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

Update Readme


Update Readme to show schema for RGB and Button data

// Values for R/G/B in range 0-255
    "type": "rgb",
    "value": { 
        "red" : 230, 
        "green" : 10,
        "blue" : 70
// Only support 1 button for now, which is power button
    "type": "button",
    "value": true
Updated 19/08/2017 15:51

/language/objects wrongly claims that `self` is a type object inside `submethod BUILD`


from /language/objects#self:

Since user defined constructors are called on type objects, self will point to the type object in constructors and the submethod BUILD. Initializers are also called before object construction time. This can be used to access class attributes in initializers.

This is not actually the case. self is not a type object there, it is a partially constructed object which counts as .defined:

class A {
    submethod BUILD { say self.defined; }
};  # True

Indeed, virtual method calls with the syntax work there.
(The $.foo method call syntax doesn’t, to protect users from writing buggy code.)

Updated 19/08/2017 13:17

The shallowness of `.clone` needs a more precise explanation.


objects#Object_Cloning says:

This cloning is shallow since it only binds attributes to the same values contained in the original object; it does not make copies of those values.

I don’t think that’s completely accurate.

A $. attribute does get a fresh Scalar container, and the value from the old object’s attribute gets assigned into it:

class A { has $.x is rw }

my $a = => 2);
my $b = $a.clone;

$b.x = 10;

say $a;  # => 2)
say $b;  # => 10)

…whereas a @. attributes doesn’t get a fresh Array container, but rather gets the Array from the old object’s attribute bound to it.

So it’s more like how arguments are passed to $ and @ parameters in a signature, than simple := binding.
Not sure how to properly word that for the docs.

Updated 19/08/2017 08:40

Add some fancy meta tags for sharing etc


A good example of this is how Feather Icons displays custom title, descriotion and even a screenshot when their link is shared on Twitter:

<meta name="description" content="Designed on a 24x24 grid with an emphasis on functionality, consistency and simplicity.">
<meta property="og:title" content="Feather – Simply beautiful open source icons">
<meta property="og:description" content="Designed on a 24x24 grid with an emphasis on functionality, consistency and simplicity.">
<meta property="og:image" content="">
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:site" content="@colebemis">


Updated 19/08/2017 00:21

Better document bolometric_flux for BlackBody1D


Currently in BlackBody1D model python temperature = Parameter(default=5000, min=0, unit=u.K) bolometric_flux = Parameter(default=1, unit=u.erg / ** 2 / u.s) but to get it to produce the same result as blackbody_nu function, I need to give the reverse of normalization used in the model evaluation (see python bolometric_flux = const.sigma_sb * temperature ** 4 / np.pi

I can see how setting the latter as default is problematic because it depends on the other parameter, temperature. At the very least, this should be better documented, perhaps in BlackBody1D docstring. What do you think, @astrofrog ?

Updated 19/08/2017 12:52

Expand adoc description of batch transfer mode + `--recursive` flag


This relates a little to globusonline/

Someone mentioned in a ticket that he found the “Using the CLI” doc inconsistent with the reference. If my guess is right, this is because we’re not sufficiently clear in the reference documentation that command-line --recursive and batch-mode --recursive are different and mutually exclusive.

We may want to further clarify with a one or two examples on the reference page, as in shell $ globus transfer --batch ... xyz def --recursive a b c d --recursive

Updated 18/08/2017 18:52

Change conda install instructions to conda-forge channel


The Astropy conda packagers recently moved Gammapy and all the other conda packages previously maintained by the Astropy team to conda-forge.

In the past years, conda-forge has emerged as the most popular / well-organised solution for conda packaging / build / distribution. To quote from

A community led collection of recipes, build infrastructure and distributions for the conda package manager.

Here’s the conda-forge package for Gammapy:


I think this is the long-term conda solution for Gammapy as an open source project for the coming years (irrespective if others like CTA start building or distributing conda packages).

This PR updates the Gammapy docs to use conda-forge, getting rid of mentions of the astropy or openastronomy channels. (but doesn’t do other improvements, like improving our install docs or merging the content from into the existing Sphinx install docs).

Updated 18/08/2017 15:29

Version string confusion for 0-versions


Reading the section on Semantic Versioning in the project readme, it seems the description (a) is in conflict with the examples (b) in that 0-versions seem to be treated differently.


Note: When you specify a version without an operator, dep automatically uses the ^ operator by default.


