poka authored 4 days ago and Nicolas80 committed 4 days ago

* * Updated CHANGELOG.md

* * Updated CHANGELOG.md with `vault inspect` changes

* * Added extra data in `vault inspect` command
  * Secret seed/mini-secret (if no soft derivation is used)
  * Public key (hex)
  * SS58 Address
  * (potential G1v1 public key if inspecting a <Base> account with ed25519 crypto-scheme)
* Does most changes mentioned in Issue #28 (except the network id)

* * Small change to be in line with the adapted parameter descriptions

* * Refactored method get_keypair_from_secret_provider into SecretProvider internal method get_keypair
* Added long_about for `LinkAccount` and `ChangeOwnerKey` to explain the target identity/account is provided with the options of those commands.
* Applied cargo fmt

* * Renamed AccountLinkParams to SecretProvider and adapted descriptions to "target account" instead of "account to link" which makes it a bit more generic
* Added generic method get_keypair_from_secret_provider that returns the keypair or an error
* Adapted `LinkAccount` & `ChangeOwnerKey` to use those
* Not sure the best place for the SecretProvider and new method - leaving in identity for now

* * Small refactoring to move method to vault.rs

* * Small change to show the linked `<Base>` account if different from the one we inspect.

* * Updated CHANGELOG.md

* Code review:

* Adapted commands identity `LinkAccount` & `ChangeOwnerKey`
  * Added possibility to provide
    * `-a` address of vault account to link
    * `-v` name of vault account to link
  * `-S` (secret_format) doesn't have a default anymore since it conflicts with using `-a` or `-v`
  * Added an error message if none of the params are provided

* * Small clippy cleanup

* * Changed sorting of children account so that they are sorted by Path value (makes more sense when viewing derivations of one account)
* Added display of "Crypto scheme" when making `vault inspect` so that it is more coherent

* Code review:

* Cleanup of keys.rs get_keypair method since it always gets a crypto_scheme.
  * Removed "default" values in the method
  * Relying on "default" value of `-c` / `--crypto-scheme` argument (present in several places) which is ed25519

* Code review:

* Small comment cleanup
* Code cleanup removed unnecessary empty cells in "g1v1" row
* Added doc on prompt_secret_substrate_and_compute_keypair method to explain the values returned in the tuple.

* * Updated version to 0.4.1
* Added Changelog entry for the new version
* Adapted some of the examples in config.md to take into account the new `-c` argument
** Did a simple addition of "-c sr25519" so that current examples still make sense and link to the same SS58 addresses

* Added extra message when the DB parsing of DbAccountId fails, so we know for which string it failed.

* * reverted most of the commit e9cd6a99: "Add secret format in database and display"

* * Adapted catch_known of predefined derivations to properly handle both sr25519 and ed25519 versions of those keys.

* * Added some logic between arguments of `vault import`; can't provide both `password` and `no-password`
** Added extra validation of non-interactive `name` argument value (same validation as when interactive: no '<', '>', '/' characters)
* Added possibility to make a non-interactive derivation (given proper arguments are given and there is no issue found during the process)
** Added same validation for non-interactive `derivation_path` argument as when interactive
** If the same resulting address is already in the vault; interaction is still mandatory to make a choice
* Had to change the name of argument AddressOrVaultNameGroup.name => vault_name to avoid conflict in `vault derive`
** Not changing the `-v` shortcut so no impact on existing commands
* Allowing to pass "" empty string as non-interactive `name` argument and considering it as None (does a trim before checking empty; so only spaces will be considered as None as well)

* * Adding `-c` crypto scheme parameter (with default ed25519) in different places
** as global parameter (taken into account if `-S` parameter is given
** as extra parameter for identity.rs commands LinkAccount & ChangeOwnerKey
*** Also added a simple display of "target" address and crypto-scheme when performing those commands

