I suspect this is not an XML/HTML entity problem, but a font-related issue. Maybe the fonts bundled with Tor Browser simply don't contain the relevant glyphs.
This error is also causing problems on our Trac page right now: #18860 (moved) In addition to Linux TBB not supporting the "reply" button on the right side of every comment in https://trac.torproject.org/projects/tor/report I'm also not able to view the "bold" "italic" "link" or any other icons above the box that I'm typing in right now. I'm assuming that this is the same for all Linux TBB users that are using this site right now. Tails closed the ticket that I opened with them, saying that this seems to be squarely a Linux TBB issue. https://labs.riseup.net/code/issues/12154
I'm curious which way the TBB community is looking to go from here.
Does a new font need to be packaged for Linux TBB that renders Dingbats / Glyphs / Older Unicode?
Does it already have a font that renders them properly, but there isn't proper font fallbacks in place?
Are the web and app developers, including uBlock and Trac (see child tickets) at fault for including these things in their code? Is it a security vulnerability to render old unicode images?
If the third one is true, we should develop a document that explains the secure way to use them or ways to replace them with something else.
Can someone please respond to this ticket? Trac.torproject.org is broken. uBlock is broken. MANY SITES on the internet are broken. We really should talk about this.
Does a new font need to be packaged for Linux TBB that renders Dingbats / Glyphs / Older Unicode?
Well. None of the bundled fonts include the Dingbats Unicode code block.
Bundling NotoSansSymbols-Regular.ttf (832 KiB) along with some font-config trickery would be an improvement, though I am uncertain as to how real browser developers want to handle the download/bundle space vs coverage tradeoff.
Does a new font need to be packaged for Linux TBB that renders Dingbats / Glyphs / Older Unicode?
Well. None of the bundled fonts include the Dingbats Unicode code block.
Bundling NotoSansSymbols-Regular.ttf (832 KiB) along with some font-config trickery would be an improvement, though I am uncertain as to how real browser developers want to handle the download/bundle space vs coverage tradeoff.
The coverage is vast, including this very Trac. The download/bundle space addition is small. I'm sorry, but why the frack is this issue not being given the time of day? Please let me know how I can move things forward in a healthy way.
Does a new font need to be packaged for Linux TBB that renders Dingbats / Glyphs / Older Unicode?
Well. None of the bundled fonts include the Dingbats Unicode code block.
Bundling NotoSansSymbols-Regular.ttf (832 KiB) along with some font-config trickery would be an improvement, though I am uncertain as to how real browser developers want to handle the download/bundle space vs coverage tradeoff.
The coverage is vast, including this very Trac. The download/bundle space addition is small. I'm sorry, but why the frack is this issue not being given the time of day? Please let me know how I can move things forward in a healthy way.
Witing a patch, testing it and, if all works well, attaching it to this ticket + setting the status of it to needs_review would be such a way.
Whatever you'd like to call them, many of them are broken in Tor Browser and have been for a very long time. I understand that font fingerprinting needs to be addressed in a robust way because it protects against font enumeration attacks. However, there doesn't seem to be much work being done to fix the bugs that this security mitigation technique has introduced.
As was stated over two years ago, this issue seems to cause issues on the tor project trac itself! Right now as I'm on this page, the "reply to comment" icon to the right of every comment is blank due to this bug (that is, if I'm understanding the bug correctly). https://trac.torproject.org/projects/tor/ticket/18860
A proposal has been made to improve the list of TBB font whitelist / bundled fonts by soliciting user feedback. I agree that it would be a useful project to go through each of the fonts on each platform and see if there are better fonts that could be used instead. https://trac.torproject.org/projects/tor/ticket/20842 I've posted some comments over there as well about how we could potentially move this proposal into a reality.
In the mean time, assuming such a large project would take up a lot of time and resources, my quick suggestion to hopefully fix this specific ticket is to add fonts-noto-color-emoji to the list of Google Noto fonts shipped with the GNU+Linux version of TBB. This is an official Debian package now: https://packages.debian.org/buster/fonts-noto-color-emoji and the binary is available https://github.com/googlei18n/noto-emoji/releases If it would be preferable to get this in stretch-backports as well, please let me know and I'll do my best to pursue this.
Can anyone out there add fonts-noto-color-emoji to the Tor Browser for GNU+Linux so I can test to see if it fixes the multiple unicode errors that I have been consistently seeing for the last few years while using TBB as my daily driver?
Thanks for this helpful comment. The idea of building the fonts from source is pretty interesting. Right now we are shipping the fonts as they come from Google. I opened #26302 (moved) for investigating the source code approach.
There is no need to have anything in Debian in order to make progress, but thanks for the offer trying to move things forward in case it were needed.
And, no, I don't see any blockers other than someone sitting down, writing the patch, building the bundle and testing it. Am I seeing this right, that this font alone is 7MB in size? That's quite a lot...
I'd still really love it if someone could write a patch to add fonts-noto-color-emoji to Tor Browser.
As was stated over three years ago, this issue seems to cause issues on this tor project trac itself! Right now as I'm on this page, if you are using the Linux version of Tor Browser, the "reply to comment" icon to the right of every comment is blank due to this bug (that is, if I'm understanding the bug correctly). https://trac.torproject.org/projects/tor/ticket/18860