From harfbuzz at lists.freedesktop.org Tue Nov 8 17:23:45 2022 From: harfbuzz at lists.freedesktop.org (harfbuzz at lists.freedesktop.org) Date: 8 Nov 2022 11:23:45 -0600 Subject: [HarfBuzz] Your website, databases and emails has been hacked. Message-ID: <20221108112345.EFC447073B2B152A@lists.freedesktop.org> An HTML attachment was scrubbed... URL: From enometh at meer.net Wed Nov 30 09:15:50 2022 From: enometh at meer.net (Madhu) Date: Wed, 30 Nov 2022 09:15:50 -0000 Subject: [HarfBuzz] kaithi & arabic : glyphs that span number Message-ID: <20221021.172516.902473700879332390.enometh@meer.net> In April 2002, Sameer Singh posted a message on emacs-devel where he demonstrated the use of harfbuzz numerals with the kaithi script https://lists.gnu.org/archive/html/emacs-devel/2022-04/msg01297.html to yield this image: [1] https://lists.gnu.org/archive/html/emacs-devel/2022-04/png4ClBZc1zV4.png There the first glyph spans the remaining chacracters when composed. I tried to reproduce it and showed my work on this message https://lists.gnu.org/archive/html/emacs-devel/2022-05/msg00341.html but I only get [2] https://lists.gnu.org/archive/html/emacs-devel/2022-05/pngTFsflcCrfU.png The first character does not span the remaining digit characters. Can anyone tell me what I might be missing? 1. 2.txt has code points 0x110BC 0x0967 0x0966 0x096E 2. hb-view --font-file=$HOME/.fonts/NotoSansKaithi-Regular.ttf --text-file=2.txt > 2.png This produces the image in the above mentioned url [2] . I believe it should look like [1]. Similar tests for arabic numbers a text file arabic-num.txt with code points 0x0600 0x31 0x32 0x33 and a command hb-view --font-file=$HOME/.fonts/NotoSansArabic-Regular.ttf --text-file=arabic-num.txt --output-file=arabic-num.png produces a file where the first character does not span the remaining characters. I believe I used the same font-files as Sameer did: -rw-r--r-- 1 madhu users 260740 Mar 24 2020 NotoSansArabic-Regular.ttf -rw-r--r-- 1 madhu users 87260 Mar 24 2020 NotoSansKaithi-Regular.ttf ---Madhu