So it looks like we need to fix Mac and Windows, but Linux is OK.
That's actually not true. Surprisingly I have at least one Linux box where emoticons are broken. Is is running Debian testing. That said I realized that it can be a bit hard to see this if one is upgrading from 5.0.7 as it seems to work then but after shutting the browser down and restarting it no emoticons are shown. (This reminded me at one of the failure modes of our start script. :/)
So it looks like we need to fix Mac and Windows, but Linux is OK.
That's actually not true. Surprisingly I have at least one Linux box where emoticons are broken. Is is running Debian testing. That said I realized that it can be a bit hard to see this if one is upgrading from 5.0.7 as it seems to work then but after shutting the browser down and restarting it no emoticons are shown. (This reminded me at one of the failure modes of our start script. :/)
You're right, this happened to me. Once I shutdown the browser and started it again, the Emojis were no longer available.
So it looks like we need to fix Mac and Windows, but Linux is OK.
That's actually not true. Surprisingly I have at least one Linux box where emoticons are broken. Is is running Debian testing. That said I realized that it can be a bit hard to see this if one is upgrading from 5.0.7 as it seems to work then but after shutting the browser down and restarting it no emoticons are shown. (This reminded me at one of the failure modes of our start script. :/)
You're right, this happened to me. Once I shutdown the browser and started it again, the Emojis were no longer available.
So it looks like we need to fix Mac and Windows, but Linux is OK.
That's actually not true. Surprisingly I have at least one Linux box where emoticons are broken. Is is running Debian testing. That said I realized that it can be a bit hard to see this if one is upgrading from 5.0.7 as it seems to work then but after shutting the browser down and restarting it no emoticons are shown. (This reminded me at one of the failure modes of our start script. :/)
You're right, this happened to me. Once I shutdown the browser and started it again, the Emojis were no longer available.
Thanks, this is commit c885551e69ca5043d0af0de6940e250968e6f7e5, 7aa3abf90371fb82518edf5d54e862e91744c97b and ee6fd02653fc0e396bee5daa034bd00be890ccb5 on master/maint-5.5/hardened-builds.
Trac: Status: needs_review to closed Resolution: N/Ato fixed
No NotoEmoji-Regular.ttf in TBB 5.5.1 for Win, no emoji.
It appears that the Emoji font I add to the whitelist for Windows, Segoe UI Emoji, is not always available on Windows 7. So there are a couple of alternatives:
Whitelist Segoe UI Symbol instead. This covers a lot of Emoji characters, although they are monochrome.
Whitelist both. This of course allows Windows 7 and earlier to be distinguished from later versions, but there are other ways to distinguish these fonts.
Bundle NotoEmoji-Regular.ttf with Windows and remove Segoe UI Emoji from the whitelist. This makes the Emoji fonts identical on all versions of Windows. The Noto Emojis don't look the quite the same as the native Windows fonts.
I'm inclined to go for the third option. Does anyone else have an opinion?
Trac: Status: closed to reopened Resolution: fixed toN/A
No NotoEmoji-Regular.ttf in TBB 5.5.1 for Win, no emoji.
It appears that the Emoji font I add to the whitelist for Windows, Segoe UI Emoji, is not always available on Windows 7. So there are a couple of alternatives:
Windows 8 seems to be affected as well.
Whitelist Segoe UI Symbol instead. This covers a lot of Emoji characters, although they are monochrome.
Whitelist both. This of course allows Windows 7 and earlier to be distinguished from later versions, but there are other ways to distinguish these fonts.
Bundle NotoEmoji-Regular.ttf with Windows and remove Segoe UI Emoji from the whitelist. This makes the Emoji fonts identical on all versions of Windows. The Noto Emojis don't look the quite the same as the native Windows fonts.
I'm inclined to go for the third option. Does anyone else have an opinion?
Hm, this is pretty unfortunate especially as we probably want to test whether the bundling mechanism is working on Windows as well properly. But, yes, I think we could try that route. I am not really inclined to wait for that for 5.5.1, though, as #18169 (moved) and #18168 (moved) are more important and I would like to see some testing with this approach.