Hi everyone! Here is my status report for October 2023.
We have finally released Tor Browser 13.0. I prepared that release, and writing its changelog took me a surprising amount of time. The changes it contains are numerous; we hope you enjoy them! It's also the first release not to have Torbutton as a backend, and I've written a blog post about that [0]. In addition to that, in the first half of the month, I've been doing last-minute fixes, such as some preferences that we missed [1], a way to allow using Tor Browser with Tor set as a proxy but without connecting to the control port [2], fixed a build problem of Android that happened only with the release channel [3], and a few more. Speaking of Android, I also helped to find a way to improve its building time when we run our reproducible builds [4].
Then, I worked a little to make the DLL blocklist file obey the portable mode [5]. We haven't merged my proposed MR because it relied on Win32 APIs, but we might use std::filesystem instead to work around certain limitations. I haven't done that in the first place because Firefox used to disable it in libc++, but it has been enabled in Firefox 116, and we might do the same to our 115 ESR and revisit the patch.
Last week, we were at the Mullvad offices for a team meeting. It was a very productive week, and we've taken a lot of notes. We've already published a first draft [6], but we might have to go back to it and do some more formatting and add any missing information.
Finally, I've started hacking on Android because we want to bring the connection assist there as well. My dream is to unify our code base and handle both desktop and Android. It helps feature parity, it reduces the maintenance time, and in case of bugs, we can fix them on all our platforms at once. I'm not sure we'll do it for the front end eventually (there have been some discussions about that [7]), but even doing it only for the back end would be a great help. I've already managed to hook up a Tor daemon to the TorProvider I worked on last Summer and also to plumb the creation of a lyrebird instance for the domain fronting proxy. So, I'd expect it to be feasible, even though it'll take a while to create a polished patch.
I plan to keep working on Android a little bit. I've done the first exploration because I implemented the desktop part, so I know very well, but I might pass it to someone else. Then, I might go back to Mullvad Browser packaging changes.
Best, Pier
[0] https://blog.torproject.org/torbutton-retired/ [1] https://gitlab.torproject.org/tpo/applications/tor-browser/-/issues/41496 [2] https://gitlab.torproject.org/tpo/applications/tor-browser/-/issues/42160 [3] https://gitlab.torproject.org/tpo/applications/tor-browser-build/-/issues/40... [4] https://gitlab.torproject.org/tpo/applications/rbm/-/issues/40062 [5] https://gitlab.torproject.org/tpo/applications/tor-browser/-/issues/42163 [6] https://gitlab.torproject.org/tpo/applications/team/-/wikis/Goteburg-2023-Me... [7] https://gitlab.torproject.org/tpo/applications/tor-browser/-/issues/41188