[poppler] Why poppler, which supports tagged PDFs, doesn't recognize some of the tags as a whole?

[Albert Astals Cid]

El dijous, 24 de juny de 2021, a les 15:47:38 (CEST), Germán Poo-Caamaño va escriure:
> On Thu, 2021-06-24 at 10:44 +0200, Albert Astals Cid wrote:
> > El dijous, 24 de juny de 2021, a les 7:48:45 (CEST), Denis Bitouzé va
> > escriure:
> > > Hi,
> > > 
> > > the joined `test.pdf` file is properly tagged as you can check it
> > > by
> > > loading it at:
> > > 
> > >   ┌────
> > >   │ https://www.ngpdf.com/loadFile
> > >   └────
> > > 
> > > and then looking at:
> > > 
> > >   ┌────
> > >   │ https://www.ngpdf.com/editor/editFile
> > >   └────
> > > 
> > > You can see each line of the code:
> > > 
> > >   ┌────
> > >   │ \pdfdict_new:n   {l_my_action_dict}
> > >   │ \pdfdict_put:nnn {l_my_action_dict}{Type}{/Action}
> > >   │ \pdfdict_put:nnn {l_my_action_dict}{S}{/URI}
> > >   │ \pdfdict_put:nnn {l_my_action_dict}{URI}{(  
> > > https://www.latex-project.org)}
> > >   └────
> > > 
> > > is a single tag.
> > > 
> > > Nevertheless this code, if copied e.g. from Evince 3.38.1, is
> > > pasted not
> > > as it is and but as:
> > 
> > That would be a question for the Evince developers (some of them are
> > on this is i guess so you may still get an answer).
> > 
> > The fact that poppler has facilities to "see" the contents of tagged
> > pdf doesn't mean that evince is using them.
> 
> I am unsure what the report or question is about. Is it about
> presenting/seeing each tag separately or copying/pasting the test in
> the tags?
> 
> If the later, that corresponds to poppler-glib.

You think so?

I'm not convinced i mean poppler-glib already exports the Tags contents, it's just that evince always uses the "raw text" contents and never the "Tagged text" contents.

Or you would expect poppler-glib to chose which text to give you?

Cheers,
  Albert

> 
>