As a Wikisource user, I want the Epubcheck error (as described below) fixed, so that the epub export process can be more reliable and efficient.
Background: Epubcheck is a tool to validate the conformance of EPUB publications against the EPUB specifications. Right now, Epubcheck is raising errors such as:
ERROR(RSC-005): Emma.epub/OPS/c18_Emma_Volume_1_Chapter_18.xhtml(55,122): Error while parsing file: Duplicate ID '-'
We can find similar errors from Nu Html Checker.
These errors are occurring because the page numbers defined in ProofreadPage pagelists are commonly being used as element IDs in MediaWiki:Proofreadpage pagenum template:
- en: <span class="pagenum ws-pagenum" id="{{{num}}}" data-page-number="{{{num}}}" title="{{urlencode:{{{page}}}|WIKI}}">​</span>
- fr: <span class="pagenum ws-pagenum" id="{{{num}}}" title="{{{page}}}"></span>
- eu: <span class="pagenum" id="{{{num}}}" title="{{{page}}}"></span>
- bn: <span class="pagenum ws-pagenum" id="{{{num}}}" data-page-number="{{{num}}}" title="{{urlencode:{{{page}}}|WIKI}}"></span>
This works really well for actual page numbers, because they can then be used to locate individual pages by URL fragment, e.g. page 3 in https://en.wikisource.org/wiki/Emma/Volume_1/Chapter_1#3 but when a page is given any repeating string (such as, commonly, a hyphen for pages that are outside of the pagination).
One fix might be to pass a new id parameter to this template, that's ensured to be unique (within separate <pages /> invocations only — would that be sufficient?).
(The example error above should actually be avoided by not transcluding the hyphen pages, because they're blank anyway, but this issue is still valid I think in that we should not be producing duplicate IDs.)
Acceptance Criteria:
- Fix the Epubcheck bug, as described above, so such errors no longer occur.
- Look into whether the following idea fixes the issue: Pass a new id parameter to this template, that's ensured to be unique (within separate <pages /> invocations only).