fyrd/caniuse

Raw browser/feature support data from caniuse.com

caniuse

About

This repo contains raw data from the caniuse.com support tables. It serves two purposes:

1. The ability for anyone interested to update or add to the support data on the site. If you are interested in this, please read the CONTRIBUTING file.

2. Access to the site's data for other projects. For this use the fulldata-json/data-2.0.json file which includes all support data. (data.json is also available for backwards compatibility.)

The data in this repo is available for use under a CC BY 4.0 license (https://creativecommons.org/licenses/by/4.0/). For attribution just mention somewhere that the source is caniuse.com. If you have any questions about using the data for your project please contact me here: https://a.deveria.com/contact

Sponsor

Caniuse.com is created and maintained by Alexis Deveria
https://caniuse.com/
https://a.deveria.com/

caniuse's Issues

the 'current' specification "as of the date recorded it"

Sometimes we see that the standard specification is fairly largely changed so the old one is now deprecated (including Fullscreen API as he said #124, Flexible Box Layout Module, and so on).

Currently we modify the record from 'y' to 'a' when we learned it, leaving a note which says "Partial support refers to supporting an earlier draft of the spec." or something. Maybe we don't have to record exactly what earlier draft the browser supported, but I think we have to record the version of the 'current' spec, so that everyone can note the browser 'may not' support the latest spec, even when no caniuse contributor get to know the spec is updated and switch 'y' to 'a'.

It will also help the contributors, because they can learn more easily that the current caniuse record is based on up to date spec or not, and if it's not, they will update it.

CSS Animation support on Safari & Opera

Kindle browser

It would be great to have the Kindle browser added to Can I use! I can provide data for the Keyboard and Touch.

Problems with H.264 in Chrome

There seem to be problems with playing H.264 videos, (especially long ones like > 30 minutes,) in latest versions of Chrome: http://oddlystudios.com/blog/html5-video-problems-in-chrome/

I've tested version 22 through 25, both in Windows and Linux, and experienced the same problem.

history specification reference at w3c - 404

http://caniuse.com/#feat=history
specification reference return 404
http://www.w3.org/TR/2011/WD-html5-20110113/history.html

atob btoa base64

what about testcases for this ?
http://www.whatwg.org/specs/web-apps/current-work/multipage/webappapis.html#atob

Add new browser Internet Explorer Mobile

More info:

http://en.wikipedia.org/wiki/Internet_Explorer_Mobile

transitionend and animationend support

JavaScript event to detect when CSS3 animations and transitions end https://plus.google.com/u/0/110977198681221304891/posts/PcrM4jPd7sX

• W3C WD http://dev.w3.org/csswg/css3-transitions/#transition-events
• MDN transitionend https://developer.mozilla.org/en-US/docs/Mozilla_event_reference/transitionend
• MDN animationend https://developer.mozilla.org/en-US/docs/Mozilla_event_reference/animationend

Opera 10.0-10.1?

Any reason why "Opera 10.0-10.1" is listed as a single version? This is a little confusing when coming from the API, e.g.

"You can use TTF/OTF - TrueType and OpenType font support with Firefox 3.5, Chrome 4, Safari 3.1, Opera 10.0-10.1, iOS Safari 4.2-4.3, and Opera Mobile 10. http://caniuse.com/#search=ttf"

Why not just Opera 10.00?

add user-select info

Option to customize browser statistic sources

Is it a place for site-specific suggestion?

Anyway, the site uses global usage stats, which is not always desirable. For example, Russian and CIS browser shares differ from global considerably (more Opera, less IE), so when working on something for this segment it'd be useful to take this into account.

It'd be even better if we could export custom usage stats from tracking services (Google Analytics and the likes) to get the data for the very project we're working on.

Different data structure for returning "usage_global" in agents part

In the "agents" part, two different structures are used in order to return information about "usage_global". All agents use a json object (version: value), except "and_chr" (Chrome for Android") and "and_ff" ("Firefox for Android"), that use a json array. Should be change the latter case in order to be more consistent ?

agents missing currentVersion property

I am trying to extrapolate what the current version of each browser is. The currentVersion of most of the browsers in agents is an empty string. Is this a mistake?

Is the versions array in each of the agents (currently 20 in length) always 20 long?

Can I assume that the last two entries are always far future and near future and that array[array.length-3] is always the current version?

Node library for this

Hi! I have written a node library for query the data from caniuse:
https://github.com/wuchengwei/caniuse
Just FYI, thanks!

Linear gradient vendor prefixes.

IE10 requires -ms- prefix. Firefox (at least version 16) still needs -moz- prefix.

And move IE10 to "Current" already, it's been released 3.5 months ago

SVG feature coverage

In issue #2 you note some submission guidelines for improving caniuse.com feature coverage, and as I have been missing higher-fidelity detailed data about what parts of SVG are supported where (similar to the good support coverage data CSS has) I figured I'd bring it up in a separate ticket.

