Live data from Hacker News

PhantomJS 2

github.com

11–20 of 24 posts

Re: PhantomJS 2

#11

Can anyone actually explain what is so important in Version 2? All I can see in release-2.0 notes, is an upgrade to Qt5

There are better release notes in the download: --- 2015-01-23: Version 2.0.0 New features * Switched to Qt 5 and updated WebKit (issue 10448) * Implemented clearing of memory cache (issue 10357) * Added support for HTTP header change for every request (issue 11299) Improvements * Fixed rendering of CJK text by always linking the codecs (issue 10249) * Ensured onResourceReceived is still fired on an error (issue 1116…

Function.prototype.bind[1] issue was quiet painful to many, so this release will be a relief.

[1]https://github.com/ariya/phantomjs/issues/10522

Re: PhantomJS 2

#12

Can anyone actually explain what is so important in Version 2? All I can see in release-2.0 notes, is an upgrade to Qt5

Phantom 1.x was using an ancient version of WebKit.

Re: PhantomJS 2

#13
post #5

The official OSX build ( http://phantomjs.org/download.html ) crashes for me with Killed: 9

I had the same problem, so I found a small patch[1] and recompiled several binaries, including one for OSX [2]

edit: a pull request has now been submitted and accepted [3]

[1] https://qt.gitorious.org/qt/qtbase/commit/70e4428b6f1c6a4bad... [2] https://github.com/eugene1g/phantomjs/releases [3] https://github.com/ariya/phantomjs/pull/12934

Re: PhantomJS 2

#14
Any chance for headless browser/testing stack with webgl support (mesa, llvm, even less than 1 fps would be enough)?

Re: PhantomJS 2

#15
It appears they've moved forward to Qt5 but they're still running on QtWebKit. This is a bad situation since QtWebKit has been deprecated (read: nobody has publicly committed to writing any new code for it, not least Digia) in favour of QtWebEngine, which doesn't have even 10% of the API richness of the old module.

WebEngine implements process separation and is based on Blink, whereas WebKit was in-process and e.g. allowed direct access to DOM from C++.

The job of moving PJS to WebEngine, if possible, when it happens will basically be the same effort as a rewrite. This release looks like a stopgap to get the best of the old WebKit line, but calling it 2.0 seems a bit shortsighted -- 2.0 should be kept for the WebEngine series if/when it happens.

Re: PhantomJS 2

#17
post #15

It appears they've moved forward to Qt5 but they're still running on QtWebKit. This is a bad situation since QtWebKit has been deprecated (read: nobody has publicly committed to writing any new code for it, not least Digia) in favour of QtWebEngine, which doesn't have even 10% of the API richness of the old module. WebEngine implements process separation and is based on Blink, whereas WebKit was in-process and e.g. a…

The most important thing we wanted to achieve was to get a fresh version of QtWebkit. Which was not possible without upgrading to Qt5. We had plans to switch to QtWebEngine in future versions. But in current state, we're thinking about moving to Chromium.

Re: PhantomJS 2

#18

Does it support webfonts yet (making it finally useful for screenshot capture)?

Only on Linux and OS X. Windows version doesn't support WebFonts yet. We will enable them in the next minor release (2.0.1). It will be available shortly.

Re: PhantomJS 2

#19
post #14

Any chance for headless browser/testing stack with webgl support (mesa, llvm, even less than 1 fps would be enough)?

Definitely not in PhantomJS 2 with QtWebkit. May be after we will switch to Chromium.

Re: PhantomJS 2

#20

Finally! is that official?

Kind of. Before we will do official announcement we have to complete following things: - compile binaries for Linux X86/X64 - update links/information blocks on our web site.
Post reply on HN