Merge pull request #2040 from conda-forge/lint-dup-cfyaml-keys

feat: add lint for duplicate keys in conda-forge.yml

conda_smithy/lint_recipe.py

@@ -11,6 +11,7 @@
 import github
 import jsonschema
 import requests
+from ruamel.yaml.constructor import DuplicateKeyError
 
 from conda_smithy.linter import conda_recipe_v1_linter
 from conda_smithy.linter.hints import (
@@ -571,6 +572,22 @@ def run_conda_forge_specific(
                 "The feedstock has no `.ci_support` files and thus will not build any packages."
             )
 
+    # 9: No duplicates in conda-forge.yml
+    if (
+        not is_staged_recipes
+        and recipe_dir is not None
+        and os.path.exists(
+            cfyml_pth := os.path.join(recipe_dir, "..", "conda-forge.yml")
+        )
+    ):
+        try:
+            with open(cfyml_pth) as fh:
+                get_yaml(allow_duplicate_keys=False).load(fh)
+        except DuplicateKeyError:
+            lints.append(
+                "The ``conda-forge.yml`` file is not allowed to have duplicate keys."
+            )
+
 
 def _format_validation_msg(error: jsonschema.ValidationError):
     """Use the data on the validation error to generate improved reporting.

conda_smithy/utils.py

@@ -75,14 +75,14 @@ def get_feedstock_about_from_meta(meta) -> dict:
         return dict(meta.meta["about"])
 
 
-def get_yaml():
+def get_yaml(allow_duplicate_keys: bool = True):
     # define global yaml API
     # roundrip-loader and allowing duplicate keys
     # for handling # [filter] / # [not filter]
     # Don't use a global variable for this as a global
     # variable will make conda-smithy thread unsafe.
     yaml = ruamel.yaml.YAML(typ="rt")
-    yaml.allow_duplicate_keys = True
+    yaml.allow_duplicate_keys = allow_duplicate_keys
     return yaml
 
 

news/cfyaml-duplicate-keys.rst

@@ -0,0 +1,3 @@
+**Added:**
+
+* Added a lint for duplicate keys in the ``conda-forge.yml`` file. (#2040)

tests/test_lint_recipe.py

@@ -2519,6 +2519,50 @@ def test_lint_no_builds():
         assert not any(lint.startswith(expected_message) for lint in lints)
 
 
+@unittest.skipUnless(is_gh_token_set(), "GH_TOKEN not set")
+def test_lint_duplicate_cfyml():
+    expected_message = (
+        "The ``conda-forge.yml`` file is not allowed to have duplicate keys."
+    )
+
+    with tmp_directory() as feedstock_dir:
+        cfyml = os.path.join(feedstock_dir, "conda-forge.yml")
+        recipe_dir = os.path.join(feedstock_dir, "recipe")
+        os.makedirs(recipe_dir, exist_ok=True)
+        with open(os.path.join(recipe_dir, "meta.yaml"), "w") as fh:
+            fh.write(
+                """
+                package:
+                   name: foo
+                """
+            )
+
+        with open(cfyml, "w") as fh:
+            fh.write(
+                textwrap.dedent(
+                    """
+                    blah: 1
+                    blah: 2
+                    """
+                )
+            )
+
+        lints = linter.main(recipe_dir, conda_forge=True)
+        assert any(lint.startswith(expected_message) for lint in lints)
+
+        with open(cfyml, "w") as fh:
+            fh.write(
+                textwrap.dedent(
+                    """
+                    blah: 1
+                    """
+                )
+            )
+
+        lints = linter.main(recipe_dir, conda_forge=True)
+        assert not any(lint.startswith(expected_message) for lint in lints)
+
+
 @pytest.mark.parametrize(
     "yaml_block,annotation",
     [