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

Possible Viewer 2.0 Improvements

  • Reduce asset sizes:

    • [ ] Minify or dedupe report renderer js, css, and templates (included twice, for report rendering but also in a browserified bundle of ReportGenerator for html saving). Minifying just the JS before inlining would save about 75KB in viewer and lighthouse-background.

      This could either use the assets already in the page (but we’d need a non-awkward/maintainable way to do so) so they only appear once, or we could replace brfs with a transform that minifies the JS/CSS/HTML before inlining it as a string (this approach would also fix #2014)

    • [ ] Resize lh_logo_bg.png closer to the size it appears on screen (formerly ~#1569~)
    • [ ] Minify in-page CSS (formerly ~#1552~)
  • [ ] Experiment with versioned viewer/report scripts so old JSON results can still be loaded in viewer. May be less of a priority since we want to maintain some level of backwards compatibility so that DevTools can load old JSON reports.

    Could do this with a script that checks the version and redirects to a specific URL per LH version (easy), or could split viewer and report into separate bundles and load the right report bundle based on json version (harder).

  • [ ] PWAify Viewer

  • [ ] Switch Viewer to use event-based ga and logging, like report does.

  • [ ] Explore CSP options (for extension and HTML report, as well)

  • [ ] Verify that report is actually isolated to its container so that when the container is cleared the report is actually cleared. It’s likely that event listeners, references, etc remain every time a new report is loaded.

  • [ ] Viewer could easily be added to the report compile step since they’re designed to be loaded in the page at the same time. Would need to write idbKeyval and firebase externs.

Updated 24/06/2017 01:14

"its associated CSS layout box"


A video element is said to intersect the viewport when it is being rendered and its associated CSS layout box intersects the viewport.

IIUC, an element has multiple CSS boxes (content box, padding box, border box, margin box), so “[its] CSS layout box” seems ambiguous. Presumably the sentence should refer to one of those boxes specifically?

Updated 26/06/2017 19:34 1 Comments

Blocked colors in shield panel are different than laptop

  • Did you search for similar issues before submitting this one? Yes
  • Description: The colors on the blocked count in the site shield settings panel is wrong on iOS vs browser-laptop Browser-laptop: image browser-ios: image uploaded from ios

  • Device (iPhone5, iPhone6s plus, iPad 3, ?): Does not matter

  • Brave Version: 1.3.3 and 1.4
  • Steps to reproduce:
    1. Navigate to any site on iOS
    2. Open the shield panel

    3. Navigate to any site on browser-laptop

    4. Open the shield panel
  • Actual Behavior The colors on the blocked count is wrongs vs browser-laptop
  • Expected Behavior The colors should be the same
  • Is this an issue with Beta build? Yes
  • Is this an issue in the currently released version? Yes
  • Can this issue be consistently reproduced? Yeah

  • Website problems only:

  • did you check with Brave Shields down?
  • did you check in Puffin/Dolphin/iCab/Firefox (UIWebView-based browsers)?

  • Screenshot if needed:

  • Any related issues:

Updated 22/06/2017 18:27 1 Comments

Primary color on a tag in panel title


Test Case



There is a problem with panels title. I used a hyperlink in the colored panel title (Info, Success, Danger, Warning, Primary) but it has Primary color on hover/focus that makes it unpleasant and hard to read.

Expected result

Have white or gray color in hover/focus mode.

Updated 26/06/2017 10:00

Regressions with and error reporting


Our search should be finding files in non-standard paths, but this is not happening automatically. We should be able to find and use a at least in any first level path.

Additionally, when we aren’t finding a, we were reporting this to the user. This is not happening either now, instead the user receives a message that an unknown error occured.

Ref #2962

Updated 21/06/2017 16:25

Add a loader during pre-commit hook process



Could you add some loader for pre-commit hooks? When I commit my changes I have a hook that takes a time to be done, and there’s no way in current interface to understand that something is happening now. Of course, after the hook is done the commit text and staged files are gone, and it’s clear that I can commit, but in my opinion loader will be very clear and user-friendly feature)

P.S. This feature is awesome, thank you

Updated 20/06/2017 17:16 1 Comments

[YouTube] Add support for links

  • Firefox 53.0.3 and up
  • Min-Vid 0.3.8-dev (built on 6/14/2017)

  • All Windows

  • All Mac
  • All Linux

  • Latest Min-Vid add-on was previously installed on a clean profile.

