From 6919c5152eebd08ab5ca7c92b16a261e16ead77b Mon Sep 17 00:00:00 2001
From: Niels Pardon <par@zurich.ibm.com>
Date: Mon, 10 Aug 2026 19:31:33 +0200
Subject: [PATCH] [GLUTEN-12742][CORE] Fix protoc DEPENDS so incremental builds
 regenerate proto sources

Both protoc `add_custom_command`s in cpp/core/CMakeLists.txt declared their
dependency on a directory rather than on the `.proto` files, and the trailing
slash passed to `get_filename_component(... DIRECTORY)` stripped a component,
so the declared path ended up one to two levels above the inputs:

  SUBSTRAIT_PROTO_SRC_DIR  .../resources/substrait/proto
  SUBSTRAIT_PROTO_DIR      .../resources/substrait          <-- declared dep
  files protoc reads       .../resources/substrait/proto/substrait/*.proto

Editing a `.proto` in place therefore changed no mtime in the build graph, and
an incremental build silently skipped protoc and kept linking stale
`*.pb.cc` / `*.pb.h`. This has been benign so far only because every
`algebra.proto` change to date was additive or comment-only; it turns wrong as
soon as a field is renumbered or removed, because the JVM then writes one tag
while the native library probes another and protobuf shunts the bytes into
unknown fields without raising anything.

Depend on the globbed file lists instead, matching what the ClickHouse backend
already does in cpp-ch/local-engine/proto/CMakeLists.txt. The
`get_filename_component` calls become dead and are dropped.

Also add `CONFIGURE_DEPENDS` to both globs so that adding or deleting a
`.proto` re-runs CMake instead of requiring a manual re-configure.

`dev/builddeps-veloxbe.sh` does `rm -rf build`, so CI and the official build
script were unaffected; this only bit local incremental builds, IDE builds, and
any workflow reusing cpp/build.

Fixes #12742
---
 cpp/core/CMakeLists.txt | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/cpp/core/CMakeLists.txt b/cpp/core/CMakeLists.txt
index 240567b8ff1..e0d82d42041 100644
--- a/cpp/core/CMakeLists.txt
+++ b/cpp/core/CMakeLists.txt
@@ -87,7 +87,8 @@ set(PROTO_OUTPUT_DIR "${CMAKE_CURRENT_BINARY_DIR}/proto")
 file(MAKE_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/proto)
 
 # List Substrait Proto compiled files
-file(GLOB SUBSTRAIT_PROTO_FILES ${SUBSTRAIT_PROTO_SRC_DIR}/substrait/*.proto
+file(GLOB SUBSTRAIT_PROTO_FILES CONFIGURE_DEPENDS
+     ${SUBSTRAIT_PROTO_SRC_DIR}/substrait/*.proto
      ${SUBSTRAIT_PROTO_SRC_DIR}/substrait/extensions/*.proto)
 foreach(PROTO ${SUBSTRAIT_PROTO_FILES})
   file(RELATIVE_PATH REL_PROTO ${SUBSTRAIT_PROTO_SRC_DIR} ${PROTO})
@@ -99,11 +100,9 @@ set(SUBSTRAIT_PROTO_OUTPUT_FILES ${SUBSTRAIT_PROTO_HDRS}
                                  ${SUBSTRAIT_PROTO_SRCS})
 set_source_files_properties(${SUBSTRAIT_PROTO_OUTPUT_FILES} PROPERTIES GENERATED
                                                                        TRUE)
-get_filename_component(SUBSTRAIT_PROTO_DIR ${SUBSTRAIT_PROTO_SRC_DIR}/
-                       DIRECTORY)
 
 # List Gluten Proto compiled files
-file(GLOB GLUTEN_PROTO_FILES ${GLUTEN_PROTO_SRC_DIR}/*.proto)
+file(GLOB GLUTEN_PROTO_FILES CONFIGURE_DEPENDS ${GLUTEN_PROTO_SRC_DIR}/*.proto)
 foreach(PROTO ${GLUTEN_PROTO_FILES})
   file(RELATIVE_PATH REL_PROTO ${GLUTEN_PROTO_SRC_DIR} ${PROTO})
   string(REGEX REPLACE "\\.proto" "" PROTO_NAME ${REL_PROTO})
@@ -113,7 +112,6 @@ endforeach()
 set(GLUTEN_PROTO_OUTPUT_FILES ${GLUTEN_PROTO_HDRS} ${GLUTEN_PROTO_SRCS})
 set_source_files_properties(${GLUTEN_PROTO_OUTPUT_FILES} PROPERTIES GENERATED
                                                                     TRUE)
-get_filename_component(GLUTEN_PROTO_DIR ${GLUTEN_PROTO_SRC_DIR}/ DIRECTORY)
 
 set(SPARK_COLUMNAR_PLUGIN_SRCS
     ${SUBSTRAIT_PROTO_SRCS}
@@ -221,7 +219,7 @@ add_custom_command(
   OUTPUT ${SUBSTRAIT_PROTO_OUTPUT_FILES}
   COMMAND ${PROTOC_BIN} --proto_path ${SUBSTRAIT_PROTO_SRC_DIR}/ --cpp_out
           ${PROTO_OUTPUT_DIR} ${SUBSTRAIT_PROTO_FILES}
-  DEPENDS ${SUBSTRAIT_PROTO_DIR}
+  DEPENDS ${SUBSTRAIT_PROTO_FILES}
   COMMENT "Running Substrait PROTO compiler"
   VERBATIM)
 
@@ -229,7 +227,7 @@ add_custom_command(
   OUTPUT ${GLUTEN_PROTO_OUTPUT_FILES}
   COMMAND ${PROTOC_BIN} --proto_path ${GLUTEN_PROTO_SRC_DIR}/ --cpp_out
           ${PROTO_OUTPUT_DIR} ${GLUTEN_PROTO_FILES}
-  DEPENDS ${GLUTEN_PROTO_DIR}
+  DEPENDS ${GLUTEN_PROTO_FILES}
   COMMENT "Running Gluten PROTO compiler"
   VERBATIM)