Created attachment 172173 [details] A table of content with hypertext links The links generated for the table of contents in PDF are clickable in Firefox and other browsers as expected. The issue is that the same links are not detected by screen readers as tested with Orca and NVDA. - https://help.gnome.org/users/orca/stable/ - https://www.nvaccess.org/download/ Source file used to export to PDF with Universal access option selected is attached. Under links the following options were selected: Named destinations, PDF targets, and URLs relative to file system. Also the Open with Internet browser was selected. Many thanks!
Hello, We start to suspect there may be an issue with the Orca screen reader and possibly not with the generated document. We will get back to you when we have done some more experimenting to verify. Thanks
MNo repro in Version: 7.3.0.0.alpha0+ / LibreOffice Community Build ID: adf65471e889676a600a9c6d0454c75cbd549ad3 CPU threads: 8; OS: Linux 5.4; UI render: default; VCL: gtk3 Locale: ru-RU (ru_RU.UTF-8); UI: en-US TinderBox: Linux-rpm_deb-x86_64@86-TDF, Branch:master, Time: 2021-08-13_04:44:18 Calc: threaded
There is another bug 148934 but this one is for Orca/NVDA. Mikael, please confirm you checked PDF/UA in PDF export dialog. Eleonora please confirm with which reader software you tested.
I'll look into it.
Created attachment 183249 [details] Example file exported to PDF/UA with LO 7.0 from bibisect-linux repo I exported the example file from LO master to docx on win, then the docx to PDF with Word 16. Compared the simple Writer-exported PDF in Evince, when Orca is reading. In both cases the TOC entries are read up like "1 Intro Link" "2 Content Link" by Orca. Also happens with file exported from 7.0. Only way to make it *not* work was to open the PDF in Firefox - but then, Orca (in Ubuntu 18.04) does not read anything in Firefox, not even its own help page, which is plain HTML. So the whole thing must be somehow broken there. PDF reading is a bit better in Chromium, but Link is not read by Orca, not with Writer-made PDF, nor with Word-made PDF. For plain HTML, it is read. I opened the exported file in PAC tool, and it does not complain about issues related to links either, only the missing alt descriptions (which is bug 148934). "Also the Open with Internet browser was selected." - writes the reporter. So I think this is what they experienced.
From the analysis thus far, nothing much we can do here. Closing as NOTOURBUG (i.e. an issue somewhere between browser and screenreader, not the generated PDF).