21726110101

Committed 05 Feb 2026 07:49PM UTC coverage: 80.284% (-0.01%) from 80.296%

Build # 21726110101

Build Type

push

github

Committed by

web-flow

Commit Message

Support dep inference on multiple files in a single workunit (#23075)

Each file's result is still cached independently.

In large repos we schedule a large number of very fast
tasks, so the scheduling overhead dominates the runtime,
so processing many files in a batch is a win.

The Python-side code doesn't make use of this yet, so it
passes single files and receives singletons back.

A followup change can actually put this into action.

Run Details

20 of 44 new or added lines in 5 files covered. (45.45%)

2 existing lines in 2 files now uncovered.

78530 of 97815 relevant lines covered (80.28%)

3.36 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

67.62

/src/python/pants/engine/intrinsics.py

# Copyright 2024 Pants project contributors (see CONTRIBUTORS.md).
# Licensed under the Apache License, Version 2.0 (see LICENSE).

from __future__ import annotations

import dataclasses
import logging

from pants.engine.environment import EnvironmentName
from pants.engine.fs import (
    AddPrefix,
    CreateDigest,
    Digest,
    DigestContents,
    DigestEntries,
    DigestSubset,
    MergeDigests,
    NativeDownloadFile,
    PathGlobs,
    PathMetadataRequest,
    PathMetadataResult,
    Paths,
    RemovePrefix,
    Snapshot,
)
from pants.engine.internals import native_engine
from pants.engine.internals.docker import DockerResolveImageRequest, DockerResolveImageResult
from pants.engine.internals.native_dep_inference import (
    NativeDockerfileInfos,
    NativeJavascriptFilesDependencies,
    NativePythonFilesDependencies,
)
from pants.engine.internals.native_engine import NativeDependenciesRequest, task_side_effected
from pants.engine.internals.session import RunId, SessionValues
from pants.engine.process import (
    FallibleProcessResult,
    InteractiveProcess,
    InteractiveProcessResult,
    Process,
    ProcessExecutionEnvironment,
    ProcessResultWithRetries,
    ProcessWithRetries,
)
from pants.engine.rules import _uncacheable_rule, collect_rules, implicitly, rule
from pants.util.docutil import git_url
from pants.util.frozendict import FrozenDict


@rule
async def create_digest(
    create_digest: CreateDigest,
) -> Digest:
    return await native_engine.create_digest(create_digest)


@rule
async def path_globs_to_digest(
    path_globs: PathGlobs,
) -> Digest:
    return await native_engine.path_globs_to_digest(path_globs)


@rule
async def path_globs_to_paths(
    path_globs: PathGlobs,
) -> Paths:
    return await native_engine.path_globs_to_paths(path_globs)


@rule
async def download_file(
    native_download_file: NativeDownloadFile,
) -> Digest:
    return await native_engine.download_file(native_download_file)


@rule
async def digest_to_snapshot(digest: Digest) -> Snapshot:
    return await native_engine.digest_to_snapshot(digest)


@rule
async def get_digest_contents(digest: Digest) -> DigestContents:
    return await native_engine.get_digest_contents(digest)


@rule
async def get_digest_entries(digest: Digest) -> DigestEntries:
    return await native_engine.get_digest_entries(digest)


@rule
async def merge_digests(merge_digests: MergeDigests) -> Digest:
    return await native_engine.merge_digests(merge_digests)


@rule
async def remove_prefix(remove_prefix: RemovePrefix) -> Digest:
    return await native_engine.remove_prefix(remove_prefix)


@rule
async def add_prefix(add_prefix: AddPrefix) -> Digest:
    return await native_engine.add_prefix(add_prefix)


@rule
async def execute_process(
    process: Process, process_execution_environment: ProcessExecutionEnvironment
) -> FallibleProcessResult:
    return await native_engine.execute_process(process, process_execution_environment)


@rule
async def execute_process_with_retry(req: ProcessWithRetries) -> ProcessResultWithRetries:
    results: list[FallibleProcessResult] = []
    for attempt in range(0, req.attempts):
        proc = dataclasses.replace(req.proc, attempt=attempt)
        result = await execute_process(  # noqa: PNT30: We only know that we need to rerun the test after we run it
            proc, **implicitly()
        )
        results.append(result)
        if result.exit_code == 0:
            break
    return ProcessResultWithRetries(tuple(results))


