Do not run item pipelines or feed exports on scrapy check

docs/topics/contracts.rst

@@ -40,7 +40,13 @@ You can use the following contracts:
 
 .. autoclass:: ScrapesContract
 
-Use the :command:`check` command to run the contract checks.
+Use the :command:`check` command to run the contract checks. It ignores
+:setting:`ITEM_PIPELINES` and :setting:`FEEDS`, since contracts check the
+output of callbacks instead of sending it to item processing; use the ``-s``
+command-line option to set them back for a check run.
+
+.. versionchanged:: VERSION
+   :setting:`ITEM_PIPELINES` and :setting:`FEEDS` are now ignored.
 
 Custom Contracts
 ================

scrapy/commands/check.py

@@ -9,6 +9,7 @@ from unittest import TextTestRunner
 from scrapy import Spider
 from scrapy.commands import ScrapyCommand
 from scrapy.contracts import ContractsManager
+from scrapy.settings import SETTINGS_PRIORITIES
 from scrapy.utils.conf import build_component_list
 from scrapy.utils.misc import load_object, set_environ
 
@@ -70,6 +71,18 @@ class Command(ScrapyCommand):
             help="print contract tests for all spiders",
         )
 
+    def process_options(self, args: list[str], opts: argparse.Namespace) -> None:
+        super().process_options(args, opts)
+        assert self.settings is not None
+        # Contracts discard the output of callbacks, so item pipelines and
+        # feed exports get no items, and opening them only causes side
+        # effects, such as an empty output file. The priority is above spider
+        # custom settings, which also define them, and below the command line,
+        # which can hence set them back.
+        priority = (SETTINGS_PRIORITIES["spider"] + SETTINGS_PRIORITIES["cmdline"]) // 2
+        self.settings.set("ITEM_PIPELINES", {}, priority=priority)
+        self.settings.set("FEEDS", {}, priority=priority)
+
     def run(self, args: list[str], opts: argparse.Namespace) -> None:
         # load contracts
         assert self.settings is not None

tests/test_command_check.py

@@ -21,7 +21,13 @@ class DummyTestCase(TestCase):
 class TestCheckCommand(TestProjectBase):
     spider_name = "check_spider"
 
-    def _write_contract(self, proj_path: Path, contracts: str, parse_def: str) -> None:
+    def _write_contract(
+        self,
+        proj_path: Path,
+        contracts: str,
+        parse_def: str,
+        custom_settings: str = "",
+    ) -> None:
         spider = proj_path / self.project_name / "spiders" / "checkspider.py"
         spider.write_text(
             f"""
@@ -33,6 +39,7 @@ class CheckSpider(scrapy.Spider):
 
     custom_settings = {{
         "DOWNLOAD_DELAY": 0,
+        {custom_settings}
     }}
 
     def parse(self, response, **cb_kwargs):
@@ -51,8 +58,9 @@ class CheckSpider(scrapy.Spider):
         contracts: str = "",
         parse_def: str = "pass",
         use_reactor: bool = True,
+        custom_settings: str = "",
     ) -> None:
-        self._write_contract(proj_path, contracts, parse_def)
+        self._write_contract(proj_path, contracts, parse_def, custom_settings)
         args = ["check"]
         if not use_reactor:
             args += ["-s", "TWISTED_REACTOR_ENABLED=False"]
@@ -122,6 +130,51 @@ class CheckSpider(scrapy.Spider):
         """
         self._test_contract(proj_path, contracts, parse_def)
 
+    def test_check_no_item_processing(self, proj_path: Path) -> None:
+        (proj_path / self.project_name / "pipelines.py").write_text(
+            """
+from pathlib import Path
+
+class MarkerPipeline:
+    def open_spider(self, spider):
+        Path("pipeline.txt").touch()
+
+    def process_item(self, item, spider):
+        return item
+""",
+            encoding="utf-8",
+        )
+        self._append_settings(
+            proj_path / self.project_name,
+            """
+FEEDS = {"items.jsonl": {"format": "jsonlines"}}
+""",
+        )
+        contracts = """
+        @returns items 1
+        """
+        parse_def = """
+        yield {'key1': 'val1'}
+        """
+        custom_settings = f'"ITEM_PIPELINES": {{"{self.project_name}.pipelines.MarkerPipeline": 100}},'
+        self._test_contract(
+            proj_path, contracts, parse_def, custom_settings=custom_settings
+        )
+        assert not (proj_path / "items.jsonl").exists()
+        assert not (proj_path / "pipeline.txt").exists()
+
+        ret, _, err = proc(
+            "check",
+            "-s",
+            f'ITEM_PIPELINES={{"{self.project_name}.pipelines.MarkerPipeline": 100}}',
+            "-s",
+            'FEEDS={"items.jsonl": {"format": "jsonlines"}}',
+            cwd=proj_path,
+        )
+        assert ret == 0, err
+        assert (proj_path / "items.jsonl").exists()
+        assert (proj_path / "pipeline.txt").exists()
+
     def test_SCRAPY_CHECK_set(self, proj_path: Path) -> None:
         parse_def = """
         import os