[Bazel/C++] Augment cc_dist_library to generate lists of source fil…

…es (#10026)

This change largely moves internal logic that creates the the `CcFileList` provider from `build_systems.bzl` to `cc_dist_library.bzl`.

There are also some associated changes to the particular `cc_dist_library` targets, since the output didn't really make sense after previous BUILD.bazel refactoring. There is also a target now for `libprotoc`.

pkg/BUILD.bazel

@@ -305,9 +305,9 @@ gen_file_lists(
     out_stem = "src_file_lists",
     src_libs = {
         # source rule: name in generated file
-        "//:protobuf": "libprotobuf",
-        "//src/google/protobuf/compiler:protoc_lib": "libprotoc",
-        "//:protobuf_lite": "libprotobuf_lite",
+        ":protobuf": "libprotobuf",
+        ":protoc": "libprotoc",
+        ":protobuf_lite": "libprotobuf_lite",
     },
 )
 
@@ -343,8 +343,8 @@ cc_dist_library(
     }),
     tags = ["manual"],
     deps = [
-        "//:protobuf_lite",
         "//src/google/protobuf:arena",
+        "//src/google/protobuf:protobuf_lite",
         "//src/google/protobuf/io",
         "//src/google/protobuf/io:io_win32",
         "//src/google/protobuf/stubs:lite",
@@ -362,15 +362,15 @@ cc_dist_library(
     }),
     tags = ["manual"],
     deps = [
-        "//:protobuf",
-        "//:protobuf_lite",
         "//src/google/protobuf:arena",
         "//src/google/protobuf/compiler:importer",
         "//src/google/protobuf/io",
         "//src/google/protobuf/io:gzip_stream",
         "//src/google/protobuf/io:io_win32",
         "//src/google/protobuf/io:printer",
         "//src/google/protobuf/io:tokenizer",
+        "//src/google/protobuf:protobuf",
+        "//src/google/protobuf:protobuf_lite",
         "//src/google/protobuf/stubs",
         "//src/google/protobuf/stubs:lite",
         "//src/google/protobuf/util:delimited_message_util",
@@ -382,6 +382,22 @@ cc_dist_library(
     ],
 )
 
+cc_dist_library(
+    name = "protoc",
+    tags = ["manual"],
+    deps = [
+        "//src/google/protobuf/compiler:code_generator",
+        "//src/google/protobuf/compiler:command_line_interface",
+        "//src/google/protobuf/compiler/cpp",
+        "//src/google/protobuf/compiler/csharp",
+        "//src/google/protobuf/compiler/java",
+        "//src/google/protobuf/compiler/objectivec",
+        "//src/google/protobuf/compiler/php",
+        "//src/google/protobuf/compiler/python",
+        "//src/google/protobuf/compiler/ruby",
+    ],
+)
+
 ################################################################################
 # Distribution sources
 ################################################################################

pkg/build_systems.bzl

@@ -1,6 +1,7 @@
 # Starlark utilities for working with other build systems
 
 load("@rules_pkg//:providers.bzl", "PackageFilegroupInfo", "PackageFilesInfo")
+load(":cc_dist_library.bzl", "CcFileList")
 
 ################################################################################
 # Macro to create CMake and Automake source lists.
@@ -31,21 +32,6 @@ def gen_file_lists(name, out_stem, **kwargs):
 # Aspect that extracts srcs, hdrs, etc.
 ################################################################################
 
-CcFileList = provider(
-    doc = "List of files to be built into a library.",
-    fields = {
-        # As a rule of thumb, `hdrs` and `textual_hdrs` are the files that
-        # would be installed along with a prebuilt library.
-        "hdrs": "public header files, including those used by generated code",
-        "textual_hdrs": "files which are included but are not self-contained",
-
-        # The `internal_hdrs` are header files which appear in `srcs`.
-        # These are only used when compiling the library.
-        "internal_hdrs": "internal header files (only used to build .cc files)",
-        "srcs": "source files",
-    },
-)
-
 ProtoFileList = provider(
     doc = "List of proto files and generated code to be built into a library.",
     fields = {
@@ -65,56 +51,11 @@ def _flatten_target_files(targets):
             files.append(tfile)
     return files
 
