feat: Add deployment_from configuration and srv3 config for hoodi

.github/workflows/regression.yml

@@ -45,6 +45,10 @@ jobs:
             network: hoodi
           - config: config_samples/ethereum/hoodi/vaults/hoodi_vaults_easy_track_config.json
             network: hoodi
+          - config: config_samples/ethereum/hoodi/srv3/hoodi_srv3_config.json
+            network: hoodi
+          - config: config_samples/ethereum/hoodi/srv3/hoodi_srv3_easy_track_config.json
+            network: hoodi
     steps:
       - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6.0.3
 

CLAUDE.md

@@ -68,10 +68,12 @@ explorer_hostname:  "api.etherscan.io"
 explorer_token_env_var: "ETHERSCAN_EXPLORER_TOKEN"
 github_repo:        { url, commit, relative_root }
 dependencies:       { "dep_name": { url, commit, relative_root } }
-bytecode_comparison: { constructor_calldata, constructor_args, libraries }
+bytecode_comparison: { constructor_calldata, constructor_args, libraries, deployment_from, extra_sources }
 allowed_diffs:      { bytecode: {...}, source: {...} }  # optional; see below
 ```
 
+`bytecode_comparison.extra_sources` (`{ "0xaddr": ["path/to/File.sol"] }`, optional) names additional source files to fetch from the configured GitHub repo/commit and add to the compilation — for contracts whose explorer-verified source set omits a file the pinned GitHub source imports (e.g. a newly-added interface). The files are fetched the same way as all others (honoring `dependencies`), so verification stays honest.
+
 `allowed_diffs` declares known/expected diffs per contract so a run still passes while everything else stays verified (validated by `diffyscan/utils/allowed_diffs.py`). Each rule needs a `reason`. **Prefer the tightest facet** — bytecode: `immutables`, `byte_ranges`, `cbor_metadata`, `constructor_args`/`constructor_calldata`; source: `line_ranges`, `files`. `any: true` is a blanket wildcard that hides all future drift — avoid it unless a diff genuinely can't be scoped, and say why in the `reason`. The deprecated `--allow-source-diff`/`--allow-bytecode-diff` CLI flags are just shorthands for `any: true`. `tests/test_no_wildcard_regression.py` guards against new wildcards.
 
 ### Environment variables

README.md

@@ -89,7 +89,7 @@ diffyscan config_samples/lido_dao_sepolia_config.json
 
 When no path is given, diffyscan looks for `config.json`, `config.yaml`, or `config.yml` in the current directory. When a directory is given, all `.json`, `.yaml`, and `.yml` files inside it are processed.
 
-Alternatively, create a new config file. Configs can be written in JSON or YAML. The `bytecode_comparison` section is optional and only needed for manual overrides when explorer metadata is missing or you want to override it:
+Alternatively, create a new config file. Configs can be written in JSON or YAML. The `bytecode_comparison` section is optional and only needed for manual overrides when explorer metadata is missing or you want to override it. `deployment_from` can be used per contract when constructor simulation depends on `msg.sender`:
 
 **JSON** (`config.json`):
 
@@ -138,6 +138,9 @@ Alternatively, create a new config file. Configs can be written in JSON or YAML.
           "0xC01fC1F2787687Bc656EAc0356ba9Db6e6b7afb7"
         ]
       ]
+    },
+    "deployment_from": {
+      "0x045dd46212A178428c088573A7d102B9d89a022A": "0xE92329EC7ddB11D25e25b3c21eeBf11f15eB325d"
     }
   }
 }