Regression introduced by bce5268a21.
The bit size of the initialisation vector for AES GCM has been
introduced in NSS version 3.52 in the CK_GCM_PARMS struct via the
ulIvBits field.
Unfortunately, Firefox 68.8.0 and 76.0 do not set this field and thus it
gets initialised to zero, which in turn causes IV generation to fail.
I found out about this because WebRTC stopped working after updating to
NSS 3.52 and so I started bisecting.
Since there wasn't an obvious error in Firefox hinting towards NSS but
instead just the video stream ended up as a "null" stream, I didn't
suspect the NSS update to be the culprit at first. So I verified a few
times and then also started bisecting the actual commit in NSS that
caused the issue.
This turned out to be the problematic change:
https://phabricator.services.mozilla.com/D63241
> One notable change was caused by an inconsistancy between the spec and
> the released headers in PKCS#11 v2.40. CK_GCM_PARAMS had an extra
> field in the header that was not in the spec. OASIS considers the
> header file to be normative, so PKCS#11 v3.0 resolved the issue in
> favor of the header file definition.
Since the test I've used[1] was a bit flaky, I still didn't believe the
result of the bisect to be accurate, but after running the test several
times leading same results I dug through the above change line by line
to get more clues.
It fortunately didn't take that long to stumble upon the ulIvBits change
(which is actually documented in the NSS 3.52 release notes[4], but I
managed to blatantly ignore it for some reason) and started checking the
Firefox source tree for changes regarding that field.
Initialisation of that new field has been introduced[2] in preparation
for the 76 release, but subsequently got reverted[3] prior to the
release, because Firefox 76 is expected to be shipped with NSS 3.51,
which didn't have the ulIvBits field.
The patch I'm adding here is just a reintroduction of that change,
because we're using NSS 3.52. Not initialising that field will break
WebRTC and WebCrypto, which I think the former seems to gain in
popularity these days ;-)
Tested the change against the mentioned VM test[1] and also by testing
manually using Jitsi Meet and Nextcloud Talk.
[1]: https://github.com/aszlig/avonc/tree/884315838b6f0ebb32b/tests/talk
[2]: https://hg.mozilla.org/mozilla-central/rev/3ed30e6b6de1
[3]: https://hg.mozilla.org/mozilla-central/rev/665137da70ee
[4]: https://developer.mozilla.org/en-US/docs/Mozilla/Projects/NSS/NSS_3.52_release_notes
Signed-off-by: aszlig <aszlig@nix.build>
* The 'arm.patch' patch doesn't apply anymore.
* The 'build-arm-libopus.patch' patch isn't required anymore.
* See the mozilla phabricator link for the added patch.
Additionally, we are now *always* undconditionally applying all patches
to all architectures. That is, unless they have undesirable
side-effects, but those might not be fit for inclusion.
By applying all patches all the time, they'll be removed or replaced
when they stop applying.
with firefox 64 being the latest version, and the removal of
"tor-browser/icecat-like" variants, we can greatly simplify the common
firefox derivation.
Firefox now requires `llvm-objdump` during the build phase. The aarch64
patches do no longer apply. So far I am guessing that they have been
merged. We should verify that.
This adds support for building firefox with the gtk wayland backend. It
should work on all the flavors that use >=gtk3. Using the wayland
still allows using the X11 backend.
It works, but this state is far from ideal: GNU guys update generated source
tarballs very infrequently. Ideally, src needs to be generated by running
makeicecat over firefox src. Will do later.
Firefox >=65 will depend on icu >=63. All the older firefox versions
(and derived packages) seem to work fine with this change.
Also the system path environment patch will fail to apply since there
was a trivial whitespace change in the source file. By adding `-l` to
patch we can avoid having to track two patches that do basically the
same. Having patchFlags per file without resorting to pre-/postPatch
would be nicer but there doesn't seem to be a facility for that right
now.
Misc cleanups, but mainly this:
Before:
- `version` could mean either Firefox or TorBrowser version,
- `configureFlags` was hacky.
Now:
- `ffversion` is Firefox version, `tbversion` is TorBrowser version,
- `configureFlags` is much less hacky.
Fixes#30285
Some things done:
- Add macOS frameworks needed
- Fix RUST_BINDGEN handling. We need to pass all of NIX_CFLAGS_COMPILE
to rust bindgen
- Add custom install phase for darwin
