PhantomJS 2 + Qt 5 plan

Ariya Hidayat

unread,

Apr 4, 2013, 10:25:30 AM4/4/13

to phan...@googlegroups.com

Folks,

As you already know, we plan to migrate to Qt 5 and call the next
iteration as PhantomJS 2.0 (for specific, track
https://github.com/ariya/phantomjs/issues/10448).

Since this is a major release, it gives us an additional chance to
clean-up other stuff as well, including obsolete/broken API. You are
welcomed to file an issue for every annoyance which has been bothering
you all these years.

If anyone is interested in tracking the general 2.0 milestone, see
https://github.com/ariya/phantomjs/issues?milestone=12. A gentle
reminder: there is no need to ask what's the status of a particular
issue, if there is some update/change/tweaks, it will be posted
anyway.

At some point the master branch will be ported to Qt 5. I believe
Vitaly has been playing with the Qt 5 version already, so expect some
cool stuff :)

There will be a 1.x branch to continue from the current master branch
(more or less 1.9 + new stuff). This will be denoted as 1.10 in the
version number. This branch will never be released. It can serve as a
test playground for new features implemented while the Qt5-based
version is not ready yet.

1.9 will be still maintained for *critical* bug fixes. In fact, I plan
to roll out 1.9.1 in a week or so.

With all the effort in parallel, I don't think it's reasonable to
follow the usual 3-month release cycle, at least for this particular
one. Let's designate the next equinox (Sep 22) as the target release
date for 2.0. For testing purposes (among others, downstream projects
like CasperJS, Poltergeist, Grunt, ...), I still do think it makes to
produce binaries as the form of preview/alpha/whatever.

Since everyone will be asking this, with respect to the new Blink from
Google (http://blog.chromium.org/2013/04/blink-rendering-engine-for-chromium.html),
we don't plan to switch to it in the near future. Having said that, I
think using Chromium (as a whole) will be a wise move (but not for the
immediate future). I am quite familiar with Chromium code base and I
have a rough understanding of what needs to be tweaked/implemented to
do that, however let's keep this more or less for the longer term
(https://github.com/ariya/phantomjs/issues/10209), probably for
PhantomJS 3.x timeline.

Thank you!

Best regards,

--
Ariya Hidayat, http://ariya.ofilabs.com
http://twitter.com/ariyahidayat
http://gplus.to/ariyahidayat

Vitaliy Slobodin

unread,

Apr 4, 2013, 1:37:56 PM4/4/13

to phan...@googlegroups.com

Hi, guys.

Indeed, I'm playing with Qt5 now. It works! And, well, PhantomJS crashes often (surprise! :))
I'm trying to fix it as soon as I encounter a crash. But there are a lots of work to make it stable.

I'll write a more detailed message on this weekend.

Regards,
Vitaly.

Ariya Hidayat wrote:

Folks,

As you already know, we plan to migrate to Qt 5 and call the next
iteration as PhantomJS 2.0 (for specific, track
https://github.com/ariya/phantomjs/issues/10448).

Since this is a major release, it gives us an additional chance to
clean-up other stuff as well, including obsolete/broken API. You are
welcomed to file an issue for every annoyance which has been bothering
you all these years.

If anyone is interested in tracking the general 2.0 milestone, see

https://g ithub.com/ariya/phantomjs/issues?milestone=12. A gentle

Nickolay

unread,

Apr 7, 2013, 3:35:58 AM4/7/13

to phan...@googlegroups.com

Good news!

Hope to see this issue fixed: https://github.com/ariya/phantomjs/issues/10444
Such basic API is expected to just work, indeed.

Also hope 2.0 will contain stable webkit version. For example we are still using phantom 1.6.0,
because we found everything >1.7.0 to be unstable (I blame webkit version). We are running Siesta
(bryntum.com/products/siesta/) and we have big test suites that runs fine on all browsers and phantom 1.6.0
and sporadically fails on phantom > 1.6.0.

Regards, Nickolay

dimisec

unread,

Apr 23, 2013, 2:57:12 AM4/23/13

to phan...@googlegroups.com

Would it be worth to make a "qt5" branch in the GitHub tree? It could make easier for folks to try the Qt5-based build and report issues etc.?

Dmitri

Message has been deleted

James Greene

unread,

Aug 2, 2013, 12:04:26 AM8/2/13

to phan...@googlegroups.com

The only one I'm aware of at the moment is Vitaliy's separate (non-forked) repo:

https://github.com/Vitallium/phantomjs-qt5/

Sincerely,
James Greene

On Wed, Jul 31, 2013 at 11:59 AM, Eric Heydenberk <eheyd...@monetate.com> wrote:

Hey everyone.

Is there currently a branch on ariya/phantomjs or on one of the forks to track the work to use Qt5? Is there a detailed discussion or comments somewhere about what the challenges of doing so might be? I'm interested in contributing but don't want to duplicate work and don't know where to begin other than copying the source for Qt5 over and chasing down the errors that result. In the meantime, I've been tracking this branch: https://github.com/Vitallium/phantomjs/tree/upgrade-qt-to-4.8.5 and building phantomjs binaries against it.

Thanks!

--
You received this message because you are subscribed to the Google Groups "phantomjs" group.
To unsubscribe from this group and stop receiving emails from it, send an email to phantomjs+...@googlegroups.com.
Visit this group at http://groups.google.com/group/phantomjs.

