Developers › Desktop Team blog
Full ECMAScript (JavaScript) 5.1 support
This build adds full support for the ECMAScript (JavaScript) 5.1 specification.
At time of writing the test262 test suite is reporting one error from a total of 10927 tests on the Windows and Mac platforms. As a comparison, Opera 11.51 on Windows has 3865 failures. See the known issues below for information on Linux.
Keep in mind that the test262 test suite is undergoing active development and far from complete. Even a perfect score in the current version of the test suite does not guarantee perfect ES 5.1 conformance. Results may also vary depending on when the test was run, as tests could be added, removed, fixed or otherwise altered at any time. Case in point: The only test that fails in this snapshot, S11.1.1_A2, is invalid.
That leaves Opera with a "perfect score" on Windows, Mac and 64-bit Linux in the current test262, while still keeping the unstable nature of the test suite in mind.
In addition to full ES 5.1 support, we have added a native implementation of ECMAScript typed arrays. This offers increased performance when dealing with things like binary data, and will be noticeable when, for example, using emulators or playing games using Canvas.
This build also adds support for <video muted> to mute HTML5 videos by default, and Linux users can enjoy GTK+3 support which means that Opera looks native again on Gnome 3 based systems (like Fedora).
Highlights
Known issues
WARNING: This is a development snapshot: It contains the latest changes, but may also have severe known issues, including crashes, and data loss situations. In fact, it may not work at all.
Download
At time of writing the test262 test suite is reporting one error from a total of 10927 tests on the Windows and Mac platforms. As a comparison, Opera 11.51 on Windows has 3865 failures. See the known issues below for information on Linux.
Keep in mind that the test262 test suite is undergoing active development and far from complete. Even a perfect score in the current version of the test suite does not guarantee perfect ES 5.1 conformance. Results may also vary depending on when the test was run, as tests could be added, removed, fixed or otherwise altered at any time. Case in point: The only test that fails in this snapshot, S11.1.1_A2, is invalid.
That leaves Opera with a "perfect score" on Windows, Mac and 64-bit Linux in the current test262, while still keeping the unstable nature of the test suite in mind.
In addition to full ES 5.1 support, we have added a native implementation of ECMAScript typed arrays. This offers increased performance when dealing with things like binary data, and will be noticeable when, for example, using emulators or playing games using Canvas.
This build also adds support for <video muted> to mute HTML5 videos by default, and Linux users can enjoy GTK+3 support which means that Opera looks native again on Gnome 3 based systems (like Fedora).
Highlights
- CORE-28638 Implement full support for new ECMAScript 5 features
- CORE-38784 Native implementation of ECMAScript typed arrays for increased performance
- CORE-40275 Implement <video muted>
- DSK-343260 Support GTK+3 on Linux
- CORE-33418 Update to the Unicode 6.0.0 specification
Known issues
- Two valid tests in the test262 test suite fail on 32-bit Linux (works on 64-bit systems)
WARNING: This is a development snapshot: It contains the latest changes, but may also have severe known issues, including crashes, and data loss situations. In fact, it may not work at all.
Download
Changelog
Desktop
- DSK-342300 Manually stopping Link should completely "unlink" the client from the synchronization process
- DSK-318061 Dragging address badge to address or search field replaces page address with page's meta description
- DSK-332429 Intranet hosts not remembered in opera:config#UserPrefs|IntranetHosts and info bar doesn't go away on first click
Core
- CORE-39787 SVG element styled as table-row-group leaks memory
- CORE-40441 Identical floated tables not always positioned in the same place alongside floated block
- CORE-10745 Style updates to <col> don't cause reflow
- CORE-40472 Missing horizontal scrollbar on container with overflow:auto when a multicol container contains more columns than specified
- CORE-40470 Properties recognized as a style rule when setting properties on style object for other rules
- CORE-40409 IDN domains do not work
- CORE-38784 Native implementation of ES typed arrays
- CORE-40410 Crash at studivz.net
- CORE-40387 'playing' is not fired after play() is called - progress spinner on top of video (YouTube/HTML5)
- CORE-40353 Change in DOM causes focus to be moved to another contentEditable element within the page
- CORE-35502 Don't send Accept-Encoding HTTP header for <video>
- CORE-39892 Problem with horizontal padding calculation
- CORE-40275 Implement <video muted>
- CORE-39405 Spatnav then JavaScript focus() does not remove CSS :hover styling on previous item
- CORE-22646 :enabled :disabled and :checked UI selectors fail with cloned nodes
- CORE-28884 Inset box-shadow for <input>
- CORE-34144 DOMAttrModified event dispatched for unchanged attribute
- CORE-40593 arc function in canvas requires last argument (it should be optional)
Readers' comments
Comments left by My Opera members under the original post: the first 50, as archived.