  • 1.2.3 becomes the range >=1.2.3, <2.0.0
  • 0.2.3 becomes the range >=0.2.3, <0.3.0
  • 0.0.3 becomes the range >=0.0.3, <0.1.0

Following (a), a constraint"0.2.0" would be equivalent to "^0.2.0", yet dep will not, as it seems, fetch a version 0.3.0. Digging deeper,

Refer to the semver library documentation for more info.

The semver library seems to accept 0.3.0 as valid for the constraint ^0.2.0.

If dep behaves as intended in that it does not accept 0.3.0 for a version string saying "0.2.0", it would be helpful if the documentation could explain this in more detail.

Versions: * Go go version go1.9rc2 darwin/amd64 * dep v0.3.0-142-g6ca8a48

Updated 18/08/2017 21:21 2 Comments

[DOCS] Bulk API copy as curl link defines json content type but documentation says it must be ndjson


Found in and later versions.

If you click on “Copy as curl” on the POST _bulk example, the following is copied to clipboard. This specifies the wrong content type; according to the documentation it should be Content-Type: application/x-ndjson.

curl -XPOST 'localhost:9200/_bulk?pretty' -H 'Content-Type: application/json' -d'
{ "index" : { "_index" : "test", "_type" : "type1", "_id" : "1" } }
{ "field1" : "value1" }
{ "delete" : { "_index" : "test", "_type" : "type1", "_id" : "2" } }
{ "create" : { "_index" : "test", "_type" : "type1", "_id" : "3" } }
{ "field1" : "value3" }
{ "update" : {"_id" : "1", "_type" : "type1", "_index" : "test"} }
{ "doc" : {"field2" : "value2"} }
Updated 18/08/2017 13:57 1 Comments

Chrome doesn't work with Linux subsystem for Windows


I’m trying to test this under the Windows Subsystem for Linux and get an immediate error:

~$ lighthouse --output-path /mnt/c/Users/ray/Desktop/lh.html
Runtime error encountered: Error: The environment variable CHROME_PATH must be set to executable of a build of Chromium version 54.0 or later.
    at Object.linux (/home/ray/.nvm/versions/node/v8.2.0/lib/node_modules/lighthouse/chrome-launcher/chrome-finder.js:107:15)
    at Launcher.<anonymous> (/home/ray/.nvm/versions/node/v8.2.0/lib/node_modules/lighthouse/chrome-launcher/chrome-launcher.js:113:75)
    at (<anonymous>)
    at /home/ray/.nvm/versions/node/v8.2.0/lib/node_modules/lighthouse/chrome-launcher/chrome-launcher.js:12:71
    at Promise (<anonymous>)
    at __awaiter (/home/ray/.nvm/versions/node/v8.2.0/lib/node_modules/lighthouse/chrome-launcher/chrome-launcher.js:8:12)
    at Launcher.launch (/home/ray/.nvm/versions/node/v8.2.0/lib/node_modules/lighthouse/chrome-launcher/chrome-launcher.js:98:16)
    at Object.<anonymous> (/home/ray/.nvm/versions/node/v8.2.0/lib/node_modules/lighthouse/chrome-launcher/chrome-launcher.js:42:24)
    at (<anonymous>)
    at /home/ray/.nvm/versions/node/v8.2.0/lib/node_modules/lighthouse/chrome-launcher/chrome-launcher.js:12:71

Any ideas?

Updated 18/08/2017 16:48 4 Comments

Documentation on custom lemmatization


Hi, I am trying to implement Russian support for spaCy. Russian is a highly inflectional language, which makes lemmatization a very important part in language processing pipeline. I think that before training any Brown clusters or word vectors I need to add a decent lemmatizer in the Tokenizer. I discovered the base class Lemmatizer in spacy/, but I am not sure what part of its interface I need to reimplement. There are also some hints that lemmatizer should also do POS tagging, is it true?

Updated 18/08/2017 12:55

dnsdist: Document the Truncate option to `setDynBlocksAction()`


Short description

<!– Write a small description of what this Pull Request fixes or provides, including the issue #s –>


<!– please indicate if any of these things are done/included with this Pull Request. Not all boxes need to be checked for the Pull Request to be accepted –> I have: - [x] read the document - [ ] compiled and tested this code - [x] included documentation (including possible behaviour changes) - [ ] documented the code - [ ] added or modified regression test(s) - [ ] added or modified unit test(s)

Updated 18/08/2017 10:46

Add docs precompiled libraries in challenge evaluation scripts


Some of the challenge evaluation scripts use precompiled libraries such as MSCOCO ( We want to add a section in the documentation where the challenge hosts should be guided to compile those libraries against Ubuntu 14.04 ONLY since all the EvalAI servers use Ubuntu 14.04 so it might not work with libraries that are compiled against OSx or Windows.