For more options, visit https://groups.google.com/groups/opt_out.

Vitaliy Slobodin

unread,

Aug 2, 2013, 3:02:46 AM8/2/13

to phan...@googlegroups.com

Hi!

I decided to keep our next version of PhantomJS in the separate repository, because I want to keep Webkit and Qt5 as submodules.
Webkit as a submodule give us the easy way to keep it in sync with the official version.

Eric, branch https://github.com/Vitallium/phantomjs/tree/upgrade-qt-to-4.8.5 will be merged to the master tree as soon as all errors will be solved. I'll finish it on this weekend.

So, a bit news about upgrading to Qt5:
- it works :) but really unstable.
- the only problem is QtWebkit. The static build completely broken (but I already fixed this)

Regards,
Vitaliy.

James Greene wrote:

The only one I'm aware of at the moment is Vitaliy's separate
(non-forked) repo:
https://github.com/Vitallium/phantomjs-qt5/

Sincerely,

James Greene< BR>

            iteration as P hantomJS 2.0 (for specific, track
            https://github.com/ariya/__phantomjs/issues/10448
            <https://github.com/ariya/phantomjs/issues/10448>).

            Since this is a major release, it gives us an additional
            chance to
            clean-up other stuff as well, including obsolete/broken
            API. You are
            welcomed to file an issue for every annoyance which has
            been bothering

&nbs p; you all these years.

If anyone is interested in tracking the general 2.0
milestone, see

https://g ithub.com/ariya/phantomjs/__issues?milestone=12
<http://ithub.com/ariya/phantomjs/issues?milestone=12>. A

            gentle
            reminder: there is no need to ask what's the status of a
            particular
            issue, if there is some update/change/tweaks, it will be
            p osted
            anyway.

            At some point the master branch will be ported to Qt 5. I
            believe
            Vitaly has been playing with the Qt 5 version already, so
            expect some
            cool stuff :)

            There will be a 1.x branch to continue from the current
            master branch
            (more or less 1.9 + new stuff). This will be denoted as

&nbsp ; 1.10 in the

            version number. This branch will never be released. It
            can serve as a
            test playground for new features implemented while the
            Qt5-based
            version is not ready yet.

            1.9 will be still maintained for *critical* bug fixes. In
            fact, I plan
            to roll out 1.9.1 in a week or so.

With all the effort in parallel, I don't t hink it's

            reasonable to
            follow the usual 3-month release cycle, at least for this
            particular
            one. Let's designate the next equinox (Sep 22) as the
            target release
            date for 2.0. For testing purposes (among others,
            downstream projects
            like CasperJS, Poltergeist, Grunt, ...), I still do think
            it makes to

&nb sp; produce binaries as the form of preview/alpha/whatever.

            Since everyone will be asking this, with respect to the
            new Blink from
            Google

(http://blog.chromium.org/__2013/04/blink-rendering-__engine-for-chromium.html
<http://blog.chromium.org/2013/04/blink-rendering-engine-for-chromium.html>),

we don't plan to switch to it in the near future. Having
said that, I

think using Chromium (a s a whole) will be a wise move

            (but not for the
            immediate future). I am quite familiar with Chromium code
            base and I
            have a rough understanding of what needs to be
            tweaked/implemented to
            do that, however let's keep this more or less for the
            longer term

(https://github.com/ariya/__phantomjs/issues/10209
<https://github.com/ariya/phantomjs/issues/1 0209>),

            probably for
            PhantomJS 3.x timeline.

            Thank you!

            Best regards,

            --
            Ariya Hidayat, http://ar iya.ofilabs.com

            <http://iya.ofilabs.com>
            http://twitter.com/__ariyahidayat
            <http://twitter.com/ariyahidayat>
            http://gplus.to/ariyahidayat

    --
    You received this message because you are subscribed to the Google
    Groups "phantomjs" group.
    To unsubscribe from this group and stop receiving emails from it,
    send an email to phantomjs+...@googlegroups.com

<mailto:phantomjs%2Bunsu...@googlegroups.com>.

Stefano Pongelli

unread,

Aug 26, 2013, 5:39:49 AM8/26/13

to phan...@googlegroups.com

Hi

good to hear! how is it going? ..does this mean we will see 2.0 next month? :)

(no webkit 536+ in 1.9.2, right?)

Cheers,

Stefano

Message has been deleted

James Greene

unread,

Sep 12, 2013, 11:56:03 AM9/12/13

to phan...@googlegroups.com

Initial discussion among the team is to delay the PhantomJS 2.0 release and integrate Qt WebEngine instead of doing double the effort. The end result will be awesome for users but I'm sure more than a few are going to get irritated in waiting longer for a new build with all the HTML5 goodies. :-\

Since Digia is hoping to release a preview this autumn with Qt 5.2, we would theoretically be able to release a PhantomJS 2.0 build (or maybe an alpha/beta build) based on the preview version so long as it seems to work well. Then when the official release comes next spring (Qt 5.3), we can just do a minor/patch release bump depending on how many changes go in.

Sincerely,
James Greene

On Thu, Sep 12, 2013 at 10:49 AM, Dmitri S <dsav...@gmail.com> wrote:

Regarding http://blog.qt.digia.com/blog/2013/09/12/introducing-the-qt-webengine/ - how are the QT plans to base their future engine on Chromium going to affect PhantomJS? Is it something for 3.0?