeStamps -XX:+PrintGCDetails'

only the first jvm option is recognized

current work around is to first source, then submit with the following command source spark-submit \ --conf 'spark.executor.extraJavaOptions=-XX:+PrintGCTimeStamps -XX:+PrintGCDetails'

Updated 27/03/2017 03:35

from_igraph() drops vertex character names


See this StackOverflow post. I reproduced the issue for the current CRAN version (0.9.0) and commit a77fc20d13a3ad654148fe37a7ac3e5ae16bb5a1.

g = make_empty_graph() + vertices(c("a", "b")) + edge("a", "b")
x = from_igraph(g)

The drake package internally relies on igraph objects with meaningful vertex names, and DiagrammeR has such powerful visuals.

Updated 24/03/2017 22:15

Change the name of the Brackets Node executable



I had trouble to install some extensions on OSX. The problem was that the node command wasn’t found. I think the problem is the executable name of node process of Brackets: on Windows is node.exe, and on Windows 10 I don’t have any problem, while on macOS and Linux is Brackets-node. I don’t have proof this is the real problem but to my eyes seems very likely.

On there were some discussion to rename the process on Windows. Maybe we should instead rename the others?

Steps to Reproduce

  1. Go to the extension manager
  2. Try to install the npm registry extension

Expected behavior: the extension should be installed

Actual behavior: the installation fails


OSX 1.9

Updated 23/03/2017 19:51

Strings are not always localized


Reported by @glueckpress on the call-for-feedback post:

One thing I noticed when testing the plugin with different languages is that it doesn’t support user language yet. When my site language is default en_US, and I switch my user language to de_DE, the core dashboard widget would localize the news feed. This plugin does not do that (yet), it always shows news from the English feed.

I’m seeing the news items localized (e.g., WordPress 4.7.3 Sicherheits- und Wartungs-Release), but not the rest of the text (e.g., Attend an upcoming event near...)

Updated 24/03/2017 09:17 2 Comments

[Request] Add Material UI


Library name: Material UI Git repository url: npm package name or url (if there is one): material-ui License (List them all if it’s multiple): MIT Official homepage: Wanna say something? Leave message here: add package plz

thx for the work

Notes from cdnjs maintainer: Please read the and document first.

We encourage you to add a library via sending pull request, it’ll be faster than just opening a request issue, since there are tons of issues, please wait with patience, and please don’t forget to read the guidelines for contributing, thanks!!

Updated 26/03/2017 12:25

Non-determinisstic matrix scaling


Upon adding the matrix to the table, the matrix is often (not always) scaled incorrectly and also breaks the size of the whole table (except for the agg. switcher column) image After next re-layout call, the scaling is again correct. Seems that when the matrix is added and the re-layout is called, the data are not there yet.

Updated 26/03/2017 15:51 3 Comments

[V1] Can't start the app in production mode


gluestick version 1.0.0-rc.13

gluestick new TestApp
cd TestApp
NODE_ENV=produciton gluestick start

It throws:

Starting to optimize CSS...
[GlueStick][start-server] Building server entry.
      throw er; // Unhandled 'error' event

Error: spawn /Users/ferrannp/Desktop/TestApp/node_modules/.bin/pm2 ENOENT
    at exports._errnoException (util.js:1012:11)
    at Process.ChildProcess._handle.onexit (internal/child_process.js:182:32)
    at onErrorNT (internal/child_process.js:348:16)
    at _combinedTickCallback (internal/process/next_tick.js:74:11)
    at process._tickCallback (internal/process/next_tick.js:98:9)
Updated 27/03/2017 09:50

Elsevier failing metadata


Hello CL

There is apparently a problem with these journals from Elsevier. I have asked to provide me with the failing xmls.

Is there something in the setup for these four journals that would have prevented the upload of metadata for them? We’ve repeatedly encountered errors when we try to upload data.


Updated 23/03/2017 09:12 1 Comments

setWriteListener causes race


Another variation of #1047 has been detected.

