Bugzilla

As for the other failures:
* The gv-junit failure seems unrelated to this change, as it's on android (`minidump-analyzer` isn't on android yet).
* The bc failure is because `minidump-unwind` doesn't support fat dylibs. This was an unintentional omission.
* The xpcshell failures on test_crash_exc_guard.js looks related to the `minidump` crates giving more detail (so the test can probably be updated).
* The xpcshell crash failures look to be due to `minidump_writer`, which had to be updated when vendoring dependencies to not have duplicate crate versions.

So we need to:
1. Decide whether to dynamically load `wincrypt` functions again or simply disable support for that fallback code path on mingw.
2. Update `minidump-unwind` to support fat dylibs.
3. Update `toolkit/crashreporter/test/browser/browser_sandbox_crash.js` with the new value.
4. Potentially fix the issue in `minidump_writer` and bump the version.

As for the other failures:
* The gv-junit failure seems unrelated to this change, as it's on android (`minidump-analyzer` isn't on android yet).
* The bc failure is because `minidump-unwind` doesn't support fat dylibs. This was an unintentional omission.
* The xpcshell failures on test_crash_exc_guard.js looks related to the `minidump` crates giving more detail (so the test can probably be updated).
* The xpcshell crash failures look to be due to `minidump_writer`, which had to be updated when vendoring dependencies to not have duplicate crate versions.

So we need to:
1. Decide whether to dynamically load ~`wincrypt`~ `wintrust` functions again or simply disable support for that fallback code path on mingw.
2. Update `minidump-unwind` to support fat dylibs.
3. Update `toolkit/crashreporter/test/browser/browser_sandbox_crash.js` with the new value.
4. Potentially fix the issue in `minidump_writer` and bump the version.