PhantomJS 1.9: to be or not to be

378 views
Skip to first unread message

Ariya Hidayat

unread,
Mar 17, 2013, 10:35:13 PM3/17/13
to phan...@googlegroups.com
Folks,

This is a follow-up the previous discussion "Getting bleeding-edge
WebKit", see https://groups.google.com/d/msg/phantomjs/2Zb1Ryc5y8E/O16GOb25HUoJ
if you miss it.

My conclusion so far: even after some more twiddlings, I still can't
make my qt4-retrofit work properly. Unless someone volunteers and
gives me a helping hand, I can only two possible solutions:

(a) Continue finding ways to make qt4-retrofit work as expected
(b) Abort the idea, port PhantomJS to go straight to Qt 5.0

Consider that we're due to another release in a few days, this brings
some complication. Again, some possible different paths:

(1) Release what we have "as is", version 1.9
(2) Combine with (b), skip this one and spend some months for Qt 5
porting, version 2.0

Arguments can be made for both cases. We don't have a huge expected
list of features anyway so skipping is not causing a lot of headache.
OTOH, having a continued maintenance release also doesn't hurt.

I am personally more in favor of 2-b, in particular because most
likely we are not in a hurry for another release. In addition, there
is no benefit of delaying Qt 5 porting, also because the fresh WebKit
in Qt 5 will benefit everyone and I'd rather spend the little
affordable time to work on it.


Thoughts? Feedback?


Regards,


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

Min Zhang

unread,
Mar 18, 2013, 4:44:59 AM3/18/13
to phan...@googlegroups.com
Hi Ariya,

Thanks for your great job! I'm currently using your PhantomJS to build a performance testing tool. Sorry I'm not an expect on QT so I may not help you on this building problem.
However, as a user, I would (very) like to see if v1.9 could fix my problems. So here is my thought: how about just tagging what we have "as is" as version 1.9 but not to release it? By say "not to release", I mean just keep the website, documentation, and etc. the same as v1.8.2.

BTW, I wish these problems could be solved in 1.9 or future releases
  • Bodysize wrong value: https://code.google.com/p/phantomjs/issues/detail?id=169
  • examples\netsniff.js does not record resources with state code 302 or 304 in HAR file since they don't have "startReply".
  • Some requests (onResourceRequested) would never have response (onResourceReceived).

Again, thanks for your great work!


Regards,

-Min Zhang

James Greene

unread,
Mar 18, 2013, 7:19:53 AM3/18/13
to phan...@googlegroups.com

Although I hate to miss a release... I'd actually prefer option 2-b as well because I feel like a few of the new features in 1.9 are only half-baked but I haven't had time to clean them up yet either. If we release them as-is, we'd need to at least mark them as unstable APIs for the time being.

As to Min's suggestion of tagging v1.9 but not releasing binaries for it, I don't see any point to that as you can just pull from master or the latest SHA as of the release sate to achieve the same effect... it would just add confusion, IMHO.

Min: we're happy to accept pull requests to fix those issues. None look like they're critical, although #3 is probably solved in this v1.9 codebase by using the `onResourceError` callback.

Sincerely,
   James Greene

--
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?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

Ariya Hidayat

unread,
Mar 18, 2013, 7:22:00 AM3/18/13
to phan...@googlegroups.com
> However, as a user, I would (very) like to see if v1.9 could fix my
> problems. So here is my thought: how about just tagging what we have "as is"
> as version 1.9 but not to release it? By say "not to release", I mean just
> keep the website, documentation, and etc. the same as v1.8.2.

How is this helpful? It certainly does not reduce much work from my
side. In addition, if you want to live bleeding-edge, you can already
grab the latest master branch and build it yourself.

A general public release is more useful for downstream projects
(CasperJS, Poltergeist, Grunt.js) as their users aren't likely going
to mess around with PhantomJS directly.

Jon Leighton

unread,
Mar 18, 2013, 8:02:26 AM3/18/13
to phan...@googlegroups.com
On 18/03/13 02:35, Ariya Hidayat wrote:
> Thoughts? Feedback?

Skipping this release and going straight to Qt 5 in 2.0 gets my vote.

