It looks like there's another issue with Wikisource HTML in Parsoid:
The error translates as "error: no such index", which is PRP's proofreadpage_nosuch_index message.
The source of the page is:
<pages header=1 index="Aaj Bhi Khare Hain Talaab (Hindi).pdf" from=8 to=12 prev="[[आज भी खरे हैं तालाब|मुखपृष्ठ]]" next="[[आज भी खरे हैं तालाब/नींव से शिखर तक|नींव से शिखर तक]]" />
The index page definitely exists.
The PRP code that produces it is as follows, so perhaps this error is something to do with namespace IDs being handled differently in the two parsers?
$indexTitle = Title::makeTitleSafe( $this->context->getIndexNamespaceId(), $index ); if ( $indexTitle === null || !$indexTitle->exists() ) { return $this->formatError( 'proofreadpage_nosuch_index' ); }