migrate setup_environment.sh to CMake

[brunoerg]

Fixes #750

[brunoerg]

Thanks, @hebasto. Addressed #752 (comment). The solution is simpler than I thought.

[hebasto]

Not related to the CMake migration, but I'd suggest additionally this diff:

--- a/test/setup_environment.sh
+++ b/test/setup_environment.sh
@@ -444,7 +444,7 @@ if [[ -n ${build_bitcoind} ]]; then
 
     # Build bitcoind. This is super slow, but it is cached so it runs fairly quickly.
     pushd depends
-    make NO_QT=1 NO_QR=1 NO_ZMQ=1 NO_UPNP=1 NO_NATPMP=1
+    make NO_QT=1 NO_QR=1 NO_ZMQ=1 NO_UPNP=1 NO_NATPMP=1 NO_USDT=1
     popd
 
     # Do the build

to skip needless build of the systemtap package in depends.

[hebasto]

One more amendment is needed:

--- a/.github/actions/build-bitcoind/action.yml
+++ b/.github/actions/build-bitcoind/action.yml
@@ -23,7 +23,7 @@ runs:
         ccache --zero-stats
         cd test; ./setup_environment.sh --bitcoind; cd ..
         ccache --show-stats --verbose
-        tar -czf bitcoind.tar.gz test/work/bitcoin/src/bitcoind
+        tar -czf bitcoind.tar.gz test/work/bitcoin/build/src/bitcoind
 
     - uses: actions/cache/save@v4
       if: github.event_name != 'pull_request' && steps.ccache-cache.outputs.cache-hit != 'true'

[brunoerg]

Addressed #752 (comment) and #752 (comment). Thanks, @hebasto. Just added you as co-author.

[hebasto]

ACK 37d520d.

oops, another path issue (

[hebasto]

Are we interested in engaging in parallelism:

Suggested change

	cmake --build build --target bitcoind
	cmake --build build -j $(nproc) --target bitcoind

?

If so, the same applies to building depends.

[brunoerg]

I didn't address it because we did not have it previously, but I don't see why not engage parallelism.

[brunoerg]

Addressed it in #755. I don't see why not parallelize it.

[hebasto]

--- a/test/run_tests.py
+++ b/test/run_tests.py
@@ -60,7 +60,7 @@ parser.add_argument('--ledger-path', dest='ledger_path', help='Path to Ledger em
 parser.add_argument('--jade-path', dest='jade_path', help='Path to Jade qemu emulator', default='work/jade/simulator')
 
 parser.add_argument('--all', help='Run tests on all existing simulators', default=False, action='store_true')
-parser.add_argument('--bitcoind', help='Path to bitcoind', default='work/bitcoin/src/bitcoind')
+parser.add_argument('--bitcoind', help='Path to bitcoind', default='work/bitcoin/build/src/bitcoind')
 parser.add_argument('--interface', help='Which interface to send commands over', choices=['library', 'cli', 'bindist', 'stdin'], default='library')
 
 parser.add_argument("--device-only", help="Only run device tests", action="store_true")

[brunoerg]

Addressed #752 (comment)

[hebasto]

ACK 937c1ae.

[hebasto]

Not related to this PR changes, but it seems Coldcard/firmware@65b652a has to be considered to fix the CI.

UPD. I suggest to leave it for a follow up PR.

[brunoerg]

UPD. I suggest to leave it for a follow up PR.

Just opened #754 as draft.

[achow101]

ACK 937c1ae