Multiple try-runtime-cli tests use substrate_cli_test_utils::start_node to spawn a node to run commands against.
Unfortunately, substrate_cli_test_utils::start_node depends on the substrate-node binary being accessible by cargo_bin (calls cargo_bin("substrate-node")), which is not the case outside of substrate.
In this PR I create a new method substrate_cli_test_utils::start_node_inline.
The new method functions similarly to start_node, but can be run in environments without the ability to access a node binary with cargo_bin("substrate-node").
Honestly I am not sure if this is the best approach to solving this problem, I'm open to other suggestions if there're better ones.
I am in the process of migrating
try-runtime-cli
into an independent repo (https://github.com/paritytech/try-runtime-cli/pull/6).Multiple
try-runtime-cli
tests usesubstrate_cli_test_utils::start_node
to spawn a node to run commands against.Unfortunately,
substrate_cli_test_utils::start_node
depends on thesubstrate-node
binary being accessible bycargo_bin
(callscargo_bin("substrate-node")
), which is not the case outside ofsubstrate
.In this PR I create a new method
substrate_cli_test_utils::start_node_inline
.The new method functions similarly to
start_node
, but can be run in environments without the ability to access a node binary withcargo_bin("substrate-node")
.Honestly I am not sure if this is the best approach to solving this problem, I'm open to other suggestions if there're better ones.