bug#36171: 25.1; Emacs displays Hebrew text incorrectly

[Eli Zaretskii]

> Cc: 36171@debbugs.gnu.org
> From: Raoul Comninos <revrari@mweb.co.za>
> Date: Wed, 12 Jun 2019 21:27:10 +0200
> 
> To be clear, the problem seems to be with how Emacs reads 
> these particular documents. If I type the Hebrew myself directly into 
> Emacs there are no problems with the display. These documents were 
> originally created in other word processors: Word, Nota Bene , and 
> Google Docs. They were converted then to plain text (utf-8) which is the 
> format they are now in: plain text.

I wasn't aware that the problem disappears if you type Hebrew, and
appears only with these documents.  If so, can you send one such
document as an attachment?

Am I to understand that if you type the same exact text into Emacs,
there's no problem with the display?  If so, the reason must be in the
documents themselves: something in how the text is laid out there
causes this.  Being able to look at such a document is therefore
important for investigating this issue.

> I am currently trying the latest Emacs from git. I have also installed 
> gir1.2-harfbuzz-0.0. Is that the correct program?

Emacs needs HarfBuzz the library, not the program.  What does the
following produce inside the latest Emacs you built from Git:

  M-: (frame-parameter nil 'font-backend) RET