  • Open this testcase.

  • Right click on the 1st link and observe the context menu.
  • Right click on the 2nd link and observe the context menu.

Step 2 & Step 3 - The Min Vid context menu option is available.

Step 2 - The Min Vid context menu option is available. Step 3 - The Min Vid context menu option is not available.

  • Both links are for the same video, the only difference is that one is the long link and the other is a short link.
  • Attached a screenshot of the issue: no context menu
Updated 16/06/2017 19:29 1 Comments

[YouTube] Playlists from the Related section do not have an overlay option

  • Firefox 53.0.3 and up
  • Min-Vid 0.3.8-dev (built on 6/14/2017)

  • All Windows

  • All Mac
  • All Linux

  • Latest Min-Vid add-on was previously installed on a clean profile.

  • Open the browser with the profile from prerequisites and navigate to a

  • Click on a video.
  • Search the related videos section for a playlist thumbnail and hover it.
  • Observe the playlist thumbnail.

  • The Min Vid overlay appears.

  • The Min Vid does not appear.

  • Both context menu options exist and work as expected.

  • Attached a screen recording of the issue: overlay on related playlists
Updated 16/06/2017 19:28 1 Comments

Add "checkout" command to checkout new and existing branches


Currently there are no registered commands for checking out new and existing branches. Although the status bar tile is a pretty easy way to check out different branches, it would be nice to perform the same actions in Atom without leaving the keyboard.

I noticed there are already callback functions for checking out and creating new branches in the branch menu view, so it seems possible to register commands that use those existing callback functions (or something similar).

Updated 20/06/2017 17:17 2 Comments

False positive blocking scripts


Scripts loaded after the main load event are considered as blocking if they are prefetched in the head.

It is exemplified in this gist, where in the document’s head a html <link as="script" src="/main.js" rel="preload"> tag is present and at the end of the body the preloaded script is added by javascript document.createElement('script') script.src = '/main.js' document.head.appendChild(script)

(Opened after

Updated 15/06/2017 23:09 5 Comments

format the entire codebase


we have a little problem with inconsistent formatting through the codebase. E.g. spaces are mixed with tabs, sometimes new code accidentally used void functionname { instead of void functionname {

a clang format file should be created to be used for automatic formatting of the codebase. Sadly, has very little information as a base for that as yet.

So anyone doing this, would need to make the decision for each bullet point in that config file by reading through some old code (best is to base the decisions mostly on the legacy codebase in e.g. inexor/fpsgame or inexor/engine).

But maybe thats a good way to get in first touch with the code?

Updated 16/06/2017 08:57 2 Comments

[GITHUB] add commit message validation for PRs


Each time we have a Pull Request (and everything which goes into master has to go through PRs) we should only give the final “Okay merge-able” if the commit messages are all okay.

I think such policies are best enforced by machines.

We could add a service somewhere on a server, which would listen to the GitHub webhook API and execute some validation script (there are enough out there, e.g. validate-commit-msg or fit-commit). Could be done similar to does stuff, but it would be cool if we really enforce this (instead of just comment-spamming). I.e. we enforce currently that the Appveyor and Travis hooks run through before a Pull Request is allowed to be merged (thanks to GitHubs protected branches. So we could use the Github Status API

Currently the layout is: ``` [NETWORK] allow only remote clients

This is a longer description why we no longer allow local clients. ``` which collapses to

[NETWORK] allow only remote clients in the history.

In respect to #37 we should maybe add info about the severity/type of that change (e.g. “fix” “refactor”, ..) so we can generate the next version based on the commit messages.

Updated 16/06/2017 08:52 1 Comments

In Copy -> Function Name context menu item for C++ frames, don't include function arguments


Right now we copy the entire line, for example mozilla::cyclecollector::HoldJSObjectsImpl(void*, nsScriptObjectTracer*), which makes it impossible to look up what I copy on searchfox without trimming the argument list manually. Given that the name of the menu item suggests that it is supposed to copy the function name, I think it’s reasonable to expect mozilla::cyclecollector::HoldJSObjectsImpl to be copied to the clipboard instead.

