more resilient subsystems (#1908)

* backing: extract log target

* bitfield-signing: extract log target

* utils: fix a typo

* provisioner: extract log target

* candidate selection: remove unused error variant

* bitfield-distribution: change the return type of run

* pov-distribution: extract log target

* collator-protocol: simplify runtime request

* collation-generation: do not exit early on error

* collation-generation: do not exit on double init

* collator-protocol: do not exit on errors and rename LOG_TARGET

* collator-protocol: a workaround for ununused imports warning

* Update node/network/bitfield-distribution/src/lib.rs

* collation-generation: elevate warn! to error!

* collator-protocol: fix imports

* post merge fix

* fix compilation

polkadot/node/core/provisioner/src/lib.rs

@@ -45,6 +45,8 @@ use std::{convert::TryFrom, pin::Pin};
 use std::collections::BTreeMap;
 use thiserror::Error;
 
+const LOG_TARGET: &str = "provisioner";
+
 struct ProvisioningJob {
 	relay_parent: Hash,
 	sender: mpsc::Sender<FromJob>,
@@ -203,7 +205,7 @@ impl ProvisioningJob {
 					)
 					.await
 					{
-						log::warn!(target: "provisioner", "failed to assemble or send inherent data: {:?}", err);
+						log::warn!(target: LOG_TARGET, "failed to assemble or send inherent data: {:?}", err);
 						self.metrics.on_inherent_data_request(Err(()));
 					} else {
 						self.metrics.on_inherent_data_request(Ok(()));
@@ -459,7 +461,8 @@ fn bitfields_indicate_availability(
 				// however, in practice, that would just push off an error-handling routine which would look a whole lot like this one.
 				// simpler to just handle the error internally here.
 				log::warn!(
-					target: "provisioner", "attempted to set a transverse bit at idx {} which is greater than bitfield size {}",
+					target: LOG_TARGET,
+					"attempted to set a transverse bit at idx {} which is greater than bitfield size {}",
 					validator_idx,
 					availability_len,
 				);