-
Notifications
You must be signed in to change notification settings - Fork 883
[keygen] Remove deprecated functions from keygen cli (minus grind command)
#438
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[keygen] Remove deprecated functions from keygen cli (minus grind command)
#438
Conversation
keygen/src/keygen.rs
Outdated
| let source = if matches.try_contains_id("keypair")? { | ||
| Cow::Borrowed(matches.get_one::<SignerSource>("keypair").unwrap()) | ||
| } else if !config.keypair_path.is_empty() { | ||
| &config.keypair_path | ||
| Cow::Owned(SignerSource::parse(&config.keypair_path)?) | ||
| } else { | ||
| let mut path = dirs_next::home_dir().expect("home directory"); | ||
| path.extend([".config", "solana", "id.json"]); | ||
| path.to_str().unwrap() | ||
| Cow::Owned(SignerSource::parse(path.to_str().unwrap())?) | ||
| }; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not obvious to me: why is the Cow wrapping necessary? If this will be a common need, it's unfortunate ergonomics.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, this is unfortunate ergonomics. The issue is that matches.get_one::<SignerSource>("keypair").unwrap() gives a reference &SignerSource while SignerSource::parse(...).unwrap() gives a fully owned SignerSource. Returning &SignerSource::parse(...).unwrap() is not possible since the owned type will be freed out of scope.
The alternative would be to define an owned variable outside of the if statement, set it to SignerSource::parse(...).unwrap(), and return a reference to it. I thought this was less elegant than using Cow at first, but it seems like it could be cleaner, so I made the change. Please take a look.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks, I agree the current version is a little cleaner.
It's actually odd to me that ArgMatches::get_one returns a reference instead of an owned object. I suppose I should read the clap code to see why.
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #438 +/- ##
=========================================
- Coverage 81.9% 81.8% -0.1%
=========================================
Files 840 841 +1
Lines 228107 228265 +158
=========================================
+ Hits 186851 186931 +80
- Misses 41256 41334 +78 |
This change includes all of the following changes as well: --- Author: kirill lykov <[email protected]> Date: Tue May 28 20:20:39 2024 +0200 Add transaction bench client (anza-xyz#339) New client application that sends transactions using TPU protocol. First, this client creates accounts necessary for the transaction generation. Second, it generates transactions executing specified program and sends them to the upcoming leaders. Contrary to other clients, it doesn't use the `TpuClient` and the `ConnectionCache` to interact with validators, but uses an alternative client network layer implementation. --- Author: Alex Peng <[email protected]> Date: Tue Jun 4 02:31:16 2024 -0500 transaction-bench: fix underflow in ensure_authority_balance (anza-xyz#341) --- Author: kirill lykov <[email protected]> Date: Fri Jul 26 10:02:39 2024 +0200 add 3 tests for client network component (anza-xyz#365) This PR is the first in a series of PR introducing integration tests for client network component. It introduces the following changes: * add first basic integration test * add test_connection_denied_until_allowed * add test for connection pruning * retry connecting when pruned * Introduce IoErrorWithPartialEq to check that QuicError is as expected --- Author: kirill lykov <[email protected]> Date: Mon Aug 5 18:30:09 2024 +0200 Handle error with connection properly (anza-xyz#383) When one of the connections fails to open, we stop the whole client. It doesn't work on testnet because there most of the nodes drop TPU traffic despite of saying through gossip that it is allowed (need to ping foundation on that). --- Author: kirill lykov <[email protected]> Date: Thu Aug 8 22:25:54 2024 +0200 Read/write for AccountsFile to use on testnet (anza-xyz#381) Only the `transaction-bench/src/network/` and `transaction-bench/tests` parts of this change was merge in. Illia Bobyr: Migrated to quinn 0.11 and refactored the tests a bit. --- Author: kirill lykov <[email protected]> Date: Sat Aug 17 15:10:19 2024 +0200 Handle instruction error in client-test-program (anza-xyz#411) Due to upstream changes errors are not propagated properly anymore when calling try_from_slice. This PR fixes the problem exactly the same way as in anza-xyz#387 --- Author: kirill lykov <[email protected]> Date: Tue Aug 20 10:01:36 2024 +0200 Add structure to collect statistics about send txs erros (anza-xyz#367) There is no way to check why certain packets have not been delivered. By propagating this information to the caller, we can use it not only in the client code but also in tests. This change Introduces a structure which accumulates relevant counters per node IP and returns this information to the client code. Illia Bobyr: Migrated to quinn 0.11 and refactored a bit. Refactored-by: Illia Bobyr <[email protected]> --- Author: kirill lykov <[email protected]> Date: Mon Aug 26 18:37:07 2024 +0200 Add rate limiting test to client (anza-xyz#368) This PR introduces integration test to check that when rate-limiting has happen we increment correct counters on client side. --- Author: kirill lykov <[email protected]> Date: Tue Aug 27 16:34:57 2024 +0200 Improve stream sending (anza-xyz#372) * stop using multistream connection * add throttling test * remove stream's finish call. It will be called when stream is dropped anyways, but if we call finish explicitly we will wait for ACK, so the behavior will be like if we had RW=1 tx. This change makes a x1000 difference on a connection with high latency. Illia Bobyr: Migrated to quinn 0.11, added cancellation in order to support clean shutdown, and refactored the tests a bit. Refactored-by: Illia Bobyr <[email protected]> --- Author: kirill lykov <[email protected]> Date: Thu Sep 5 18:54:01 2024 +0200 use testing_utilities (anza-xyz#427) We use modified copy-pasted code from streamer to setup server/client in integration tests for network component. Later original code was updated to be reusable. This PR removes old copy-pasted code and uses modified streamer's testing_utilities code. --- Author: kirill lykov <[email protected]> Date: Tue Sep 10 12:06:08 2024 +0200 Move leader updater to network and add documentation (anza-xyz#426) --- Author: kirill lykov <[email protected]> Date: Tue Sep 10 16:30:34 2024 +0200 Add documentaiton to network code (anza-xyz#425) --- Author: kirill lykov <[email protected]> Date: Tue Sep 10 20:58:04 2024 +0200 Don't ignore `test_no_host` (anza-xyz#435) --- Author: kirill lykov <[email protected]> Date: Wed Sep 11 20:34:09 2024 +0200 Address Illia's comments (anza-xyz#438)
Problem
The keygen cli uses clap-v3 and solana-clap-v3-utils, but still uses deprecated functions from clap-v2.
Summary of Changes
Replaced deprecated clap-v2 functions with clap-v3 functions. I organized the changes so that each commit is associated with different commands.
This PR removes all the deprecated functions for all commands except for the
grindcommand. Thegrindcommand seems like it will need a bit more structural changes, so I will make the changes in a follow-up PR.The first and last commits here adds and removes the
deprecatedfeature from clap-v3. If this feature is set, then clap-v3 warns on the use of deprecated functions. I removed this feature since thegrindcommand is not yet updated.Fixes #