Function parse_rpm_xmlish

src/packagedcode/rpm_installed.py:39–62

Yield PackageData built from an RPM XML'ish file at ``location``. This is a file created with the rpm CLI with the xml query option.

(location, datasource_id, package_type, package_only=False)

def parse_rpm_xmlish(location, datasource_id, package_type, package_only=False):
    """
    Yield PackageData built from an RPM XML'ish file at ``location``. This is a file
    created with the rpm CLI with the xml query option.
    """
    if not location or not os.path.exists(location):
        return

    # there are smetimes weird encodings. We avoid issues there
    with open(location, 'rb') as f:
        rpms = as_unicode(f.read())

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

    for rpm_raw_tags in collect_rpms(rpms):
        tags = collect_tags(rpm_raw_tags)
        yield build_package(
            rpm_tags=tags,
            datasource_id=datasource_id,
            package_type=package_type,
            package_only=package_only,
        )


def collect_rpms(text):