Nicolas Perriault

unread,
Mar 18, 2013, 10:00:16 AM3/18/13
to phan...@googlegroups.com
On Mon, Mar 18, 2013 at 3:35 AM, Ariya Hidayat <ariya....@gmail.com> wrote:

> also because the fresh WebKit
> in Qt 5 will benefit everyone and I'd rather spend the little
> affordable time to work on it.

I fully support the idea a 2.0 with fresh Webkit, especially if your
motivation is there :)

+1 for 2-b

++

--
Nicolas Perriault
https://nicolas.perriault.net/
Phone: +33 (0) 660 92 08 67

dimisec

unread,
Mar 18, 2013, 10:45:21 AM3/18/13
to phan...@googlegroups.com
another +1 for 2-b -- looking forward to the up-to-day WebKit in PhantomJS. 

Thanks for your hard work and the great super-useful project!

Ariya Hidayat

unread,
Mar 19, 2013, 12:34:12 AM3/19/13
to phan...@googlegroups.com
Actually, there is a legitimate reason to still release 1.9: update of
GhostDriver (see https://github.com/ariya/phantomjs/pull/11146).
However, I think it's Ivan's call to decide whether this needs to go
out ASAP or whether it can wait (I'm fine either way).

Ivan, what do you think?

Raj Dudi

unread,
Mar 19, 2013, 6:58:52 AM3/19/13
to phan...@googlegroups.com
I was waiting for 1.9 as it has the bug fix for the issue 10985 and many of my tests are waiting for phantomjs to have that fix.
In case we do not release 1.9 soon, then I will have to build it by my own and use it.
I dont think that there is any workaround to resolve this issue with current release of 1.8

Ivan De Marino

unread,
Mar 19, 2013, 8:51:34 AM3/19/13
to phan...@googlegroups.com
Sorry for the delay in answering.

Yes, for our WebDriver users 1.9 is essential: it bring GD 1.0.3 out!
I have spent too much time on it to NOT release it.

After this, I agree: Qt 5 is so tasty we should move.
But 1.9 is definitely needed.

I have "sold 1.9" already to many, based on the fact that I saw Ariya making this page: http://phantomjs.org/release-1.9.html
If it doesn't go out, I'll have to produce alternative builds that, in my opinion, will end up being very confusing (it's difficult as it is to make people understand what GD is respect to PhantomJS).


--
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?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.





--
Ivan De Marino
Coder, Technologist, Cook, Italian

blog.ivandemarino.me | www.linkedin.com/in/ivandemarino | twitter.com/detronizator

Ivan De Marino

unread,
Mar 19, 2013, 8:54:41 AM3/19/13
to phan...@googlegroups.com
A bit more wording.

It would be advisable to inform people that 1.9 is our last release on Qt 4.8 and that from now on work will focus on moving to Qt 5.
This way people will know in advance what to expect.

Many might say "yes, but my issue...", but we can always point out at those information to say "we have warned you".

Ariya Hidayat

unread,
Mar 19, 2013, 11:53:36 AM3/19/13
to phan...@googlegroups.com
> It would be advisable to inform people that 1.9 is our last release on Qt
> 4.8 and that from now on work will focus on moving to Qt 5.
> This way people will know in advance what to expect.

I like the idea. We'll brand 1.9 as "GhostDriver update" release and
it's the last one in the 1.x series.


Thank you!

Ivan De Marino

unread,
Mar 19, 2013, 6:41:37 PM3/19/13
to phan...@googlegroups.com
Ah, also this is now needed for Selenium CI tests :)

I know it's their project and they should deal with that, but it's nice to have them also stress-testing our driver and they now need a 1.9.0.

Qt5: any plan to use the new QJSEngine for the main execution context in PhantomJS, and ditch the "special page" approach?
The only functionality I think we will have to port there is the stack-tracing we have got at the moment.


--
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?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.


Emil Andersson

unread,
Mar 20, 2013, 11:33:30 AM3/20/13
to phan...@googlegroups.com
Hi,

I am very interested in issue in the new request abort feature[1]. Will there be a 1.9 today containing it or should I make my own binaries from the git repository?

Thanks,
Emil

Reply all
Reply to author
Forward
0 new messages