Adapted for latest rand APIs

.github/workflows/ci.yaml

@@ -7,12 +7,15 @@ on:
     tags:
       - "v*"
   pull_request:
+  workflow_dispatch:  # needed to be in 'main', for any branch to manually 'Run Workflow' in GitHub UI.
 
 jobs:
   build:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
+      - name: Set up submodules
+        run: git submodule update --init --depth 1 --recursive
       - name: Free disk space
         run: |
           .ci/free_disk_space.sh
@@ -64,6 +67,9 @@ jobs:
         uses: actions/checkout@v3
         with:
           fetch-depth: '0'
+      - name: Set up submodules
+        run: |
+          git submodule update --init --depth 1 --recursive
       - name: Build new binary
         working-directory: ./examples/nrf52
         run: |
@@ -74,6 +80,9 @@ jobs:
       - name: Save binary
         run: |
           mv ./examples/nrf52/target/thumbv7em-none-eabihf/release/ble_bas_peripheral ./examples/nrf52/target/thumbv7em-none-eabihf/release/original.elf
+      - name: "[PR] transitioning from 'nrf-sdc' to 'nrf52' needs this." #hack
+        if: ${{ github.event.pull_request }}
+        run: git restore examples/nrf52/Cargo.lock
       - name: If it's a PR checkout the base commit
         if: ${{ github.event.pull_request }}
         run: git checkout ${{ github.event.pull_request.base.sha }}

.github/workflows/tests.yaml

@@ -56,6 +56,9 @@ jobs:
         run: |
           gh pr checkout "${{ github.event.inputs.prNr }}"
           echo "commit=$(git rev-parse --verify HEAD)" >> "$GITHUB_OUTPUT"
+      - name: Set up submodules
+        run: |
+          git submodule update --init --depth 1 --recursive
       - name: Set pending
         env:
           COMMIT: ${{ steps.checkout.outputs.commit }}
@@ -102,6 +105,9 @@ jobs:
         run: |
           gh pr checkout "${{ github.event.inputs.prNr }}"
           echo "commit=$(git rev-parse --verify HEAD)" >> "$GITHUB_OUTPUT"
+      - name: Set up submodules
+        run: |
+          git submodule update --init --depth 1 --recursive
       - name: Set pending
         env:
           COMMIT: ${{ steps.checkout.outputs.commit }}

.gitmodules

@@ -0,0 +1,4 @@
+[submodule "sub/nrf-sdc"]
+	path = sub/nrf-sdc
+	url = https://github.com/alexmoon/nrf-sdc
+	shallow = true

IDEAS-RC.md

@@ -0,0 +1,20 @@
+# Ideas (regarding `rand` RC)
+
+Changes in the security dependencies may open ways to make the `trouble` code base simpler. These are listed here so they won't get forgotten, and can be introduced as secondary PR's, after `rand` 0.10 support is in.
+
+
+## No double-Chacha'ing RNG seeds
+
+The platform specific (hardware) RNG's could be used as-is, for providing a seed to the `trouble` RNG. Currently, at least with nRF52, the hardware RNG generates a seed, used for creating a `ChaCha` RNG, which again generates a seed, for another. This must be an unnecessary hoop, right?
+
+Looking at these, the author noticed that actually all we'd need in getting things started is a *seed*, not an RNG. Doing this would further simplify the examples, most of all removing the need for `rand_core` dependency in them (rand stuff would remain as an internal dependency of `trouble`). Seems like a good direction (could already be brought in by this PR).
+
+
+## Use of `getrandom`?
+
+Changes in the APIs have made cryptographically strong RNG generation, without passing the `TRng` type around, possible.
+
+Is this something we'd like to experiment with?
+
+The benefit is mainly removing a bunch of generic parameters, especially in tests. As cons, test cases would need to be changed, since a new kind of RNG would be in use.
+

KNOWN-RC.md

@@ -0,0 +1,12 @@
+# Known issues (in RC libraries)
+
+## `p256` and/or `elliptic-curve` does not build
+
+See [https://github.com/RustCrypto/elliptic-curves/issues/1590](https://github.com/RustCrypto/elliptic-curves/issues/1590).
+
+## ESP32 security samples do not run
+
+The `esp-hal` library supports `rand_core` 0.6 and 0.9. We'd need it to support 0.10.
+
+Adding this to `esp-hal` wouldn't be a big effort. Can be done as a PR but likely it'll just get there anyhow, once `rand_core` 0.10 proceeds down the RC aile.
+

README.md

@@ -44,6 +44,23 @@ See the [documentation](https://embassy.dev/trouble) and the [rustdoc](https://d
 Trouble is guaranteed to compile on stable Rust 1.80 and up. It *might*
 compile with older versions but that may change in any new patch release.
 
+## Preparations
+
+>The below concerns you only, if you plan to build:
+>
+>- `benchmarks`
+>- `examples/nrf52`
+>- `examples/nrf54`
+
+Run this command:
+
+```
+./prep.sh
+```
+
+The `nrf-sdc` dependency is brought in as a submodule (instead of a git dependency). This allows us to have a shallow clone, saving lots (700MB!) of disk space.
+
+
 ## Examples
 
 See `examples` for example applications for different BLE controllers.