Updated 19/06/2017 17:33 3 Comments

Feature Request: option to print deposit comments on deposit summary


Printing deposit summary

Collected Value Title Data
Page Name /SystemSettings.php
Screen Size 900x1440
Window Size 799x926
Page Size 1100x1156
Platform Information Linux oin 3.2.61-grsec-modsign #1 SMP Tue Aug 12 09:58:26 UTC 2014 x86_64
PHP Version 5.6.29
ChurchCRM Version 2.6.3
Reporting Browser Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.36
Prerequisite Status Missing Prerequisites: [“PHP 7.0+”,“PHP Phar”,“Mod Rewrite”,“FileInfo Extension for image manipulation”]
Integrity check status
Updated 14/06/2017 00:19

Add QR code to Publish Dialog.


A QR code might be a great way to view a project on a mobile device right after sharing. As it stands, it’s somewhat cumbersome to type out the URL of a published project on a mobile device since the URL is quite long. In addition, when using Thimble during a demo session like this weekend, a user can’t just go back to their projects list to continue the work we started.

One helpful participant at the Faire this weekend (Jim) took the time to compile some ways that we could do this and published the information as a thimble project:


Cheers, Jim!

Apparently iOS will be adding in native QR support in the near future, so it’s a capability that a lot more people will have access to soon.

Updated 15/06/2017 23:30 4 Comments

Refactor keyboard selection code to be unit-testable


The skipWrapper and getNewSelection functions have been extremely prone to mistakes. See and

This is partly because they need to account for nested Wrapper components (such as the case with arrays in React Fiber).

We need to figure out a way to make them unit-testable, and actually write a regression suite ensuring keyboard navigation works in all directions from different points of the tree. I’m not sure how to approach it—you’d need to figure out the best way to do it.

Updated 09/06/2017 19:31

[query] Add type-hinting functions


It would be convenient to have a way to constrain the types of variables explicitly:

[:find ?x :where
 [(uuid ?v)]
 [?x _ ?v]]

This hinting/constraining would be useful for queries like the above — find all entities associated with any UUID — but also allows for ground (and potentially other functions) to accept ambiguous EDN input. We don’t have a good way right now to distinguish between 10 the Long and 10 the entid.

These functions are very straightforward to implement.

Updated 16/06/2017 19:17 1 Comments

generate html API docs for artifacts, driver, etc


generated API docs could be very helpful, preventing the need to dive into multiple gatherers just to see what they provide in their artifacts, for instance. Ideally these would be generated from source with a single npm command.

  • It may not be necessary (or even helpful) to have html docs for everything, but docs for key pieces would aid those developing Lighthouse core and/or their own extensions of it. Driver, the artifacts returned by all the core gatherers, and maybe things like the Audit and Gatherer base classes would be a great first start. OTOH, things like GatherRunner may not be helpful to many (but it does have that sweet execution outline)

  • One issue is that the existing jsdoc declarations are a little spread out and not always complete. Driver is more or less documented in place, but the artifacts returned by gatherers often aren’t, and instead you’d have to look to the closure externs for those types. Hopefully efforts to document all of this will help suggest how to better document the different parts of our architecture in a way that’s maintainable and stays current.

  • It’s not clear how the docs should be laid out. It probably makes sense to do more of a logical grouping than one based on file location…e.g. you don’t want to have to navigate to 20 different pages to see all the different artifact types, but Driver should probably be a page on its own. Again we want something maintainable, but also something that humans will be able to easily browse.

Updated 07/06/2017 22:24 2 Comments

Add UI to indicate file import/unzip is happening


When you drag-and-drop files into Thimble, it can take some time to process them, especially if they are big, or are in an archive that has to be inflated. We should show some kind of UI that lets the user know that things are happening.

In Brackets we do the async file import operation here:

And when we call the callback, it’s done. We could alter the filetree somehow here and add some kind of spinner or other animation.

cc @flukeout

Updated 07/06/2017 02:01

Can't edit fontWeight in RN Style Editor


Here’s the current value:

<img width=“183” alt=“screen shot 2017-06-06 at 3 59 17 pm” src=“”>

If I try to set it to 200 I get this: <img width=“308” alt=“screen shot 2017-06-06 at 3 59 22 pm” src=“”>

If I try to set it to string '200' I get this instead: <img width=“313” alt=“screen shot 2017-06-06 at 3 59 35 pm” src=“”>

How do we fix it?

Updated 06/06/2017 15:00

Change position of "Download" action in the right-click menu.