* * Fix for `vault migrate` that should use sr25519 scheme to be consistent

* * Small fix in doc of function
* Applied `cargo fmt`

* add non interactive mode

* remove unused methods

* Add secret format in database and display

* apply nico review

* feat: Can choose between ed25519 ans sr25519

Adapt linux build now depending on libssl-dev. Introduce tests job (#45, #46) (!42)

* Adapt linux build now depending on libssl-dev (#45)

native-tls transitive dependency introduced in !41
which requires libssl-dev package installed
for the build to find ssl components

Switching to non-slim Debian image which comes
with libssl-dev pre-installed

Introduce linux tests job: to run tests (#46)
and prevent similar issue in the future

Update changelog

* Adapted vault display to have a space before address value (for derivations); which fixes the incorrect selection when double-clicking an address in the terminal

* Added a CHANGELOG.md file
* Adapted config.md example of vault usage
* Adapted some of the command descriptions to be more consistent

* Re-did the mapping of DbAccountId so that we don't have to rely on derive macro FromJsonQueryResult that resulted in using JSON_TEXT in database column instead of VARCHAR
   * Tricky part I missed the first time was in impl of sea_orm::TryGetable::try_get_by where we should return Err TryGetError::Null in case the value is not present
   * Was otherwise throwing an exception whenever `parent` field (Option<DbAccountId>) was None when trying to persist the case in DB.

   * Now checking if GcliError Input message was for doing "ctrl+c" or "Esc" key in a prompt; in wich case we terminate the execution with the message.
   * Small correction in vault_account::compute_suri_account_tree_node where it's better to return GcliError instead of panic using unwrap if the password was incorrect.
* Renamed `clap` arguments "value_name" in different places where it wasn't clear what kind of value was expected.
   * Using "ADDRESS" everytime we expect an AccountId
   * Using "USERNAME" everytime we expect a Username (most were requesting "TARGET")
* Small corrections in `clap` help messages for `vault derive`
* Added print of selected Vault item when using vault::try_fetch_key_pair method; just before requesting it's associated password (whenever we will try to sign for an action mostly)

* Added support to show g1v1 base58 public key from (ed25519) ss58 address, AccountId, ed25519::Public with tests for those
* Renamed SecretFormat.Cesium to SecretFormat.G1v1
   * Adapted `clap` display help to reflect the change
   * Still supporting using "cesium" as "-S" secret format argument and mapping it to SecretFormat.G1v1
* Adapted display of vault
   * Changed column "Format" into "Crypto", showing either "ed25519" or "sr25519" instead of "cesium" or "substrate"
   * Adapted 1st row header to "SS58 Address/G1v1 public key"
   * For "ed25519" account, showing a 2nd row just below containing "└ G1v1: <base58 public key value>"
   * Adapted test cases and added a G1v1 account example
* Adapted (String) Display of vault_account to include "g1v1_pub_key" value when crypto_scheme is Ed25519

Co-authored-by: Hugo Trentesaux <hugo.trentesaux@lilo.org>

Co-authored-by: Hugo Trentesaux <hugo.trentesaux@lilo.org>

Co-authored-by: Hugo Trentesaux <hugo.trentesaux@lilo.org>

   * Put back simple `address` and `name` arguments in main.rs (with conflicts_with statement to prevent their usage at the same time)
   * Declared conflicts between indexer <=> no-indexer, indexer <=> network, url <=> network
   * Kept AddressOrVaultNameGroup args for vault usage since it's the only way to define one required arg within a group cleanly

* Relaxed "update" restrictions in vault_account.rs before_save checks

* Started to investigate being able to make subkey command and trying to use sp_core::crypto::CryptoType instead of custom KeyPair
* Big refactoring to have all in one vault_account db table; using recursion through `parent` property
* Added AccountTreeNode object and related methods to be able to more easily handle whole tree of accounts
* Removed now unused vault_derivation
* Replaced remaining usages of `rpassword` and removed the dependency
* Added `indoc` dependency to support cleaner display of multi-line strings in tests

Small cleanup for data's db connection; adding method connect_db that returns a &DatabaseConnection (or panics with explanation message)

Also changed some methods visibility so only connect_db is visible for usage.