Updated 18/08/2017 06:03

Pre-cache key assets with Workboxjs.


First time attempting to contribute, so bare with me please if i’ve missed anything or done anything wrong.

But I noticed this was merged ( #23101 ) so in addition, thought it would make sense to pre-cache those main assets.

Also not too sure how it works but I think Jekyll is serving/parsing the markdown doc files dynamically. It would be a cool idea if we could maybe get them to static html files? That way we can have offline support with this too!

Screenshot to show speed difference: - throttled slow 3G - throttled slow 3G w/ workboxjs

Also ran grunt test and everything passed fine.

Updated 20/08/2017 06:48 8 Comments

obtautomer fails to detect basic keto-enol tautomerism



The command line tool obtautomer fails to convert enols into their keto forms (probably the classic example for tautomerism), e.g. for the following structures (with and without the -c flag):

C=C(O)C – expect CC(=O)C CC(O)=CC(=O)C – expect CC(=O)CC(=O)C C1CCCC=C1O – expect C1CCCCC1=O

Thanks, Christian

Environment Information

Open Babel version: 2.4.1 Operating system and version: Fedora 26

Updated 18/08/2017 17:09 3 Comments

Change the message on Challenge Creation


Currently, the error message on missing test annotation file for a phase shows the following error: <img width=“1440” alt=“screen shot 2017-08-17 at 4 55 26 pm” src=“”>

We want to change it to

No test annotation file is present in the zip file for challenge phase 'test-dev2017'. Please try uploading again after adding test annotation file!
Updated 17/08/2017 20:58

Add note about JSON/JSONB serialization changes


Tried to backport myself, but ran into the error below. For some reason, didn’t run into any issues with 5-1-stable

Jonathans-MacBook-Pro:activerecord jon$ git cherry-pick 835246e622dc0431af3cb951db22ef78876006af
[5-0-stable 497d1f8501] Add note about JSON/JSONB serialization changes
 Date: Thu Aug 17 13:43:43 2017 -0400
 1 file changed, 9 insertions(+)
Jonathans-MacBook-Pro:activerecord jon$ git push dnt 5-0-stable
Counting objects: 5, done.
Delta compression using up to 4 threads.
Compressing objects: 100% (5/5), done.
Writing objects: 100% (5/5), 1.40 KiB | 1.40 MiB/s, done.
Total 5 (delta 4), reused 0 (delta 0)
remote: Resolving deltas: 100% (4/4), completed with 4 local objects.
remote: error: GH006: Protected branch update failed for refs/heads/5-0-stable.
remote: error: You're not authorized to push to this branch. Visit for more information.
 ! [remote rejected]       5-0-stable -> 5-0-stable (protected branch hook declined)
error: failed to push some refs to ''
Updated 17/08/2017 20:26 1 Comments

How to get rid of automatic subscription of $client/{clientID} after connection?



  • OS: CentOS7
  • Erlang/OTP:
  • EMQ: 2.3


After each client connects, web-hook logs shows that the client automatically subscribe a topic - ‘$client/{clientID}’ as below.

Body: {"action":"client_connected","client_id":"aaabbbcccc","conn_ack":0} Body: {"action":"client_subscribe","client_id":"aaabbbcccc","username":"111","topic":"$client/aaabbbcccc","opts":{"qos":1}} Could you guide me how to remove the automatic subscription behavior unless it is mandatory feature for any system monitoring reasons?


Updated 18/08/2017 12:51 2 Comments

Removal of Python 2 should be more documented, with a workaround



The removal of Python 2 in jupyter/scipy-notebook provoked lots of errors and astonishment in my scripts. I speak of the following pull requests : and .

As I probably won’t be alone to have this problem, perhaps you could precise in the a workaround (or your maybe planned derived image ?). For instance, I use jupyter/scipy-notebook:b4dd11e16ae4 instead.

Thank you for all the good work.

Updated 18/08/2017 14:39 1 Comments

Document how to support Serde as an optional library feature


This is mentioned under feature flags and in the source.