I can’t tell you how many times I’ve accidentally downloaded a file instead of deleting it. They both start with D and Delete used to be the right option (and will probably get used more often)…

<img src=“” width=“250”/>

Let’s move Download above Delete

Updated 02/06/2017 19:50 1 Comments

Figure out what exactly we should be downloading from NCBI in terms of bacterial and archaeal genomes


We want all published genomes of bacteria and archaea, but unfortunately NCBI’s standards and file system are not easy to understand. We’ve been skirting this by just using the data from a different CRISPR database, but eventually we’d like to not rely on them and just download directly from NCBI so that we can update whenever we want.

If we could get a list of accession numbers for all the available bacteria and archaea, that would be great. Do we want all the NC_ accessions? What about NZ_? What’s the difference? Is there overlap?

Anyone who knows about this and/or feels like doing some reading up on NCBI is very welcome to help us here!

Feeds into #196.

Updated 21/06/2017 21:35 9 Comments

When the API gives us an NTEE code, convert it to human language


When we ask the Propublica API for an ntee_code, it returns things like D45, N14, and H56, which are not very interesting on their own. HOWEVER, those letters and numbers correspond to specific categories like O21: Boys Clubs, K: Food, Agriculture, and Nutrition, and D50: Zoo, Zoological Society. It would be much more useful and readable if we translated something like I44 to one cell “Crime, Legal Related” and another cell “Prison Alternatives.”

I’ve attached a document published by the IRS that gives the key to all these codes, but maybe there’s a structured dataset somewhere?

This is for you, dear open source contributor, to find out.


Updated 01/06/2017 22:30

Property Inspector is Not Resilient to Non-existent Property


val can be null here in _inspectResponse. Which throws and causes havoc. Reproed in Outlook 365.

What happens here is that the property keys enumerate in the property view as if they’re own properties and not on the prototype. But when I try to expand one of the properties. It is now on the prototype (not own) and therefore getIn returns null.

I’m not quite sure how that’s happening. Perhaps there’s a getter in the user code. However, since getIn is explicitly invoking getters it needs to be more resilient to the assumptions changing.

Updated 02/06/2017 17:24 5 Comments

Review/Edit Section 0.0 and 0.1 of Open Leadership Training (README)


Review, comment, edit google doc for Section 0.0 and 0.1 of open leadership training. Please use “suggesting” function google doc so we can track changes

This issue is a “sub-issue” of #5

Updated 02/06/2017 19:35

Telemetry should use pylintrc of catapult


catapult’s pylintrc is more restricted than Telemetry’s one. This is because there are certain pylint rule that are broken in Telemetry code base. If we can get rid of those, we can just remove catapult/telemetry/pylintrc file.

To see which lint rules are currently broken: 1) Remove catapult/telemetry/pylintrc 2) Make any no-op change to telemetry/telemetry/ 3) Commit the change 4) Do a git cl upload & see all the error message.

Updated 12/06/2017 13:51 1 Comments

Should “Naming form controls” section include a warning about names overriding form properties?



    <input name="tagName">
    <input name="nodeType">
    <input name="parentElement">
    <input name="children">
    <input name="onsubmit">
    <input name="clientWidth">
    <input name="noValidate">
    <input name="method">
    <input name="action">

This will effectively break all the corresponding form properties, e.g. form.tagName will refer to the input, instead of returning "FORM". This can cause bugs, so I’m wondering if the spec should actively warn against this.

Maybe it would make sense to add a warning to the “Naming form controls” section, to make sure that authors don’t create form controls with names that would collide with built-in form properties.

Updated 30/05/2017 20:42 2 Comments

Replace unsafe formatting lib


Currently there are a lot of “defformatstring”, “formatstring” and “tempformatstring” used through the codebase to simplify printf-formatting of a string and some variables into a string (e.g. defformatstr(name, "user-number-%d", id_number);.)

This is sometimes slow, but mostly its terrible unsafe in comparison to the fmt lib (aka cppformat). And since we switch to fmt for new code, we should convert the old formatting code as well. It would be stupid to maintain our own formatting library if there already is a well-established alternative and the goal is to drop it as soon as possible.