About this article
| Published in | Desktop Team blog |
|---|---|
| Date | September 13, 2011 |
| Author | Haavard |
| Original | Internet Archive |
Krzysztof HankiewiczlordmorgulTuesday, September 13, 2011 3:08:01 PM
I love you guys.
arnaud lautierGrouickTuesday, September 13, 2011 3:08:28 PM
DustinKhadgarTuesday, September 13, 2011 3:12:02 PM
Алексейkudesnik84Tuesday, September 13, 2011 3:12:23 PM
Helder Jefferson Ferreira da LuzDante003Tuesday, September 13, 2011 3:13:15 PM
Алексейkudesnik84Tuesday, September 13, 2011 3:13:25 PM
BS-HarouTuesday, September 13, 2011 3:13:40 PM
HenryAOTEAROAnzTuesday, September 13, 2011 3:16:59 PM
AndreioperaterrestrialTuesday, September 13, 2011 3:21:07 PM
Patrick O'Reillypaddy2kTuesday, September 13, 2011 3:25:16 PM
Fabrice Bellard's browser based virtual machine zips along with this build. http://bellard.org/jslinux/
spancherBobTuesday, September 13, 2011 3:30:08 PM
QuHnoTuesday, September 13, 2011 3:34:41 PM
David TsunamydavidtsunamyTuesday, September 13, 2011 3:38:44 PM
One bug: Flashplayer not hidden on Google Reader:
I am using Opera on OS X Lion.
BinbiK-GergoTuesday, September 13, 2011 3:40:30 PM
Michael A. Puls IIburnout426Tuesday, September 13, 2011 3:41:37 PM
ClashCityRockerclashcityrockerTuesday, September 13, 2011 3:53:24 PM
Anyone told Microsoft about Microsoft's score? I'm guessing they won't be using test262 in their marketing anymore now, and move to some other cherry-picked benchmark.
Swapnil RustagiSwapnil99proTuesday, September 13, 2011 3:56:50 PM
Andrew FrazierAndrewDFrazierTuesday, September 13, 2011 3:56:52 PM
mtaki14Tuesday, September 13, 2011 4:00:52 PM
p/s
How do you post image previews here?
nimu_channimuchanTuesday, September 13, 2011 4:03:19 PM
Jordan TrillesCristallixTuesday, September 13, 2011 4:06:18 PM
Marco MaierschwiebieTuesday, September 13, 2011 4:07:32 PM
AvolaTuesday, September 13, 2011 4:07:47 PM
blackcoderTuesday, September 13, 2011 4:12:14 PM
Swapnil RustagiSwapnil99proTuesday, September 13, 2011 4:18:55 PM
Originally posted by K-Gergo:
What the hell do you want. Why are you unnecessary posting off-topic comments. Opera is working on WebGL; all know that. And won't be telling you anything about when it will be integrated into Opera snapshots. Is there any moderator free to delete the comment of K-gergo?
Szőts Ákosszotsaki2Tuesday, September 13, 2011 4:27:35 PM
In one of the previous builds this was fixed, but after I installed this snapshot, this hit again instantly.
OS: Linux x86_64
hup.hu: http://img690.imageshack.us/img690/6929/operahalfrendering.png
imageshack.us: http://img683.imageshack.us/img683/8158/operahalfrendering1.jpg
In Dragonfly for the site hup.hu (http://tinyurl.com/6zth5p7 ) I see 11 requests and all of them were 200 or 304 except an ad company (403 because of Privoxy).
pepkin88Tuesday, September 13, 2011 4:28:11 PM
Thank you
MichałPlMichalTuesday, September 13, 2011 4:30:15 PM
Jordan TrillesCristallixTuesday, September 13, 2011 4:30:47 PM
inset box-shadow doesn't work on <select> element and border-radius works only on the border on this element.
lastly, In the previous snapshot you add border-radius to < fieldset> when the < legend> element is here, but now when you have a border to the < fieldset> element the border continue under the < legend> element texte when he shouldn't do this.