-def _combine_cc_file_lists(file_lists):
-    hdrs = {}
-    textual_hdrs = {}
-    internal_hdrs = {}
-    srcs = {}
-    for file_list in file_lists:
-        hdrs.update({f: 1 for f in file_list.hdrs})
-        textual_hdrs.update({f: 1 for f in file_list.textual_hdrs})
-        internal_hdrs.update({f: 1 for f in file_list.internal_hdrs})
-        srcs.update({f: 1 for f in file_list.srcs})
-    return CcFileList(
-        hdrs = sorted(hdrs.keys()),
-        textual_hdrs = sorted(textual_hdrs.keys()),
-        internal_hdrs = sorted(internal_hdrs.keys()),
-        srcs = sorted(srcs.keys()),
-    )
-
 def _file_list_aspect_impl(target, ctx):
     # We're going to reach directly into the attrs on the traversed rule.
     rule_attr = ctx.rule.attr
     providers = []
 
-    # Extract sources from a `cc_library` (or similar):
-    if CcInfo in target:
-        # CcInfo is a proxy for what we expect this rule to look like.
-        # However, some deps may expose `CcInfo` without having `srcs`,
-        # `hdrs`, etc., so we use `getattr` to handle that gracefully.
-
-        internal_hdrs = []
-        srcs = []
-
-        # Filter `srcs` so it only contains source files. Headers will go
-        # into `internal_headers`.
-        for src in _flatten_target_files(getattr(rule_attr, "srcs", [])):
-            if src.extension.lower() in ["c", "cc", "cpp", "cxx"]:
-                srcs.append(src)
-            else:
-                internal_hdrs.append(src)
-
-        providers.append(CcFileList(
-            hdrs = _flatten_target_files(getattr(rule_attr, "hdrs", [])),
-            textual_hdrs = _flatten_target_files(getattr(
-                rule_attr,
-                "textual_hdrs",
-                [],
-            )),
-            internal_hdrs = internal_hdrs,
-            srcs = srcs,
-        ))
-
     # Extract sources from a `proto_library`:
     if ProtoInfo in target:
         proto_srcs = []
@@ -197,19 +138,28 @@ def _create_file_list_impl(ctx, fragment_generator):
     for srcrule, libname in ctx.attr.src_libs.items():
         if CcFileList in srcrule:
             cc_file_list = srcrule[CcFileList]
+
+            # Turn depsets of files into sorted lists.
+            srcs = sorted(cc_file_list.srcs.to_list())
+            hdrs = sorted(
+                depset(transitive = [
+                    cc_file_list.textual_hdrs,
+                    cc_file_list.hdrs,
+                ]).to_list(),
+            )
+
             fragments.extend([
                 fragment_generator(
                     srcrule.label,
                     libname + "_srcs",
                     ctx.attr.source_prefix,
-                    [f.short_path for f in cc_file_list.srcs],
+                    [f.short_path for f in srcs],
                 ),
                 fragment_generator(
                     srcrule.label,
                     libname + "_hdrs",
                     ctx.attr.source_prefix,
-                    [f.short_path for f in (cc_file_list.hdrs +
-                                            cc_file_list.textual_hdrs)],
+                    [f.short_path for f in hdrs],
                 ),
             ])
 
@@ -247,7 +197,7 @@ def _create_file_list_impl(ctx, fragment_generator):
             # keys are the destination:
             files.update(srcrule[PackageFilesInfo].dest_src_map)
 
-        if files == {} and DefaultInfo in srcrule and CcInfo not in srcrule:
+        if files == {} and DefaultInfo in srcrule and CcFileList not in srcrule:
             # This could be an individual file or filegroup.
             # We explicitly ignore rules with CcInfo, since their
             # output artifacts are libraries or binaries.