It would be nice to have an example or at least mention under Using derive.

As I understand it, doing so involves this:

serde = { version = "1.0", optional = true, features = ["derive"] }

and this:

extern crate serde;

and this:

#[cfg_attr(feature="serde", derive(Serialize, Deserialize))]
pub struct MyStruct { ...
Updated 18/08/2017 08:14 5 Comments

Docs Link Checker Does Not Fail Build When Detecting Broken Links


We run a link checker against the documentation within the build to make sure there are no broken links introduced. If broken links are detected, a notification is printed in the logs and left for someone to manually check.

A detected broken link should automatically fail the build.


Visually checking a log for broken links is not practical.

Expected Behavior

Broken link = broken build

Actual Behavior

Broken link != broken build

Possible Fix

Either make the link checker force a fail, or have a process check the logs to search for the broken links text, then fail as appropriate.

Steps to Reproduce

1) Break every link in the docs, or just one, it depends on how many links you like to destroy and how much of a point you want to prove. 2) Deliver the broken link docs 3) Watch the build pass 4) Watch the manual checking process get forgotten and all the broken links get rolled out into the release.


All I wanted to do was follow a link. A single link.

Updated 17/08/2017 11:49

Slider doc needs more dcumentation



I think it’s not enough to say that the handle of a slider can be styled: styled how? Can I pass an SVG image to use as my handle? Can I only change the colour? If so, is it through color or background-color?

Updated 17/08/2017 14:01 4 Comments

Absolute basics introduction


Whilst the documentation on your site is good, I’m finding it makes lots of assumptions that limits who will be able to use Babel. It would be useful for me if you could have an introduction page (even if it’s largely links to other sites) explaining how Babel ties in with node (which I’ve never used to develop a project) and npm (which I thought was just a package manager but now seems to be some kind of build manager like ant or maven), where transpiling is done (I got the impression from someone that it’s done in the browser but I can’t confirm this from your site) and things like that. I’m sure once you’ve started using Babel, it’s really simple, but for me there’s currently a prohibitively steep learning curve because I haven’t got time to search around for an overview of the basics. (My background is in more “traditional” web development, so I know how to use JavaScript, CSS, etc. on a traditional webserver, so an overview of the differences would be particularly useful to developers like me.)

Updated 17/08/2017 08:01

Document kafka plugins' default codec change as breaking in breaking changes documentation


I am filing this under the main LS repo because we don’t have breaking changes section per plugin, but at the base Logstash level.

We changed the default codec for the kafka input from json to plain here. Btw, I think we meant to say “Changed default codec to plain instead of json” (and not “Changed default codec to plain instead of SSL” :) )

This can cause downstream pipelines to break because if users who expect json handling continue to use their existing pipeline, the events will end up in a plain large text body message field instead of individually parsed json fields - unless they know to explicitly set codec to json now.

Users tend to review breaking changes from the breaking changes section in the Logstash guide so it will be helpful to document it there (plus we already have other plugin breaking changes there as well).

Updated 18/08/2017 15:47 5 Comments

Eris doesn't automatically use Bluebird.


“install them in the same directory as Eris, and Eris will automatically use them” - the docs

Yet when I search for “bluebird” in this repository I get no results. So are the docs misleading or did someone forget to use bluebird? the docs the search results

Updated 17/08/2017 04:19 1 Comments

Clarify doc comments for special types for route-related 'values' parameters


In the various URL helper extension methods there is a values parameter that can be a RouteValueDictionary, IDictionary<string, object>, or a POCO (object with properties), and it does the right conversion based on those. The doc comments of these overloads lack information on the two special types, so we should add details on those.

@javiercn - can you propose new text for this type of parameter, and then make the change everywhere? Should check the Routing repo in case there is a similar need there.