At the moment, I make do with some prior work in the field done by Jeff Schiller (@codedread), where he has somewhat manually run the SVG test suite a couple of times over the years and across a swath of browsers (and plugin configurations). The data is a little unwieldy to access, find and read, though – here's a JSON gist I made of it today: https://gist.github.com/2803171

I believe "P" = pass, "F" = fail, "FU" = also fail and "H" = partial support (with aggregated notes in tests.json). The ones with crash: true crashed the testing browser.

The more interesting parts would of course be translating the SVG test suite format to your automatic tester, which it should be somewhat well suited to, as each of the listed tests have isolated test file, rendered look and test notes, as per your test suite requests (since browser makers use them for their own compliance and regression testing as well).

If there is anything else I can do to help make this reality, do tell.

Referencing Research Papers

The data URI table rightly states that some browsers don't support this feature for navigable types like text/html. It doesn't mention that this lack of support is intentional, and for security reasons.

The paper Phishing by Data URIs is relevant here. Is it appropriate to mention (and link) this resources (and those like it) within the Notes of related tables?

Current version of IE should be 9

IE 10 only has 0.85% share according to the usage table, and IE 9 has 15.18%. Also, IE10 has only been released on Win8, and the Win7 version is still a "preview".

(Related question: how do you determine the "current" version of a browser? Except for Android and IE, it always seems to be the one with the largest usage.)

begin tracking Google Blink

Add HTML5 attributes

HTML5 attributes http://dev.w3.org/html5/spec/single-page.html#global-attributes accesskey, draggable, dropzone, hidden, spellcheck, tabindex, translate, etc.

:last-child selector Works in IE 7 and IE 8

12.1 is the current version of Opera Mobile

caniuse.com is currently showing 12.0 as the current version of Opera Mobile; 12.1 has been out for a while.

It appears as though I should be able to correct this myself, but that would require editing the Single Line of Hell that is data.json, and Contributing.md gives me the impression that's not meant to be editable.

will you update data.json file?

@Fyrd I notice that data.json has not been updated for a long time. Will you update it?

Website does not work when cookies are disabled

I think at the very least you should get a reasonable message that this site only works with cookies, and for what reason (say what features do I, as a user, get in return for enabling/accepting cookies).

Steps to reproduce:

1. Disable cookies in your browser
2. Go to http://caniuse.com/

Show Browser Release Date / Add age-aligned mode

Hi,

Think about extending the UA info adding the release date.

Example, when you see feature X is not working with android 2.3 would be very very handy to quickly find when android 2.3 was public released.

What do you think?

---

Edit: Please share your love for this open issue, leave here your reactions or comments. We really want to see this feature online soon. Thanks!

CSS3 "appearance" property

CSS3 appearance property http://www.w3.org/TR/2004/CR-css3-ui-20040511/#appearance0
http://www.w3schools.com/cssref/css3_pr_appearance.asp
https://developer.mozilla.org/en-US/docs/CSS/-moz-appearance

Actually works prefixed just on Webkit.
Firefox related bug https://bugzilla.mozilla.org/show_bug.cgi?id=649849

I am not sure, but the property might be currently hold or removed from standards http://dev.w3.org/csswg/css3-ui/

No information on widows and orphans

Add a "tags" array for links

Currently the links array contains only the link title and url. The tags ("info", "reference", "demo", etc.) are, I suppose, added by hand during updates. It would be nice if these could also be specified in the json data files.

Website design could use improvement

I really love the features of the site, however the brown tones and edge gradients give me nausea. Could you switch to white/gray/black backgrounds with color only used as a flourish?

Thanks for a really great resource!

'Partial support' and 'Disabled by default'

Sometimes there are some notes like 'Partial support in Firefox refers to being disabled by default' as in SPDY item (http://caniuse.com/#feat=spdy). However, items including Web Audio, the 'disabled by default' browsers are treated as 'no support', as Firefox 21 partially supports it but with the option media.webaudio.enabled.

I think it's better to clarify this in the guidelines for contributing document.

I think additional

• d - not enabled by (d)efault

is also good, because sometimes some browsers disable some features because they are yet in WD status, while perfectly support current specifications.

Browser comparison doesn't display prefixes

http://caniuse.com/#compare=chrome+25,chrome+26
In this example, I should be able to see that Chrome 26 removed CSS prefixes for transitions and gradients.

BlobBuilder gone in Chrome

window.BlobBuilder do not exist on Chrome 24 anymore. I dont know if this started from a earlier version.

backslashes in URLs

"spec":"http:\/\/www.w3.org\/TR\/WOFF\/",
"url":"http:\/\/hacks.mozilla.org\/2009\/10\/woff\/",

Is it for escaping? According to http://www.ietf.org/rfc/rfc4627.txt it's not really required.

Add Chrome For Android

While many Android devices use the Android browser, increasingly people are using Google Chrome for Android instead — which has a different level of support for HTML5, CSS3 and related technologies. There is documentation at https://developers.google.com/chrome/mobile/docs/overview for what is supported. It would be very handy for this browser to be listed in the Can I Use charts.

inputmode attribute

IE6,7,8 and Firefox OS support it.

