[15][qweb report pdf] utf-8 encoding problem

[thartmann15]

Impacted versions: odoo 15 on ubuntu 20.04 with wkhtmltopdf 0.12.6 (with patched qt)

Steps to reproduce: Create a PDF using qweb report which has umlauts or some other non-ASCII characters.

Current behavior: The umlauts are not correctly displayed in the PDF.

problem: PDFs created using qweb reports don't display umlauts (or non-ASCII utf-8 characters) correctly.

the underlying issue: the PDFs are created by running wkhtmltopdf on some temporary html files.
in odoo 14, the umlaut 'ü' was encoded in the temporary body html as ' & # 252 ; ' (spaces inserted in order to display it on github).
in odoo 15, the umlaut 'ü' is used as it is.

there are two possible solutions:

(A) One can explicitly specify the encoding as argument to wkhtmltopdf. add
command_args.extend(['--encoding', 'utf-8'])
to _build_wkhtmltopdf_args(...) in ir_actions_report.py

(B) One can force lxml to output '& # 252 ;' . To this end, use
'body': Markup(lxml.html.tostring(node, encoding='ASCII')).decode('utf-8'),
and
body = ''.join(lxml.html.tostring(c, encoding='ASCII').decode('utf-8') for c in body_parent.getchildren())
in _prepare_html(...) in ir_actions_report.py

Could you please adopt one of the two solutions ?