Updated 17/08/2017 18:34 5 Comments

[Docs] Make docs WebP First


This is based as per the discussion with @kball

For regular images ( img tag ) … Simple onerror Event

<img src="image.webp" onerror="this.onerror=null; this.src='image.png'">

For Background images, we can trigger as discussed in this article

const body = document.querySelector('body');

async function supportsWebp() {
  if (!self.createImageBitmap) return false;
  const webpData = 'data:image/webp;base64,UklGRh4AAABXRUJQVlA4TBEAAAAvAAAAAAfQ//73v/+BiOh/AAA=';
  const blob = await fetch(webpData).then(r => r.blob());
  return createImageBitmap(blob).then(() => true, () => false);

(async () => {
  if(await supportsWebp()) {
     body.classname = 'webp';
  else {
     body.classname = 'no-webp';

and then we can just

.no-webp .element {
  background-image: url("image.jpg");

.webp .element {
  background-image: url("image.webp");
Updated 17/08/2017 11:14 8 Comments

[Docs cleanup] Remove link text from inline links where appropriate


Historically we’ve included link text in situations where inline links would result in the doc build adding “the section called” to the link text.


To resolve this issue, we’ve coded the links like this: <<link-id,link text>>

With this commit, the build no longer adds “the section called” to the link text. Link text is only required if you want to provide link text that is different from the topic title.

Manually coded link text takes more effort to maintain over time. Therefore, where it makes sense, we should remove the link text and allow the build to resolve the text. There are some situations where we will want to the link text to be different from the published title, but we should clean up any links that aren’t in this category.

Updated 16/08/2017 16:51

How to use dep according to best practice go folder structure


This is not a bug report.

I wonder how to use dep when your repo have a multiple cmd project. For example of such structure: + cmd + ls ls.go - import golang 3rd party flag parsing + httpd httpd.go - import fasthttp package posix.go In this case the ls may import go package for flag parsing, and the httpd may use fasthttp package. Both are importing the root package to use the common posix file.

Where i do the dep init? On each cmd and on root project? Or something else? Please advice.

Updated 16/08/2017 16:09 1 Comments

Proc::Async Example needs to be modified to run on Windows 10


On page:

The first example: ```# command with arguments my $proc =‘echo’, ‘foo’, ‘bar’);

subscribe to new output from out and err handles:

$proc.stdout.tap(-> $v { print “Output: $v” }, quit => { say ‘caught exception ’ ~ .^name }); $proc.stderr.tap(-> $v { print “Error: $v” });

say “Starting…”; my $promise = $proc.start;

wait for the external program to terminate

await $promise; say “Done.”; ```

Will not run properly on Windows. Apparently this is because in Windows, echo is a function of the executable cmd.exe.

So, to make this work on Windows, one has to change the line: my $proc ='echo', 'foo', 'bar'); to: my $proc ='cmd.exe', '/c', 'echo foo bar'); Then it will run properly.

It would be great if the example could be updated to mention this

Updated 19/08/2017 03:03 5 Comments

addRoutes() adds unusable router and extend() will loose all prefixes


Adding a let r2 = new Router({prefix: 'a'}).loadMethods().get('/', (ctx, next) => {ctx.body='supposed to be at /api/a/'; return next();}) with let r1 = new Router({prefix: 'api'}).addRoutes(r2); does not result in a callable route. Actually koa-better-router crashes with not having any routes.

And using extend -> let r1 = new Router({prefix: 'api'}).extend(r2); will loose all prefix info from r2. Putting all ‘/’ routes at the same base level.

Any known workaround?

Updated 18/08/2017 00:07 13 Comments

Docs needed: HTTP bridge HTTP basic auth


The HTTP now supports a form of HTTP/Basic authentication where the token the client presents is directly used and plugged into WAMP ticket authentication.

The feature landed via, and should be documented shortly at least. Probably also with a working example in crossbar-examples.

Updated 15/08/2017 20:14

Update guide for removal of lints


8183f636305cef4adaa9525506c33cbea72d1745 removed lints from rocket_codegen due to stability concerns. Lints will be made available in a new rocket_lints crate. Unfortunately, the guide makes no mention of this. This should be fixed.

Updated 15/08/2017 19:03

Update Browser & Device Support docs


Select one … (check one with “x”) [ ] bug [ ] feature request [x] enhancement

Expected behavior

Clearer documentation on what browsers we support.

Questions to answer

  • How do we define a version? Should this be release or something else?
  • We show that we support mobile devices
    • Does this mean that we also support all of the listed browsers that are available on mobile devices?
    • Is this all mobile devices?

Actual behavior

As of now, we say

Clarity only supports the latest two versions of the listed browsers, with the exception of Internet Explorer.

We are getting questions/issues on support for Safari 9.1 and Mobile Safari.

Reproduction of behavior


Environment details

  • Clarity version: 0.10.0

  • Browser: all

Updated 16/08/2017 13:45

Fork me on GitHub