Function collect_rpms

src/packagedcode/rpm_installed.py:65–120

Yield lists of RPM raw tags, one list for each RPM from an XML'ish ``text``. The XML'ish format is in fact multiple XML documents, one for each package wrapped in an root element. Therefore, we add a global root element to make this a valid XML document before parsing i

(text)

def collect_rpms(text):
    """
    Yield lists of RPM raw tags, one list for each RPM from an XML'ish ``text``.

    The XML'ish format is in fact multiple XML documents, one for each package
    wrapped in an <rpmHeader> root element. Therefore, we add a global root
    element to make this a valid XML document before parsing it.

    Each XML document represents one RPM package and has this overall shape:
        <rpmHeader>
        ....
          <rpmTag name="Name"><string>perl-Errno</string></rpmTag>
          <rpmTag name="Version"><string>1.30</string></rpmTag>
        ......
          <rpmTag name="License"><string>GPL+ or Artistic</string></rpmTag>
        ....
          <rpmTag name="Filesizes">
            <integer>6863</integer>
            <integer>2629</integer>
          </rpmTag>
          <rpmTag name="Filemodes">
            <integer>33188</integer>
            <integer>33188</integer>
          </rpmTag>
        ...
        </rpmHeader>

    After wrapping in a top level element we get this shape:
    <rpms>
        <rpmHeader/>
        <rpmHeader/>
    </rpms>

    When parsed with xmltodict we end up with this structure:
    {'rpms':  {'rpmHeader': [
        {'rpmTag': [
            {'@name': 'Name', 'string': 'boost-license1_71_0'},
            {'@name': 'Version', 'string': '1.71.0'},
            {'@name': 'Filesizes', 'integer': ['0', '1338']},
            {'@name': 'Filestates', 'integer': ['0', '0']},
        ]},
        {'rpmTag': [
            {'@name': 'Name', 'string': 'foobar'},
            {'@name': 'Version', 'string': '1.0'},
            ...
        ]},
    }}

    Some of these structures are peculiar as there are parallel lists (for
    instance for files) and each name comes with a type.
    """
    text = f'<rpms>{text}</rpms>'
    parsed = xmltodict.parse(text, dict_constructor=dict)
    rpms = parsed['rpms']['rpmHeader']
    for rpm in rpms:
        yield rpm['rpmTag']