Tasks: - [ ] add fmt/3.0.0@memsharded/testing to the * (to tell conan to download this dependency prior to building inexor) - [ ] add add_require_conan_lib_function(fmt) to the others in cmake/require_thirdparty_libs.cmake
* thats a macro to create a function called require_fmt(<othertarget>) * for cmake. so cmake knows about - [ ] add require_fmt(............) (twice, you’ll see) to any module-folders (inexor/..) CMakeLists.txt that is using this lib. * every module needs to require the dependency to its project. - [ ] Find all old-style formatstring and replace them directly with fmt * include the fmt-header in the specific file * convert the API from printf-style to the fmt style * consider stating using namespace fmt; if you are using it a lot in a cpp file.

Updated 21/06/2017 08:21

Feature Request: Do not list files which are not readable by nginx


I made a log folder of the system accessible from a page which use fancyindex. On the standard configuration fancyindex will list every files in the folder (since nginx has execution right on the folder). Now from the web some files are unreadable. I would like to be able to ignore all unreadable files without entering manually each of them.

As a side note I discovered that when fancyindex_ignore forw with strings (nginx not compiled with PCRE), the function ignore anything that start with the string. Is it a bug? Can you add this info in the doc if it’s intended? For now it help me a lot with logs who have a timestamp in their filename.

Updated 16/06/2017 22:01

Consider reloading the preview when files are uploaded to project.


This is a bit of a fringe scenario, but let’s say I’m uploading an image file that is already referenced in the markup via an img tag or maybe as a background-image in CSS. When the upload completes my preview doesn’t refresh to show me the image. I have to manually refresh it, or make other changes that cause it to refresh.

This may make the editor feel broken because most other changes happen instantly.

Updated 30/05/2017 15:31 1 Comments

Scroll position in the Play Queue or History sections is not reset after certain actions are made

  • Firefox 53.0.2 and up
  • Min-Vid v. 0.3.7 (custom repo build)

  • All Windows

  • All Linux
  • All Mac

  • The “xpinstall.signatures.required” is set to “false” and latest Min Vid add-on is installed on a new clean profile.

  • Play Queue and History have enough items (~20 items each) so that the scroll bar appears.

  • Open the browser with the profile from prerequisites and navigate to

  • Send a video to the Player click the “Open Queue” button.
  • Scroll down to the bottom of the list and click the “History” button.
  • Close and reopen the Play Queue.
  • Navigate to the History section and then back to the Play Queue.
  • Observe the scroll position.

  • Scroll position is at the top of the list.

  • Scroll position is at the bottom of the list.

  • The issue is also reproducible in the History section. (scroll down History, click Play Queue, close and reopen the Queue, click on History and observe the scroll position)

  • Attached a screen recording of the issue: odd behavior with scroll position after some action are made
Updated 25/05/2017 09:35 2 Comments

OpenAPI v3: code template for info#termsOfService should be `"termsUrl"`


From Info:

termsOfService | string | A URL to the Terms of Service for the API. MUST be in the format of a URL.

Variable name of info#termsOfService is just terms (termsOfService: terms). This is just a hint, the users are supposed to provide their own values. Change it to termsUrl to promote best practices.

Using terms does not create any problems with our validator.

Updated 24/05/2017 00:10

What makes a critical request critical?


Consider this more of a question or discussion rather than an issue. In some cases, I find this text in the Critical Request chain section to be somewhat misleading:

The Critical Request Chains below show you what resources are required for first render of this page.

Imagine a page with a CSS reference in the head, and all script references just before the end body tag. Lighthouse will report both the CSS reference and all scripts to be part of the critical request chain. Even if none of those scripts are required for actual rendering.

So technically, those scripts are not required for a first render of the page. They may be required, however, to finish DOM construction, but does that really block rendering of the first paint? It was my understanding that scripts put at the end of the body tag do not block progressive rendering of a page?

Another instance where I was confused is typekit’s font service. By default their font loading is async, yet every request to their network is considered render-blocking by Lighthouse. Which seems incorrect, since the whole point of async font loading is to be non render-blocking?

I’m just trying to solidify my understanding of what is considered “critical” or render-blocking, because right now it seems that resources that do not block a first paint, are still marked as such.

