No output when using webagg backend

Apologies for those of you seeing this for a second time. I've
received no response on matplotlib-users, so I'm turning to the
experts.

In one message, I wrote:

    I'm trying to use the webagg backend for the first time, and seem to
    be bumping into a common problem - nothing appears. The first time I
    tried, I got a complaint that tornado was missing, so I installed it.
    After that, I get to try plotting. Here's my example:

    >>> import matplotlib
    >>> matplotlib.use('webagg')
    >>> import matplotlib.pyplot as plt
    >>> plt.plot(range(10))
    [<matplotlib.lines.Line2D object at 0x2f56690>]
    >>> plt.show()
    Created new window in existing browser session.
    Press Ctrl+C to stop server
    ^CServer stopped

    When I execute plt.show(), a new tab opens in my browser (Chrome, on
    Linux) with this URL: http://127.0.0.1:8988. That page has a single
    link (text: "Figure 1"). If I click that link, I get this URL:
    http://127.0.0.1:8988/1, but nothing appears. The page source contains
    a bunch of JavaScript references. For instance:

    <script src="/_static/jquery/js/jquery-1.7.1.min.js"></script>
    <script src="/_static/jquery/js/jquery-ui.min.js"></script>
    <script src="/_static/mpl.js"></script>
    <script src="/1/mpl_interface.js"></script>

    I see those JavaScript files at appropriate places in the
    .../backends/web_backend directory. What am I missing?

then followed that up with a bit more information:

    I tried this at home on my Mac. Tornado was already installed, so I
    updated it to 3.2.1, and ran a Matplotlib app I have which allows me
    to set the backend on the command line. I get the exact same behavior
    as at work on my Linux desktop. I get a page at 127.0.0:8988 with a
    "Figure 1" link. Clicking that takes me to 127.0.0.1:8988/1, which is
    a blank page.

    Chrome is my default web browser on both Linux and the Mac. I manually
    visited the page from both Firefox and Safari, clicked the link,
    and... nothing. The same behavior as on Chrome.

    Then I had the bright idea to look at the JavaScript console. It shows
    this error message:

      Uncaught SyntaxError: Unexpected token &

    Clicking that error took me to this line:

    var fig = new figure(&#39;1&#39;, websocket_url_prefix);

    Are those entities supposed to be '1'? What bit of code generates
    that JavaScript?

Can someone provide a bit of direction? I can't find the code that
generates the above snippet of JavaScript, but Chrome, at least,
doesn't like it.

Thx,

Skip

I'm guessing that the complete lack of response both here and on
matplotlib-users means nobody uses the webagg backend?

Skip Montanaro

···

On Wed, May 7, 2014 at 9:04 AM, Skip Montanaro <skip@...503...> wrote:

I'm trying to use the webagg backend for the first time, and seem to
be bumping into a common problem - nothing appears.

No, it gets used. We just have a huge backlog of issue tickets to go through.

At first glance, I think this problem has already been reported and fixed in master. Can you try building from source the master branch on github and seeing if that fixes your problem?

Cheers!
Ben Root

···

On Mon, May 12, 2014 at 9:24 AM, Skip Montanaro <skip@…503…> wrote:

On Wed, May 7, 2014 at 9:04 AM, Skip Montanaro <skip@…503…> wrote:

I’m trying to use the webagg backend for the first time, and seem to

be bumping into a common problem - nothing appears.

I’m guessing that the complete lack of response both here and on

matplotlib-users means nobody uses the webagg backend?

Skip Montanaro


"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE

Instantly run your Selenium tests across 300+ browser/OS combos.

Get unparalleled scalability from the best Selenium testing platform available

Simple to use. Nothing to install. Get started now for free."

http://p.sf.net/sfu/SauceLabs


Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

Sorry, wasn't aware of the backlog.

Yes, I downloaded and built 1.4.x. That does seem to solve the problem.

Thanks!

Skip

···

On Mon, May 12, 2014 at 10:31 AM, Benjamin Root <ben.root@...553...> wrote:

No, it gets used. We just have a huge backlog of issue tickets to go
through.

At first glance, I think this problem has already been reported and fixed in
master. Can you try building from source the master branch on github and
seeing if that fixes your problem?