If HttpInput.setReadListener is first called from a container dispatched thread when there is no input available, then when the container dispatched thread eventually calls unhandle, interest in filling will be set with a call to channel.asyncReadFillInterested()

Then is HttpOutput.setWriteListener is called from a non container dispatched thread, then because the channel is in ASYNC_WAIT state, it is woken up so that onWritePossible can be called for the first time. This is done by dispatching a thread to call channel.handle.

We now have a race between the actual execution of the channel.handle and possible arrival of a succeeded callback for the asyncReadFillInterested call.

This race has at least two bad outcome. Firstly the succeeded callback does a call to fillAndParseForContent, which is not protected by mutual exclusion, so it may enter the parse together with the thread that calls the handle method and ends up doing a call to the application code that does a read(). The parser is not reentrant, so this can result in arbitrary corruption while parsing.

Secondly, if the call to fillAndParseForContent returns false, the succeeded callback does not know if this is because insufficient data has been received to be content or if the content was received but the app decided it does not yet need to read anymore. So the callback may then call channel.asyncReadFillInterested() as it guesses it needs more data. However, the other thread may have already called unhandle and registered fill interest, so a Pending read exception may be generated.

Updated 23/03/2017 05:03 3 Comments

Creating an update with --autokarma and without the --stable and --unstable-karma flags sets autokarma off


Today I created an update with the CLI like this:

$ bodhi updates new --user bowlofeggs --type bugfix --notes "The bodhi-client package no longer depends on yum." --bugs 1135681 --close-bugs --autokarma bodhi-2.4.0-3.fc26
  Update ID: FEDORA-2017-e2e877073e
    Release: Fedora 26
     Status: pending
       Type: bugfix
      Karma: 0
  Autokarma: False  [-3, 3]
    Request: testing
       Bugs: 1135681 - None
      Notes: The bodhi-client package no longer depends on yum.
  Submitter: bowlofeggs
  Submitted: 2017-03-22 22:30:16
   Comments: bodhi - 2017-03-22 22:30:16 (karma 0)
             This update has been submitted for testing by

Note that the Autokarma is False in the resulting text, and if I go to that update Autokarma is displayed as being off there as well. I would expect that the --autokarma flag would set it to be on.

1298 may be related to this (and may even be the same underlying problem) - it reports that setting --stable-karma without specifying --autokarma sets autokarma to be on. Given this, my guess is that something is using the karma thresholds to determine whether to enable autokarma instead of the --autokarma flag.

Updated 23/03/2017 12:59 1 Comments

The CLI should not traceback when trying to save an already active override


I tried to use the CLI to alter the duration on an override I made today, and it crashes:

$ bodhi overrides save --duration 3 --notes "This is needed to build Bodhi." --user bowlofeggs python-pyramid-1.8.2-2.fc26
Traceback (most recent call last):
  File "/usr/bin/bodhi", line 11, in <module>
    load_entry_point('bodhi-client==2.4.0', 'console_scripts', 'bodhi')()
  File "/usr/lib/python2.7/site-packages/click/", line 722, in __call__
    return self.main(*args, **kwargs)
  File "/usr/lib/python2.7/site-packages/click/", line 697, in main
    rv = self.invoke(ctx)
  File "/usr/lib/python2.7/site-packages/click/", line 1066, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/usr/lib/python2.7/site-packages/click/", line 1066, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/usr/lib/python2.7/site-packages/click/", line 895, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/usr/lib/python2.7/site-packages/click/", line 535, in invoke
    return callback(*args, **kwargs)
  File "/usr/lib/python2.7/site-packages/bodhi/client/", line 229, in save_buildroot_overrides
    resp = client.save_override(nvr=nvr, duration=duration, notes=notes)
  File "/usr/lib/python2.7/site-packages/bodhi/client/", line 78, in wrapper
    raise BodhiClientException(problems)
bodhi.client.bindings.BodhiClientException: Buildroot override for python-pyramid-1.8.2-2.fc26 already exists

It should be possible to use the CLI to edit active overrides, just like the web UI can do.

Updated 22/03/2017 21:49

Fork me on GitHub