On Sunday, January 21, 2018 at 5:42:09 AM UTC-5, Komiya Takeshi wrote: > > AFAIK, Sphinx goes to resolve xrefs independently for other documents. > And it dumps the result after processing.
I don't think that's relevant. My proposal was to copy the *unresolved* nodes into the TOC and resolve them later. Sure, it's duplicating work, but going back and fixing up the TOC is probably much harder; too much so to justify the extra effort. > For this reason, we can never access the resolved xrefs whole of the > project. > So I feel introducing xtext might be difficult. It turns out the tricky part is getting the TOC to resolve them. Normally, sections get turned into a TOC that gets shoved aside and is not subject to the usual transforms. However, that's fixable. > It might be better to implement better astext() instead. > I don't see how this can work. Leaving aside that this would be a fundamental change to the bowels of *docutils* (which already makes the change questionable, because *docutils* has no notion of external references), it can't solve the problem. The TOC *must* store the unresolved xref, because it gets built on reading the document, and the xref can't possibly be resolved at that point. I have a working (read: appears to WFM) patch at https://github.com/sphinx-doc/sphinx/pull/4475. -- Matthew -- You received this message because you are subscribed to the Google Groups "sphinx-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to sphinx-users+unsubscr...@googlegroups.com. To post to this group, send email to sphinx-users@googlegroups.com. Visit this group at https://groups.google.com/group/sphinx-users. For more options, visit https://groups.google.com/d/optout.