CSS "writing-mode" property

This property is not listed

https://developer.mozilla.org/en-US/docs/CSS/writing-mode
http://www.w3.org/TR/css3-writing-modes/

Standardize bugs array to use objects with "url" and "description" properties

That would make it easier to use a common format for bugs (eg: "Browser X only implements feature Y partially [bug 1234]") and make it easier to find the right issues in the bug trackers to follow (especially since searching the bug trackers directly can be quite daunting)

Allow filtering by browser version

The information is in the database, the hierarchical checkbox code is already there, so it would make sense to have the option to display browser versions for specific targeting. Maybe only three per browser (current, previous and upcoming) to avoid cluttering the interface.

Also, it would be really neat to have a threshold slider that would select/deselect the browsers based on usage stats. Again, both the data and the code to implement this are already present on the site (in the "versions shown" filter box), so it seems definitely doable.

mistake in http://caniuse.com/#search=transition

firefox 16 version still support only with moz prefix

tested in this example

http://webdesign.tutsplus.com/tutorials/htmlcss-tutorials/css3-transitions-and-transforms-from-scratch/

A Suggestion for pulling json data directly from GitHub Repository

Hi,

I just came across this article which talks about pulling file content directly from the github repository. (http://www.jamesward.com/2012/06/15/dynamically-rendering-github-files-in-web-pages)..

Probably, you guys know about this... if not, you can just give it a try..(if this approach makes life easy for you guys )

As of now my understanding is that the json data from github repository is updated to a database and from the database back to github as json periodically. (which I read from the Contributing.md file )

Cheers,
Ram

Show browsers in a more time friendly grid.

This is more of a suggestion than an issue. Since it doesn't really effect the purpose of the site it really isn't that important, but I think it is something worth considering...

Right now all of the browsers are aligned based on version. This is great, but it doesn't show the time relative to the other browsers. Browsers with a more frequent release schedule look like they have supported something for longer, when they really haven't supported it for very long, compared to poor picked on IE. It would be great if the tables could show the feature support in a more time friendly way. (Using rowspans perhaps?)

Let me know what you think.

Submitting new feature data sets

Hi,

I tried to get to this in an html5please issue discussion, but it was a bit off-topic there:

• How do I go about submitting new feature data?
• If it's not possible via pull requests, why not?

Thanks,
Sebastian

Canvas v5 methods

It would be nice to have some information about canvas v5 methods (e.g context.setLineDash() - supported in chrome).
http://www.w3.org/html/wg/drafts/2dcontext/html5_canvas/

Opera 12.10 only adds *partial* support for the Fullscreen API

According to Opera, version 12.10 implements an older version of the Fullscreen API spec:

Opera 12.10 adds partial support for the Fullscreen API that allows video, games or web pages to use the whole screen to remove distractions like browser chrome that can divert your attention from skateboarding kittens or shooting aliens. (We say "partial" because new "HTML5 the living-on-the-edge standard" specs chop and change. This release implements the Fullscreen API editor's draft 7 February 2012, while the standard has now mutated in the latest July 2012 version.)

Source: http://my.opera.com/ODIN/blog/opera-12-10-is-out

I compared the February 2012 version (implemented by Opera 12.10) with the latest editor’s draft (July 2012). From what I can see, the API stayed the same. However, the :fullscreen-ancestor pseudo-class has been dropped, and a ::backdrop pseudo-element has been added.

I'm not sure though, if this little change is enough to label Opera's implementation "partial" on http://caniuse.com/#feat=fullscreen (current label is "yes").

Chrome 22 or later for Windows supports touch event

Chrome for Windows supports touch event if hardware has touch-capability. (it is same as Firefox 18 or later for Windows)

More detail is http://code.google.com/p/chromium/issues/detail?id=138195.

so could you update the following json file for Chrome touch event support?

https://github.com/Fyrd/caniuse/blob/master/features-json/touch.json

Why do you use version "0" for Chr/And and FF/And

@Fyrd I'm just wondering, why not just use version "18" for Chr/And and "15" for FF/And?

Github edit backlinks

Hi,

I saw your post about allowing contributions through Github. We have a similarish project going on.

After adding a Github edit baclinks, using Github's thorugh-the-web file editor, the number of contributors went through the roof:

http://opensourcehacker.com/2012/01/08/readthedocs-org-github-edit-backlink-and-short-history-of-plone-documentation/

I'd suggest something similar could be added for caniuse.com, so that pages would a link back to Github file where the data can be contributed.

Please re-merge PR #107

Data update of 8e56639 overwrote d26a6e6 a few days ago.
Please re-merge PR #107.

webrtc

Currently Browser support is chrome 23 and firefox nightly.

It would be nice to have some entries for this

jQuery not loading

Could be the side effects of the hurricane, but jQuery isn't loading from Microsoft's servers. Is it possible for you to host a local version and fall back to that?