Updated 26/06/2017 08:32 9 Comments

Lighthouse 2.0.0 error

C:\Users\xmr\Desktop>node -v && npm -v

C:\Users\xmr\Desktop>lighthouse --view
  ChromeLauncher Waiting for browser. +0ms
  ChromeLauncher Waiting for browser... +0ms
  ChromeLauncher Waiting for browser...√ +484ms
  status Initializing… +1s
  method <= browser ERR:error Network.setBlockedURLs +407ms
  status Loading page & waiting for onload +0ms URL, Viewport, ViewportDimensions, ThemeColor, Manifest, ChromeConsoleMessages, ImageUsage, Accessibility, EventListeners, AnchorsWithNoRelNoopener, AppCacheManifest, DOMStats, OptimizedImages, ResponseCompression, TagsBlockingFirstPaint, WebSQL
  statusEnd Loading page & waiting for onload +9s
  status Retrieving trace +0ms
  status Retrieving devtoolsLog and network records +201ms
  status Retrieving: URL +16ms
  status Retrieving: Viewport +0ms
  status Retrieving: ViewportDimensions +0ms
  status Retrieving: ThemeColor +0ms
  status Retrieving: Manifest +0ms
  status Retrieving: ChromeConsoleMessages +437ms
  status Retrieving: ImageUsage +0ms
  status Retrieving: Accessibility +0ms
  status Retrieving: EventListeners +129ms
  status Retrieving: AnchorsWithNoRelNoopener +64ms
  status Retrieving: AppCacheManifest +0ms
  status Retrieving: DOMStats +0ms
  status Retrieving: OptimizedImages +0ms
  status Retrieving: ResponseCompression +47ms
  status Retrieving: TagsBlockingFirstPaint +0ms
  status Retrieving: WebSQL +0ms
  method <= browser ERR:error Network.setBlockedURLs +842ms
  status Loading page & waiting for onload +0ms ServiceWorker, Offline, StartUrl
  statusEnd Loading page & waiting for onload +67ms

Using Chrome 58 64-bit on Windows 7 64-bit.

Updated 23/05/2017 20:08 2 Comments

Delete project


We need to provide a red “delete project” button in the setting page. Similar to what Github, Dockerhub etc. have (require writing the name into a box).

We need to delete all references to the project apparently and also in future when activity tracking is merged all references must be deleted as per out policies.

Updated 18/05/2017 06:17

Add option to skip component


During development, it is often helpful to skip parts of a study. If a study is written in code, the skip option can be used to achieve this: It terminates a component immediately after it is run, effectively skipping it.

This option, however, is not yet represented in the builder interface. To complete this issue, users should be able to skip a component from the UI. The easiest way of achieving this would probably be to add a checkbox in the ‘advanced options’ section, but you might find it more intuitive to add an option to the drop-down menu visible for the selected component in the sidebar.

As a bonus, you might think about representing the skipped state in the sidebar. That’s a stretch goal, however, and in no way necessary to complete this issue.

Updated 14/05/2017 17:13

Clarify / add hint to responses UI


In the lab.js builder interface, small hint components provide additional frequently-used information where it might be needed, so that users don’t need to look up things in the documentation too often. Right now, these hints are not as frequent as they could be, and one area where they are missing is the part of the UI where users define the responses available to participants. For example:

  • It is not entirely clear what the filter and target fields are for.
  • Similarly, the use of the label field is not directly apparent.

These questions are covered in the documentation:

Your mission, should you choose to accept it, would be to add the most important parts of this information to the builder UI. A hint would be a great help to many users – should you find a way to improve the UI along the way, you are in no way limited to the hint alone.

Updated 15/05/2017 19:16 4 Comments

Add blank line between title and message in commits



There should be a blank line between the “title” of a commit message and the full description of the message.

Steps to Reproduce

  1. Stage a file for committing
  2. Write a title (should have a character count on this line), hit enter.
  3. Write a description (should have ∞ as the character count)
  4. Commit the change
  5. View the log

Expected behavior:

Resulting commit message matches what other tools produce.

Actual behavior:

Atom fails to insert a blank line between the title and the detailed description.

Reproduces how often:

100% of the time.



Tested on a custom Atom build based on merging and into

Atom : 1.18.0-dev-969a4da6c Electron: 1.6.9 Chrome : 56.0.2924.87 Node : 7.4.0

git: git version

GitKraken: v2.5.0

Additional Information

Atom: image

git CLI: image Note how it highlights any text placed on that line in orange/red!

As an example of how other implementations handle this, here’s GitKraken: image

Result: image

When viewing the commits on GitHub it handles the lack of a blank line in the same manner as the other tools, but as the blank line is what the official git client does and what other tools do it would be good to follow their lead. It also makes parsing of the messages simpler 😉.

