From 167d5561624c7a3750ced2b06886ce49f6d743f5 Mon Sep 17 00:00:00 2001
From: Eric Arellano <ericarellano@me.com>
Date: Fri, 28 Jan 2022 17:41:14 -0700
Subject: [PATCH] Add `SpecsPaths`

# Rust tests and lints will be skipped. Delete if not intended.
[ci skip-rust]

# Building wheels and fs_util will be skipped. Delete if not intended.
[ci skip-build-wheels]
---
 src/python/pants/engine/fs.py              | 15 ++++-
 src/python/pants/engine/internals/graph.py | 65 +++++++++++++++-------
 2 files changed, 60 insertions(+), 20 deletions(-)

diff --git a/src/python/pants/engine/fs.py b/src/python/pants/engine/fs.py
index 3ff4f9ad5c1..c6fd924489a 100644
--- a/src/python/pants/engine/fs.py
+++ b/src/python/pants/engine/fs.py
@@ -256,15 +256,28 @@ def write_digest(
 
 @dataclass(frozen=True)
 class SpecsSnapshot:
-    """All files matched by command line specs.
+    """A snapshot of all files matched by command line specs.
 
     `@goal_rule`s may request this when they only need source files to operate and do not need any
     target information. This allows running on files with no owning targets.
+
+    Use `SpecsPaths` if you don't need a digest or you plan to filter the specified files and get
+    the digest of only that subset (via `PathGlobs`).
     """
 
     snapshot: Snapshot
 
 
+class SpecsPaths(Paths):
+    """All file names matched by command line specs.
+
+    `@goal_rule`s may request this when they only need source file names to operate and do not
+    need any target information. This allows running on files with no owning targets.
+
+    Use `SpecsSnapshot` if you plan to operate on every specified file and will need the digest.
+    """
+
+
 def rules():
     # Keep in sync with `intrinsics.rs`.
     return (
diff --git a/src/python/pants/engine/internals/graph.py b/src/python/pants/engine/internals/graph.py
index e1f2cd56a68..f78031e7915 100644
--- a/src/python/pants/engine/internals/graph.py
+++ b/src/python/pants/engine/internals/graph.py
@@ -36,6 +36,7 @@
     PathGlobs,
     Paths,
     Snapshot,
+    SpecsPaths,
     SpecsSnapshot,
 )
 from pants.engine.internals import native_engine
@@ -687,6 +688,33 @@ async def resolve_addresses_from_specs(specs: Specs) -> Addresses:
 # -----------------------------------------------------------------------------------------------
 
 
+@rule(desc="Find all file names from input specs", level=LogLevel.DEBUG)
+async def resolve_specs_paths(
+    specs: Specs, owners_not_found_behavior: OwnersNotFoundBehavior
+) -> SpecsPaths:
+    targets, filesystem_specs_paths = await MultiGet(
+        Get(Targets, AddressSpecs, specs.address_specs),
+        Get(
+            Paths,
+            PathGlobs,
+            specs.filesystem_specs.to_path_globs(
+                owners_not_found_behavior.to_glob_match_error_behavior()
+            ),
+        ),
+    )
+    all_sources_paths = await MultiGet(
+        Get(SourcesPaths, SourcesPathsRequest(tgt[SourcesField]))
+        for tgt in targets
+        if tgt.has_field(SourcesField)
+    )
+    files = OrderedSet()
+    dirs = OrderedSet()
+    for paths in (filesystem_specs_paths, *all_sources_paths):
+        files.update(paths.files)
+        dirs.update(paths.dirs)
+    return SpecsPaths(tuple(files), tuple(dirs))
+
+
 @rule(desc="Find all sources from input specs", level=LogLevel.DEBUG)
 async def resolve_specs_snapshot(
     specs: Specs, owners_not_found_behavior: OwnersNotFoundBehavior
@@ -696,31 +724,30 @@ async def resolve_specs_snapshot(
     Address specs will use their `SourcesField` field, and Filesystem specs will use whatever args
     were given. Filesystem specs may safely refer to files with no owning target.
     """
-    targets = await Get(Targets, AddressSpecs, specs.address_specs)
-    all_hydrated_sources = await MultiGet(
-        Get(HydratedSources, HydrateSourcesRequest(tgt[SourcesField]))
-        for tgt in targets
-        if tgt.has_field(SourcesField)
-    )
-
-    filesystem_specs_digest = (
-        await Get(
+    targets, filesystem_specs_digest = await MultiGet(
+        Get(Targets, AddressSpecs, specs.address_specs),
+        Get(
             Digest,
             PathGlobs,
             specs.filesystem_specs.to_path_globs(
                 owners_not_found_behavior.to_glob_match_error_behavior()
             ),
-        )
-        if specs.filesystem_specs
-        else None
+        ),
+    )
+    all_hydrated_sources = await MultiGet(
+        Get(HydratedSources, HydrateSourcesRequest(tgt[SourcesField]))
+        for tgt in targets
+        if tgt.has_field(SourcesField)
+    )
+    result = await Get(
+        Snapshot,
+        MergeDigests(
+            (
+                filesystem_specs_digest,
+                *(hydrated_sources.snapshot.digest for hydrated_sources in all_hydrated_sources),
+            )
+        ),
     )
-
-    # NB: We merge into a single snapshot to avoid the same files being duplicated if they were
-    # covered both by address specs and filesystem specs.
-    digests = [hydrated_sources.snapshot.digest for hydrated_sources in all_hydrated_sources]
-    if filesystem_specs_digest:
-        digests.append(filesystem_specs_digest)
-    result = await Get(Snapshot, MergeDigests(digests))
     return SpecsSnapshot(result)