Does i should make bug repport?
sleepyheadbooz0rTuesday, September 13, 2011 4:40:04 PM
AlternatiwealternatiweTuesday, September 13, 2011 4:44:16 PM
http://my.opera.com/haavard/blog/2011/07/01/test262
It's a little bit strange to me. It seems the tests are barely changed since that.
nick_el_sonnickelsonTuesday, September 13, 2011 4:46:32 PM
cascarudexTuesday, September 13, 2011 4:49:55 PM
HaavardhaavardTuesday, September 13, 2011 4:50:32 PM
Originally posted by alternatiwe:
Indeed, my points from that article still stand. In fact, I have repeated them in this blog post: The test suite is not finished, and using these results to market a browser as "the most JS compliant browser" would be misleading.You may notice that the post in my blog does not mention Opera. It is about how test262 results are being interpreted.
Edit: I have posted a new blog post on the subject.
Francisco HerrerafherreravTuesday, September 13, 2011 4:50:41 PM
ugly95Tuesday, September 13, 2011 4:55:03 PM
I had a similar issue. Although I'm pretty sure my 1060 build didn't crash when I closed it to install 1065.
I had the same 'crashed while trying to show the crash dialogue for a previous crash' message.
I couldn't even uninstall 1065. When I tried to uninstall through the Windows Add/Remove Programs I got the same crash message.
I had to roll back to 1060, back-up all my bookmarks and settings, then do an uninstall of 1060, do a clean install of 1065, then restore my backed up settings and it worked.
John A. Bilicki IIIjabcreationsTuesday, September 13, 2011 4:57:26 PM
If you folks want some more bragging rights how about edging out the WebKit team to be first for 100% compliance on the SVG test suite...
http://www.w3.org/Graphics/SVG/Test/20061213/htmlObjectHarness/full-index.html
...and be the first browser to score 100% compliance on Jeff Schiller's page here? (not many tests left and you might be able to squeeze it in for the 12.0 release
http://www.codedread.com/svg-support.php
kapsiTuesday, September 13, 2011 5:04:54 PM
BrandonViperAFKTuesday, September 13, 2011 5:09:01 PM
This build is a bit crashy for me though, it seems to crash somewhat consistently while I am adding pages to the speeddial.
Ola P. KleivenolakTuesday, September 13, 2011 5:22:41 PM
Originally posted by Cristallix:
No need to, already fixed internally.ouzowtfouzoWTFTuesday, September 13, 2011 5:42:57 PM
Danielweep86Tuesday, September 13, 2011 5:43:22 PM
What does this mean?
Opera will be more compatible, or what?
metudeTuesday, September 13, 2011 5:43:28 PM
Originally posted by avola:
+1 same for me. But i think i found reason; Opera Voice Library.
David Manndnm240Tuesday, September 13, 2011 5:52:10 PM
Quite impressive.
+1 for targeting complete SVG compliance, though I understand it may not be highest priority
Esteban Manchado VelázquezemanchadoTuesday, September 13, 2011 5:56:13 PM
ClashCityRockerclashcityrockerTuesday, September 13, 2011 6:06:17 PM
Helena PolovinčákováhelenkaTuesday, September 13, 2011 6:18:03 PM
Standard user, Lion (updater should ask for authorization)
ugly95Tuesday, September 13, 2011 6:25:18 PM
That makes sense. I had voice enabled when I updated the first time and had the crashing issue, and didn't enable it on the clean install, which seems to work fine.
And in 1065 when I try to enable voice I get a crash.
Michael ThompsonMichael83815Tuesday, September 13, 2011 6:25:56 PM
netmainTuesday, September 13, 2011 6:32:32 PM
seems like opera link causing crash here, everytime i activate it opera crash..
keep up the good work
w7 32bits SP1