Updated 15/05/2017 08:25 3 Comments

Several audits do not provide critical information


Lighthouse audits are failing to provide some critical information on where best practices are being broken on the audited page.

Here are some issues I’ve found so far: - The color contrast audit provides no information on the failing elements - The “rel=‘noopener’” audit provides no information on the failing elements - The “images have alt” audit provides no information on the failing elements - The “uses passive listeners” audit no longer provides a code snippet or column number, only line number

Updated 23/05/2017 20:44 1 Comments

Not working with react-native


The image shows how far i can get. screen shot 2017-05-11 at 10 35 27 am

Nothing happens when clicking on the View in the inspector nor on the children objects in the developer tools. When using the inspector in the simulator i can highlight elements but no details show up. When using the chrome developer tools the react-devtools windows stays on Connecting to React… Also tried the react-devtools chrome extension which keeps saying This page doesn’t appear to be using React. Also tried the stand alone react native debugger but inspector does not show up. I remember at at one stage the inspector in the react native debugger was working. i didn’t really need it at the time but now i really need to inspect elements style. Any help would be appreciated.


Also maybe related: after reloading the app i always get the error: Expected to run on UI thread, which i only started getting after the last update.

“react”: “16.0.0-alpha.6”, “react-native”: “0.44.0”, Mac OSX Sierra, Android API 25

Updated 29/05/2017 14:01 6 Comments

Clearing partial data destroys all tabs


Description: Clearing partial data destroys all tabs

Did you search for similar issues before submitting this one? Yes

Expected behavior: Should show a warning about tabs being lost or should retain the tabs based on selected options

  • Device (iPhone5, iPhone6s plus, iPad 3, ?): All

  • Brave Version: 1.3.3 & 1.3.2 existing issue

  • Steps to reproduce:

    1. Open few tabs and visit different sites
    2. Open settings -> Clear browsing data -> Select Cache & Saved Logins -> Clear browsing data
    3. Tap on done, All opened tabs are lost

** Website problems only: N/A - did you check with Brave Shields down? N/A - did you check in Puffin/Dolphin/iCab/Maxthon (UIWebView-based browsers)? N/A

  • Screenshot if needed:

  • Any related issues:

Updated 02/06/2017 20:33

Button Name Audit


Minor note: not sure how to fix this audit.

Buttons have an accessible name. An accessible name helps convey the purpose of a button. Without one, a screen reader will only announce the word “button”.

The description doesn’t explain how to fix it or link out to learn more :(

From @robdodson

put text inside them, or give them a label, or give them an aria-label

Updated 23/05/2017 20:43 9 Comments

Log SQL queries in development mode


It would be nice to see all the SQL queries sent to the server printed to standard Docker output (in development mode only).

More info:


Note: The config/ file is being executed inside the db service only when data volume is missing while creating a new database from scratch.


Want to back this issue? Post a bounty on it! We accept bounties via Bountysource. </bountysource-plugin>

Updated 10/05/2017 11:06

PR labels unreadable in dark themes



The labels associated with a PR use the UI theme’s text, but the background color from GitHub, leading to conflicts that can make them illegible.

Steps to Reproduce

  1. Have a branch checked out with an associated PR that has a label
  2. Attempt to read the label(s) of the PR

Expected behavior:

Labels are legible.

Actual behavior:

Since the labels use the theme color for text, if the theme uses a light color and the label has a light background it becomes near unreadable.

Atom Material (v2): image

One Dark: image

One Light: (as a usable example) image

The label on image

Reproduces how often:

100% of the time.


Windows 10 x64

Build from

Atom : 1.18.0-dev-4cbff3044 x64 Electron: 1.3.15 Chrome : 52.0.2743.82 Node : 6.5.0

github@0.0.2 (bundled in the build)

Additional Information

The PR/Branch in question was

Updated 12/05/2017 06:33 1 Comments

Change "XML/SVG" option label in the editor dropdown menu.


This setting lets you edit SVG files…

<img src=“” width=“350” />

The label is totally cryptic and doesn’t give any indication of what it actually does. Let’s change it to something like…

  • Edit SVG images as Code
  • SVG Image Editing
  • …any other ideas?
Updated 07/06/2017 18:14 6 Comments

Fork me on GitHub