<html>
  <head>
    <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    Hi,<br>
    <br>
    <br>
    <blockquote
cite="mid:CAK5idxSKxBHE8_ufYAAKQqXH9+nEZEY7LTkwEWROmMrTzjwvgg@mail.gmail.com"
      type="cite">
      <pre wrap="">
I had proposed you to create an online table, where
NumPy functions are split into the following categories:
* already ported
* under development (preferably with name of the person who works on it)
* not started yet
Is it possible to get the data like this? This way other possible
contributors (maybe including me) could select an appropriate (i.e. they are
capable of) function for contribution, and other people could estimate the
progress done.</pre>
    </blockquote>
    <br>
    &gt; Yes, we're working on such thing. The temporary location is at
    <a class="moz-txt-link-freetext" href="https://bitbucket.org/fijal/hack2/src/default/numready">https://bitbucket.org/fijal/hack2/src/default/numready</a>, we'll
    probably
    create a nightly static HTML.<br>
    <br>
    Well, I just got this:<br>
    <div class="traceback">
      <ul>
        <li>
          <div class="frame" id="frame-140522643700112">
            <h4>File <cite class="filename">"/usr/local/lib/python2.7/dist-packages/Flask-0.8-py2.7.egg/flask/app.py"</cite>,
              line <em class="line">1518</em>, in <code
                class="function">__call__</code></h4>
            <pre>return self.wsgi_app(environ, start_response)</pre>
          </div>
        </li>
        <li>
          <div class="frame" id="frame-140522643698000">
            <h4>File <cite class="filename">"/usr/local/lib/python2.7/dist-packages/Flask-0.8-py2.7.egg/flask/app.py"</cite>,
              line <em class="line">1506</em>, in <code
                class="function">wsgi_app</code></h4>
            <pre>response = self.make_response(self.handle_exception(e))</pre>
          </div>
        </li>
        <li>
          <div class="frame" id="frame-140522643699600">
            <h4>File <cite class="filename">"/usr/local/lib/python2.7/dist-packages/Flask-0.8-py2.7.egg/flask/app.py"</cite>,
              line <em class="line">1504</em>, in <code
                class="function">wsgi_app</code></h4>
            <pre>response = self.full_dispatch_request()</pre>
          </div>
        </li>
        <li>
          <div class="frame" id="frame-140522643699728">
            <h4>File <cite class="filename">"/usr/local/lib/python2.7/dist-packages/Flask-0.8-py2.7.egg/flask/app.py"</cite>,
              line <em class="line">1264</em>, in <code
                class="function">full_dispatch_request</code></h4>
            <pre>rv = self.handle_user_exception(e)</pre>
          </div>
        </li>
        <li>
          <div class="frame" id="frame-140522643699664">
            <h4>File <cite class="filename">"/usr/local/lib/python2.7/dist-packages/Flask-0.8-py2.7.egg/flask/app.py"</cite>,
              line <em class="line">1262</em>, in <code
                class="function">full_dispatch_request</code></h4>
            <pre>rv = self.dispatch_request()</pre>
          </div>
        </li>
        <li>
          <div class="frame" id="frame-140522643699984">
            <h4>File <cite class="filename">"/usr/local/lib/python2.7/dist-packages/Flask-0.8-py2.7.egg/flask/app.py"</cite>,
              line <em class="line">1248</em>, in <code
                class="function">dispatch_request</code></h4>
            <pre>return self.view_functions[rule.endpoint](**req.view_args)</pre>
          </div>
        </li>
        <li>
          <div class="frame" id="frame-140522643699792">
            <h4>File <cite class="filename">"/home/dmitrey/Install/hack2/numready/numready.py"</cite>,
              line <em class="line">20</em>, in <code class="function">index</code></h4>
            <pre>pypy = read_all_numpy_funcs(sys.executable, 'numpypy')</pre>
          </div>
        </li>
        <li>
          <div class="frame" id="frame-140522643700048">
            <h4>File <cite class="filename">"/home/dmitrey/Install/hack2/numready/_numready/process.py"</cite>,
              line <em class="line">21</em>, in <code class="function">read_all_numpy_funcs</code></h4>
            <pre>assert not err</pre>
          </div>
        </li>
      </ul>
      <blockquote>AssertionError</blockquote>
    </div>
    Well, it doesn't matter essentially, from the code I just see it
    parses dir(numpypy) and doesn't provide information on the funcs
    been worked on, thus maybe my own efforts to port a func will be
    just a waste of time, because someone other will make it done before
    me.<br>
    <br>
    So, I've tried today's nightly build;<br>
    functions <b>ravel</b>, <b>flatten</b> and <b>where</b>, that are
    extremely important (they occur very often) are unimplemented yet;<br>
    <br>
    Extremely<b> </b>important function<b> dot </b>for matrix
    multiplication has very strange behavior:<br>
    <br>
    &gt;&gt;&gt;&gt; np.dot(np.ones((2,4)),np.array([1,2,3,4]))<br>
    20.0<br>
    CPython result:<br>
    &gt;&gt;&gt; np.dot(np.ones((2,4)),np.array([1,2,3,4]))<br>
    array([ 10.,  10.])<br>
    <br>
    I haven't tested any further, it already enough to make a sad
    conclusion about current numpypy progress and quality. In either
    way, I would like to have possibility to contribute some funcs, what
    should I do? Can I obtain git edit rights (or you're using something
    else, like mercury or svn?)<br>
    Also, I would like to stay tuned or even participate in numpypy
    discussion, but mail list already has high flow wrt JIT backends and
    other info that has no interest to me, I guess reading it is only a
    waste of time for many other PyPy-numpy users. Is it possible to
    create other mail list, e.g. PyPy-Math, for discussion of numpy,
    scipy ports and other related info?<br>
    <br>
    -----------------------<br>
    Regards, D.<br>
    <a class="moz-txt-link-freetext" href="http://openopt.org/Dmitrey">http://openopt.org/Dmitrey</a><br>
    <br>
    <br>
    <blockquote
cite="mid:CAK5idxSKxBHE8_ufYAAKQqXH9+nEZEY7LTkwEWROmMrTzjwvgg@mail.gmail.com"
      type="cite">
      <blockquote type="cite">
        <pre wrap="">
Also, is it possible to install recent numpypy without reinstallation of
whole PyPy?
</pre>
      </blockquote>
      <pre wrap="">
Nope.



</pre>
    </blockquote>
    <br>
  </body>
</html>