Investigate cross-compiling Tor Browser for Windows with clang-cl
It seems using clang-cl for cross-compiling Tor Browser for Windows gets closer and closer to be a real alternative. The Chrome bug is: https://bugs.chromium.org/p/chromium/issues/detail?id=495204 and has some information for starters.
Activity
The Google people made this work, so we have an alternative road to the mingw-w64/clang setup we could try.
Tracking LLVM master causes still issues sometimes (https://bugs.chromium.org/p/chromium/issues/detail?id=781873) but they get resolved pretty fast it seems.
Trac:
Keywords: N/A deleted, ff59-esr, TorBrowserTeam201712, GeorgKoppen201712 added
Priority: Medium to High-
The announcement and some discussion with details is on: https://groups.google.com/a/chromium.org/forum/#!msg/chromium-dev/cIA9fBb9vBE/73cOvZu9BQAJ
Replying to gk:
The Google people made this work, so we have an alternative road to the mingw-w64/clang setup we could try. from Martin: Yes, the clang/llvm based toolchain is much less mature in all aspects, so I wouldn't recommend it unless you have interest in it for another reason (my reason is support for windows on non-x86 platforms). or you're talking not about mingw-w64 actually?
-
For those following along at home: https://bugzilla.mozilla.org/show_bug.cgi?id=1390583 has all the details. So far I've been able to cross-compile our ESR52-based Tor Browser code in a similar setup to the one we use for our shipped bundles and the result is running on a Windows 8 system.
Replying to gk:
For those following along at home: ROFL :D Who are those? Reply here, folks! https://bugzilla.mozilla.org/show_bug.cgi?id=1390583 has all the details. Those, who follow, already know you spamming that Bugzilla ticket :) So far I've been able to cross-compile our ESR52-based Tor Browser code in a similar setup to the one we use for our shipped bundles It seems, x86_64-w64-mingw32/ucrt/lld/clang (not -cl ;) ) is not similar to stable win32 builds. ;) and the result is running on a Windows 8 system. Why win 8? :) (BTW, Tom ni you about debugging on win)
Replying to cypherpunks:
ROFL :D Who are those? Everyone who hopes that Stylo will make it into the next TB release so that we don't have to wait yet another year just to get some decent speed boost.
Replying to gk:
For those following along at home: https://bugzilla.mozilla.org/show_bug.cgi?id=1390583 has all the details. Why don't you file your findings under https://bugzilla.mozilla.org/show_bug.cgi?id=1257268 to allow Mozillians to work on them in parallel?
Copying myself because we're likely to want to use this clang-cl for cross-compiling libwebrtc too, in #25483 (moved).
Trac:
Cc: tom to tom, dcf-
I think this investigation is done. As I said in comment:10 https://bugzilla.mozilla.org/show_bug.cgi?id=1390583 has the details and in general cross-compiling Firefox with clang/mingw-w64 is working for our purposes. For details on how to set this toolchain up, see: https://lists.torproject.org/pipermail/tbb-dev/2018-March/000825.html.
We don't plan to switch to clang/mingw-w64 for Tor Browser 8 as there are likely some timeconsuming loose ends like reproducibility, DEP+ASLR etc. support and others. The plan is to start to work on this for Tor Browser 8.5.
This includes the missing magic to compile Stylo as well. That work in particular will be done in https://bugzilla.mozilla.org/show_bug.cgi?id=1390583.
Trac:
Status: new to closed
Resolution: N/A to fixed mentioned in issue #25304 (moved)