@rule
async def digest_subset_to_digest(digest_subset: DigestSubset) -> Digest:
    return await native_engine.digest_subset_to_digest(digest_subset)


@rule
async def session_values() -> SessionValues:
    return await native_engine.session_values()


@rule
async def run_id() -> RunId:
    return await native_engine.run_id()


__SQUELCH_WARNING = "__squelch_warning"


# NB: Call one of the helpers below, instead of calling this rule directly,
#  to ensure correct application of restartable logic.
@_uncacheable_rule
async def _interactive_process(
    process: InteractiveProcess, process_execution_environment: ProcessExecutionEnvironment
) -> InteractiveProcessResult:
    # This is a crafty way for a caller to signal into this function without a dedicated arg
    # (which would confound the solver).  Note that we go via __dict__ instead of using
    # setattr/delattr, because those error for frozen dataclasses.
    if __SQUELCH_WARNING in process.__dict__:
        del process.__dict__[__SQUELCH_WARNING]
    else:
        logging.warning(
            "A plugin is calling `await _interactive_process(...)` directly. This will cause "
            "restarting logic not to be applied. Use `await run_interactive_process(process)` "
            "or `await run_interactive_process_in_environment(process, environment_name)` instead. "
            f"See {git_url('src/python/pants/engine/intrinsics.py')} for more details."
        )
    return await native_engine.interactive_process(process, process_execution_environment)


async def run_interactive_process(process: InteractiveProcess) -> InteractiveProcessResult:
    # NB: We must call task_side_effected() in this helper, rather than in a nested @rule call,
    #  so that the Task for the @rule that calls this helper is the one marked as non-restartable.
    if not process.restartable:
        task_side_effected()

    process.__dict__[__SQUELCH_WARNING] = True
    ret: InteractiveProcessResult = await _interactive_process(process, **implicitly())
    return ret


async def run_interactive_process_in_environment(
    process: InteractiveProcess, environment_name: EnvironmentName
) -> InteractiveProcessResult:
    # NB: We must call task_side_effected() in this helper, rather than in a nested @rule call,
    #  so that the Task for the @rule that calls this helper is the one marked as non-restartable.
    if not process.restartable:
        task_side_effected()

    process.__dict__[__SQUELCH_WARNING] = True
    ret: InteractiveProcessResult = await _interactive_process(
        process, **implicitly({environment_name: EnvironmentName})
    )
    return ret


@rule
async def docker_resolve_image(request: DockerResolveImageRequest) -> DockerResolveImageResult:
    return await native_engine.docker_resolve_image(request)


@rule
async def parse_dockerfile_info(
    deps_request: NativeDependenciesRequest,
) -> NativeDockerfileInfos:
    path_infos_pairs = await native_engine.parse_dockerfile_info(deps_request)
    return NativeDockerfileInfos(FrozenDict(path_infos_pairs))


@rule
async def parse_python_deps(
    deps_request: NativeDependenciesRequest,
) -> NativePythonFilesDependencies:
    path_deps_pairs = await native_engine.parse_python_deps(deps_request)
    return NativePythonFilesDependencies(FrozenDict(path_deps_pairs))


@rule
async def parse_javascript_deps(
    deps_request: NativeDependenciesRequest,
) -> NativeJavascriptFilesDependencies:
    path_deps_pairs = await native_engine.parse_javascript_deps(deps_request)
    return NativeJavascriptFilesDependencies(FrozenDict(path_deps_pairs))


@rule
async def path_metadata_request(request: PathMetadataRequest) -> PathMetadataResult:
    return await native_engine.path_metadata_request(request)


def rules():
    return [
        *collect_rules(),
    ]

