Skip to content

💫 Toolkit to help you get started with SDD or any other process!

fix(auth): normalize whitespace in auth-config env-var/id references at store time (#3691)

Ali jawwad71e6201790bc
diff --git a/src/specify_cli/authentication/config.py b/src/specify_cli/authentication/config.py
index 8d1faf80c..829940d6f 100644
--- a/src/specify_cli/authentication/config.py
+++ b/src/specify_cli/authentication/config.py
@@ -13,6 +13,7 @@ import stat
from dataclasses import dataclass
from fnmatch import fnmatch
from pathlib import Path
+from typing import Any
from urllib.parse import urlparse
@@ -53,6 +54,19 @@ def _is_valid_host_pattern(pattern: str) -> bool:
return pattern.startswith("*.") and "*" not in pattern[2:]
+def _norm(value: Any) -> Any:
+ """Strip surrounding whitespace from a whitespace-insignificant string
+ config reference (env-var names, tenant/client ids) before it is stored.
+
+ These fields are validated on their ``.strip()``ed form, so an accidentally
+ padded value passes validation but then silently breaks the verbatim
+ ``os.environ.get(...)`` / URL lookups downstream. Normalizing at store time
+ mirrors how ``hosts`` is already handled (``h.strip().lower()``). Non-string
+ values (e.g. ``None``) pass through unchanged.
+ """
+ return value.strip() if isinstance(value, str) else value
+
+
def load_auth_config(
path: Path | None = None,
) -> list[AuthConfigEntry]:
@@ -182,10 +196,10 @@ def load_auth_config(
provider=provider,
auth=auth,
token=token,
- token_env=token_env,
- tenant_id=entry_raw.get("tenant_id"),
- client_id=entry_raw.get("client_id"),
- client_secret_env=entry_raw.get("client_secret_env"),
+ token_env=_norm(token_env),
+ tenant_id=_norm(entry_raw.get("tenant_id")),
+ client_id=_norm(entry_raw.get("client_id")),
+ client_secret_env=_norm(entry_raw.get("client_secret_env")),
)
)
diff --git a/tests/test_authentication.py b/tests/test_authentication.py
index 06523c6f2..9edbd9860 100644
--- a/tests/test_authentication.py
+++ b/tests/test_authentication.py
@@ -92,6 +92,27 @@ class TestLoadAuthConfig:
assert entries[0].auth == "bearer"
assert entries[0].token_env == "GH_TOKEN"
+ def test_padded_token_env_is_normalized_and_resolves(self, tmp_path, monkeypatch):
+ # token_env is validated on its stripped form but was stored raw, so a
+ # padded env-var name passed validation yet broke the verbatim
+ # os.environ.get() lookup — resolve_token silently returned None.
+ monkeypatch.setenv("GH_TOKEN", "secret-tok")
+ cfg = tmp_path / "auth.json"
+ cfg.write_text(json.dumps({
+ "providers": [{
+ "hosts": ["github.com"],
+ "provider": "github",
+ "auth": "bearer",
+ "token_env": " GH_TOKEN ",
+ }]
+ }))
+ entries = load_auth_config(cfg)
+ assert len(entries) == 1
+ # Stored normalized (matching how hosts are normalized), ...
+ assert entries[0].token_env == "GH_TOKEN"
+ # ... so the env lookup finds the token instead of returning None.
+ assert GitHubAuth().resolve_token(entries[0]) == "secret-tok"
+
def test_valid_ado_config(self, tmp_path):
cfg = tmp_path / "auth.json"
cfg.write_text(json.dumps({
@@ -136,6 +157,26 @@ class TestLoadAuthConfig:
assert entries[0].auth == "azure-ad"
assert entries[0].tenant_id == "tid"
+ def test_padded_azure_ad_refs_are_normalized(self, tmp_path):
+ # The normalization also covers tenant_id / client_id / client_secret_env
+ # (used verbatim in the OAuth token URL/body and os.environ.get). Padded
+ # values were validated on their stripped form but stored raw.
+ cfg = tmp_path / "auth.json"
+ cfg.write_text(json.dumps({
+ "providers": [{
+ "hosts": ["dev.azure.com"],
+ "provider": "azure-devops",
+ "auth": "azure-ad",
+ "tenant_id": " tid ",
+ "client_id": " cid ",
+ "client_secret_env": " SECRET ",
+ }]
+ }))
+ entries = load_auth_config(cfg)
+ assert entries[0].tenant_id == "tid"
+ assert entries[0].client_id == "cid"
+ assert entries[0].client_secret_env == "SECRET"
+
def test_azure_cli_config(self, tmp_path):
cfg = tmp_path / "auth.json"
cfg.write_text(json.dumps({