Made an extra check of derivation path to make sure it's not linking to the same SS58 Address as "root" suri when doing `vault import|migrate`

* Cleaned-up code in several places
* Now retrieving secret value in "substrate uri" format from the different prompt_xxx_and_compute_keypair methods

* Simply using "name" for the name given to an SS58 Address in the vault

* Added a test to verify we don't need the nacl::sign::Keypair to handle cesium v1 (we can keep using scrypt to retrieve the seed and then sp_core::ed25519::Pair)
** This means we could remove some code linked to that and possibly the `nacl` dependency.

* Fixed small bug after switching from `dialoguer` to `inquire`: using `prompt_skippable()` only returns `None` if the user presses `Esc` (returned "" when user just pressed enter)
* Now using `prompt()` and checking empty string again.

* Adapted to have more coherent commands descriptions and added `long_about` more detailed description for `import` and `derivation`
* Added the `vault list for` command
* Added display of "format" of <Root> derivations in all `vault list` commands
* Only allowing derivations for format "substrate" and "seed"

* Had to bump up the dependency since version 0.6.2 had a bug for the password confirmation - see https://github.com/mikaelmello/inquire/issues/149
* Removed `dialoguer` dependency
* We may want to replace all `rpassword` usage with `inquire` in the future

* Added "/.idea" exclusion in .gitignore (for when using JetBrains IDEs)
* Added dialoguer dependency for easier user input handling (see in inputs.rs)
* Added sea-orm dependency to allow having DB entity mappings and use a local sqlite file database
* Added rstest test dependency for parameterized tests support
* Added derivation tests for each SecretFormat (including cesium v1 key derivation, using sp_core::ed25519::Pair)
* Made a lot of changes to add vault_account and vault_derivation db tables to persist vault keys & derivations
* Added support for KeyPair::Ed25519 linking to sp_core::ed25519::Pair which can be created from secret seed retrieved from nacl::sign::Keypair (which is created from cesium id + secret)
** This was necessary to allow deriving keys from "cesium v1" keys (to be reviewed - it might be a bad idea to permit that from a security point of view)
* Only kept original (substrate) keyfiles support for migration (use "vault list-files" and "vault migrate")
* Added possibility to give either "-a" Address or "-v" Vault Name as general option
* Added extra commands in Vault
** list-files: (deprecated)List available key files (needs to be migrated with command "vault migrate" in order to use them)
** migrate: (deprecated)Migrate old key files into db (will have to provide password for each key)
** 'list' now has sub-commands 'all' or 'root' to show all keys or only root keys (without derivation path)
** use: "Use specific vault key (changes the config address)", which will have the same behaviour as `gcli <-a <Address>|-v <VaultName>> config save` (left a FIXME in there to review)
** derivation: Add a derivation to an existing (root) vault key
** rename: Give a meaningful vault name to a vault key or derivation
** remove: Remove a vault key (and potential derivations if it's a root key)
* Had to bubble up "await" and "async" in a lot of places
* ...

Nicolas80 authored 4 months ago and Hugo Trentesaux committed 4 months ago

* Fixing precision issue when converting the u64 cents amount into f32.
Now composing the displayed value with integer part and fractional part separately (keeping left padding of 0 for the decimal part for the amount of decimals required)

to speed up the process
The runner is not overloaded with one build

This reverts commit 9460516b.

Does not solve the issue, the links do not work anymore:
They need `?job=$job_name` at the end in which the artifacts have been created
https://git.duniter.org/clients/rust/gcli-v2s/-/releases/0.3.1

matograine authored 5 months ago and Moul committed 5 months ago

* CI : remove GET params at the end of assets URLs

on runtime 802
changed crypto types