1	# Copyright 2024 Pants project contributors (see CONTRIBUTORS.md).
2	# Licensed under the Apache License, Version 2.0 (see LICENSE).
3
4	from __future__ import annotations	12✔
5
6	import dataclasses	12✔
7	import logging	12✔
8
9	from pants.engine.environment import EnvironmentName	12✔
10	from pants.engine.fs import (	12✔
11	AddPrefix,
12	CreateDigest,
13	Digest,
14	DigestContents,
15	DigestEntries,
16	DigestSubset,
17	MergeDigests,
18	NativeDownloadFile,
19	PathGlobs,
20	PathMetadataRequest,
21	PathMetadataResult,
22	Paths,
23	RemovePrefix,
24	Snapshot,
25	)
26	from pants.engine.internals import native_engine	12✔
27	from pants.engine.internals.docker import DockerResolveImageRequest, DockerResolveImageResult	12✔
28	from pants.engine.internals.native_dep_inference import (	12✔
29	NativeDockerfileInfos,
30	NativeJavascriptFilesDependencies,
31	NativePythonFilesDependencies,
32	)
33	from pants.engine.internals.native_engine import NativeDependenciesRequest, task_side_effected	12✔
34	from pants.engine.internals.session import RunId, SessionValues	12✔
35	from pants.engine.process import (	12✔
36	FallibleProcessResult,
37	InteractiveProcess,
38	InteractiveProcessResult,
39	Process,
40	ProcessExecutionEnvironment,
41	ProcessResultWithRetries,
42	ProcessWithRetries,
43	)
44	from pants.engine.rules import _uncacheable_rule, collect_rules, implicitly, rule	12✔
45	from pants.util.docutil import git_url	12✔
46	from pants.util.frozendict import FrozenDict	12✔
47
48
49	@rule	12✔
50	async def create_digest(	12✔
51	create_digest: CreateDigest,
52	) -> Digest:
53	return await native_engine.create_digest(create_digest)	×
54
55
56	@rule	12✔
57	async def path_globs_to_digest(	12✔
58	path_globs: PathGlobs,
59	) -> Digest:
60	return await native_engine.path_globs_to_digest(path_globs)	×
61
62
63	@rule	12✔
64	async def path_globs_to_paths(	12✔
65	path_globs: PathGlobs,
66	) -> Paths:
67	return await native_engine.path_globs_to_paths(path_globs)	×
68
69
70	@rule	12✔
71	async def download_file(	12✔
72	native_download_file: NativeDownloadFile,
73	) -> Digest:
74	return await native_engine.download_file(native_download_file)	×
75
76
77	@rule	12✔
78	async def digest_to_snapshot(digest: Digest) -> Snapshot:	12✔
79	return await native_engine.digest_to_snapshot(digest)	×
80
81
82	@rule	12✔
83	async def get_digest_contents(digest: Digest) -> DigestContents:	12✔
84	return await native_engine.get_digest_contents(digest)	×
85
86
87	@rule	12✔
88	async def get_digest_entries(digest: Digest) -> DigestEntries:	12✔
89	return await native_engine.get_digest_entries(digest)	×
90
91
92	@rule	12✔
93	async def merge_digests(merge_digests: MergeDigests) -> Digest:	12✔
94	return await native_engine.merge_digests(merge_digests)	×
95
96
97	@rule	12✔
98	async def remove_prefix(remove_prefix: RemovePrefix) -> Digest:	12✔
99	return await native_engine.remove_prefix(remove_prefix)	×
100
101
102	@rule	12✔
103	async def add_prefix(add_prefix: AddPrefix) -> Digest:	12✔
104	return await native_engine.add_prefix(add_prefix)	×
105
106
107	@rule	12✔
108	async def execute_process(	12✔
109	process: Process, process_execution_environment: ProcessExecutionEnvironment
110	) -> FallibleProcessResult:
111	return await native_engine.execute_process(process, process_execution_environment)	×
112
113
114	@rule	12✔
115	async def execute_process_with_retry(req: ProcessWithRetries) -> ProcessResultWithRetries:	12✔
116	results: list[FallibleProcessResult] = []	×
117	for attempt in range(0, req.attempts):	×
118	proc = dataclasses.replace(req.proc, attempt=attempt)	×
119	result = await execute_process( # noqa: PNT30: We only know that we need to rerun the test after we run it	×
120	proc, **implicitly()
121	)
122	results.append(result)	×
123	if result.exit_code == 0:	×
124	break	×
125	return ProcessResultWithRetries(tuple(results))	×
126
127
128	@rule	12✔
129	async def digest_subset_to_digest(digest_subset: DigestSubset) -> Digest:	12✔
130	return await native_engine.digest_subset_to_digest(digest_subset)	×
131
132
133	@rule	12✔
