feedparser depends on order of description and content

[mosasiru]

parser is affected by order by descriotion and content:encoded.
A test case is below.

<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/">
<channel>
    <title>title</title>
    <link>http://www.example.com/</link>
    <item>
        <title>title2</title>
        <description>hoge</description>
        <content:encoded><![CDATA[
                fuga
        ]]></content:encoded>
        <link>http://example.com/2.html</link>
    </item>
    <item>
        <title>title1</title>
        <content:encoded><![CDATA[
                fuga
        ]]></content:encoded>
        <description>hoge</description>
        <link>http://example.com/1.html</link>
    </item>
</channel>
</rss>

Above two entries' description and content:encoded are just same except order.
But the result is not same..

In [4]: a.entries[0].content
Out[4]: [{'base': '', 'language': None, 'type': 'text/html', 'value': 'fuga'}]
In [6]: a.entries[1].content
Out[6]:
[{'base': '', 'language': None, 'type': 'text/html', 'value': 'fuga'},
 {'base': '', 'language': None, 'type': 'text/plain', 'value': 'hoge'}]

In [5]: a.entries[0].description
Out[5]: 'hoge'
In [7]: a.entries[1].description
Out[7]: 'fuga'

It seems because
(1)content is copied to sumary
https://github.com/kurtmckee/feedparser/blob/develop/feedparser/namespaces/_base.py#L482-L483
(2)summary is set to content
https://github.com/kurtmckee/feedparser/blob/develop/feedparser/namespaces/_base.py#L428-L430
this behavior is affected by order, it seems strange to me.

[mosasiru]

The copyToSummary behavior seems to me strange.
but if you think the behavior should be kept, new context variable may be needed.