(build) fix some compile and conan problems

[yangzhg]

What problem does this PR solve?

Bolt builds in strict or isolated toolchain environments exposed several build and dependency-management issues:

• Supported GCC versions no longer require a separate libstdc++fs, but Bolt still linked against -lstdc++fs, causing linker failures in environments where that library is unavailable.
• Absolute source and build paths could be embedded in generated artifacts.
• The Makefile did not automatically consume externally supplied Conan host and build profiles.
• Several pinned Conan recipes had incompatible option names, CMake targets, or compiler settings.
• GNU and Apache source downloads lacked preferred mirrors.
• The Volcengine TOS C++ SDK did not have a local Conan recipe based on upstream release archives.

Issue Number: N/A

Type of Change

• 🐛 Bug fix (non-breaking change which fixes an issue)
• ✨ New feature (non-breaking change which adds functionality)
• 🚀 Performance improvement (optimization)
• ⚠️ Breaking change (fix or feature that would cause existing functionality to change)
• 🔨 Refactoring (no logic changes)
• 🔧 Build/CI or Infrastructure changes
• 📝 Documentation only

Description

This PR improves Bolt build compatibility across different compiler, sysroot, and Conan environments:

• Add configurable compiler prefix mapping for GCC and Clang to replace absolute source and build paths embedded in binaries, debug information, predefined macros, and coverage metadata.
• Automatically consume externally supplied Conan host and build profiles while preserving the existing profile fallback.
• Align conan graph info arguments with the actual conan install invocation.
• Standardize the minimum supported GCC version as GCC 10 and remove obsolete stdc++fs linkage from Bolt and its test targets.
• Align Arrow options with the pinned Conan recipe, including acero and the legacy hdfs_bridgs option.
• Fix Arrow integration for RapidJSON and Snappy CMake target detection.
• Build pkgconf with GNU99 extensions required by its source.
• Fix Snappy target naming and Conan dependency discovery for Paimon and its Avro external project.
• Prepend Kernel.org mirrors for the required GNU sources while retaining all original URLs as fallbacks.
• Prepend Huawei Cloud mirrors for Arrow 15.0.1, ORC 2.0.0, and Thrift 0.17.0 while retaining their original Apache URLs.
• Add a Conan recipe for all published Volcengine TOS C++ SDK releases from 2.6.0 through 2.6.28:
  • Source URLs and SHA-256 checksums are managed through conandata.yml.
  • Sources are downloaded from upstream GitHub tag archives.
  • CMake package metadata and a Conan test_package are included.

Validation performed:

• Ran scripts/install-bolt-deps.sh with an isolated Conan home and confirmed that all Conan Center patches apply successfully.
• Downloaded the added Apache mirror artifacts and verified their SHA-256 checksums.
• Exported and fetched the TOS recipes for versions 2.6.0 and 2.6.28.

Performance Impact

• No Impact: This change only affects the build system and dependency recipes and does not modify runtime critical paths.
• Positive Impact: I have run benchmarks.
• Negative Impact: Explained below (e.g., trade-off for correctness).

Release Note

Release Note:

Release Note:

• Improved Bolt build compatibility across strict and isolated toolchain environments.
• Added compiler path mapping to avoid embedding absolute source and build paths.
• Added fallback mirrors for pinned GNU and Apache dependencies.
• Added Conan recipes for Volcengine TOS C++ SDK releases from 2.6.0 through 2.6.28.

Checklist (For Author)

• I have added/updated unit tests (ctest).
• I have verified the code with local build (Release/Debug).
• I have run clang-format / linters.
• (Optional) I have run Sanitizers (ASAN/TSAN) locally for complex C++ changes.
• No need to test or manual test.

Breaking Changes

• No
• Yes (Description: ...)

[Copilot]

Pull request overview

This PR improves Bolt’s build reliability in strict/isolated toolchain environments by cleaning up obsolete linkage, aligning Conan integration and pinned recipes, adding source mirrors, and introducing a local Conan recipe for the Volcengine TOS C++ SDK.

Changes:

• Add compiler prefix/path mapping flags in CMake to avoid embedding absolute source/build paths in artifacts.
• Improve Conan profile handling/graph info generation in the Makefile; align Arrow/Snappy/RapidJSON-related recipe patches and add Apache/GNU mirror fallbacks.
• Add a full Conan recipe + test_package for tos_client across multiple upstream SDK versions.

Reviewed changes

Copilot reviewed 21 out of 21 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
scripts/conan/recipes/tos_client/config.yml	Registers multiple tos_client versions for the local recipe.
scripts/conan/recipes/tos_client/all/conanfile.py	Adds the Conan recipe for building/packaging the Volcengine TOS C++ SDK.
scripts/conan/recipes/tos_client/all/conandata.yml	Defines per-version source URLs and checksums for the TOS SDK archives.
scripts/conan/recipes/tos_client/all/test_package/conanfile.py	Adds Conan test_package to validate the exported package via CMake.
scripts/conan/recipes/tos_client/all/test_package/CMakeLists.txt	Builds and links a minimal consumer against tos_client::tos_client.
scripts/conan/recipes/tos_client/all/test_package/test_package.cpp	Minimal runtime check that exercises a symbol from the SDK.
scripts/conan/recipes/paimon-cpp/all/patches/paimon-cpp-conan-deps.patch	Adjusts Snappy discovery/targets and improves Conan/CMakeDeps integration for ExternalProject builds.
scripts/conan/patches/source-mirrors-apache.patch	Prepends Huawei Cloud mirrors for selected Apache sources.
scripts/conan/patches/gnu-source-mirrors.patch	Prepends kernel.org mirrors for required GNU sources.
scripts/conan/patches/pkgconf-2.2.0-fix-gnu99.patch	Ensures pkgconf is built with GNU99 extensions as required by its sources.
scripts/conan/patches/arrow.15.0.1-csv-support.patch	Aligns Arrow 15.0.1 options/targets and fixes Snappy/RapidJSON-related integration details.
Makefile	Consumes externally supplied Conan host/build profiles and aligns conan graph info args with install.
conanfile.py	Aligns Arrow options with the pinned Arrow recipe (e.g., acero, legacy hdfs_bridgs).
CMakeLists.txt	Adds file-prefix-map support and removes obsolete stdc++fs linkage while standardizing GCC>=10.
bolt/substrait/tests/CMakeLists.txt	Removes ${FILESYSTEM} linkage from tests.
bolt/jit/tests/CMakeLists.txt	Removes ${FILESYSTEM} linkage from tests.
bolt/dwio/txt/tests/CMakeLists.txt	Removes ${FILESYSTEM} linkage from tests.
bolt/dwio/parquet/tests/CMakeLists.txt	Removes ${FILESYSTEM} linkage from tests.
bolt/dwio/dwrf/test/utils/CMakeLists.txt	Removes legacy GCC<9 stdc++fs linkage workaround.
bolt/dwio/dwrf/test/CMakeLists.txt	Removes ${FILESYSTEM} linkage from tests.
bolt/dwio/common/tests/utils/CMakeLists.txt	Removes legacy GCC<9 stdc++fs linkage workaround.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.