134	async def session_values() -> SessionValues:	12✔
135	return await native_engine.session_values()	×
136
137
138	@rule	12✔
139	async def run_id() -> RunId:	12✔
140	return await native_engine.run_id()	×
141
142
143	__SQUELCH_WARNING = "__squelch_warning"	12✔
144
145
146	# NB: Call one of the helpers below, instead of calling this rule directly,
147	# to ensure correct application of restartable logic.
148	@_uncacheable_rule	12✔
149	async def _interactive_process(	12✔
150	process: InteractiveProcess, process_execution_environment: ProcessExecutionEnvironment
151	) -> InteractiveProcessResult:
152	# This is a crafty way for a caller to signal into this function without a dedicated arg
153	# (which would confound the solver). Note that we go via __dict__ instead of using
154	# setattr/delattr, because those error for frozen dataclasses.
155	if __SQUELCH_WARNING in process.__dict__:	×
156	del process.__dict__[__SQUELCH_WARNING]	×
157	else:
158	logging.warning(	×
159	"A plugin is calling `await _interactive_process(...)` directly. This will cause "
160	"restarting logic not to be applied. Use `await run_interactive_process(process)` "
161	"or `await run_interactive_process_in_environment(process, environment_name)` instead. "
162	f"See {git_url('src/python/pants/engine/intrinsics.py')} for more details."
163	)
164	return await native_engine.interactive_process(process, process_execution_environment)	×
165
166
167	async def run_interactive_process(process: InteractiveProcess) -> InteractiveProcessResult:	12✔
168	# NB: We must call task_side_effected() in this helper, rather than in a nested @rule call,
169	# so that the Task for the @rule that calls this helper is the one marked as non-restartable.
170	if not process.restartable:	2✔
171	task_side_effected()	2✔
172
173	process.__dict__[__SQUELCH_WARNING] = True	2✔
174	ret: InteractiveProcessResult = await _interactive_process(process, **implicitly())	2✔
175	return ret	2✔
176
177
178	async def run_interactive_process_in_environment(	12✔
179	process: InteractiveProcess, environment_name: EnvironmentName
180	) -> InteractiveProcessResult:
181	# NB: We must call task_side_effected() in this helper, rather than in a nested @rule call,
182	# so that the Task for the @rule that calls this helper is the one marked as non-restartable.
183	if not process.restartable:	1✔
184	task_side_effected()	1✔
185
186	process.__dict__[__SQUELCH_WARNING] = True	1✔
187	ret: InteractiveProcessResult = await _interactive_process(	1✔
188	process, **implicitly({environment_name: EnvironmentName})
189	)
190	return ret	1✔
191
192
193	@rule	12✔
194	async def docker_resolve_image(request: DockerResolveImageRequest) -> DockerResolveImageResult:	12✔
195	return await native_engine.docker_resolve_image(request)	×
196
197
198	@rule	12✔
199	async def parse_dockerfile_info(	12✔
200	deps_request: NativeDependenciesRequest,
201	) -> NativeDockerfileInfos:
NEW 202	path_infos_pairs = await native_engine.parse_dockerfile_info(deps_request)	×
NEW 203	return NativeDockerfileInfos(FrozenDict(path_infos_pairs))	×
204
205
206	@rule	12✔
207	async def parse_python_deps(	12✔
208	deps_request: NativeDependenciesRequest,
209	) -> NativePythonFilesDependencies:
NEW 210	path_deps_pairs = await native_engine.parse_python_deps(deps_request)	×
NEW 211	return NativePythonFilesDependencies(FrozenDict(path_deps_pairs))	×
212
213
214	@rule	12✔
215	async def parse_javascript_deps(	12✔
216	deps_request: NativeDependenciesRequest,
217	) -> NativeJavascriptFilesDependencies:
NEW 218	path_deps_pairs = await native_engine.parse_javascript_deps(deps_request)	×
NEW 219	return NativeJavascriptFilesDependencies(FrozenDict(path_deps_pairs))	×
220
221
222	@rule	12✔
223	async def path_metadata_request(request: PathMetadataRequest) -> PathMetadataResult:	12✔
224	return await native_engine.path_metadata_request(request)	×
225
226
227	def rules():	12✔
228	return [	12✔
229	*collect_rules(),
230	]

pantsbuild / pants / 21726110101

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous