When building CLucene, CMake may find the wrong Boost and zlib. We should pass the search paths to the build command for CLucene explicitly to find the correct dependencies.
Fix conflit name TCHAR in odbc sqltypes.h and clucene clucene-config.h.
Change TCHAR to TWCHAR in odbc sqltypes.h, because TCHAR in odbc is not found used in doris,
but there are too many places to call clucene's TCHAR.
thirdparty/installed/include/sqltypes.h:
`typedef char TCHAR;`
thirdparty/installed/include/CLucene/clucene-config.h:
`typedef wchar_t TCHAR;`
As part of Inverted Index DSIP steps, we'd like to contribute our inverted index implementations step by step.
First of all we need to introduce clucene to doris thirdparty libs, because inverted index implementations are based on
lucence API and index file format, also we add our features and performance improvements base on clucene, so we
need to maintain the repo ourselves
HOW to reproduce?
Add export CMAKE_BUILD_TYPE=DEBUG in custom_env.sh. Then build thirdparty in MAC.
There are two problems:
build vectorscan with DEBUG type, will got unused-but-set-variable error:
doris/thirdparty/src/vectorscan-vectorscan-5.4.7/src/nfa/mcclellancompile.cpp:1485:13: error: variable 'total_daddy' set but not used [-Werror,-Wunused-but-set-variable]
u16 total_daddy = 0;
gflags will output libgflags_debug.a instead of libgflags.a while build with DEBUG type. Then we will got error can not find library gflags error.
To avoid these errors, we set CMAKE_BUILD_TYPE while build vectorscan and gflags.
Co-authored-by: cambyzju <zhuxiaoli01@baidu.com>
Co-authored-by: Adonis Ling <adonis0147@gmail.com>
In #15037, I modified the build script of libgsasl to enable GSSAPI,
but it is still wrong, because the PATH does not include the `thirdparty/installed/bin`,
so when building libgsasl, it will report error:
`WARNING: MIT Kerberos krb5-config not found, disabling GSSAPI`
but `krb5-config` is in `thirdparty/installed/bin`.
Without GSSAPI, the libhdfs3 can not access hdfs with kerberos authentication.
On macOS, we need some extra libraries to build the codebase,
therefore two packages were introduced to the project. They are `binutils` and `gettext`.
It takes a lot of time to build these packages completely. This PR introduces a way to build the needed libraries
and other stuff are skipped to build. It can save the time to build the third-party libraries on macOS.
When doris be getFileStatus from HDFS2 server, libhdfs3 will throw exception because of the permission code returned by hdfs2 server is greater than 1<<12.
The bit 12 of permission code is aclBit which has been deprecated in hadoop3. so we remove the check code in libhdfs3, same as hadoop3 java project.
Currently, we may fail to build the third-party libraries if we keep the outdated extracted data.
Considering the following scenario, Bob added patches to some libraries and Alice updates the codebase and builds
the third-party libraries. If Alice kept the outdated extracted data, she should fail to build the third-party libraries
because the patches are not applied due to the outdated `patched_marks`.
This PR introduces a way to clean the outdated data before building the third-party libraries.
Upgrade simdjson from 1.0.2 to latest version 3.0.1 to avoid -mlzcnt compiler flag causing BE UT(macOS) failure.
simdjson is now only used by VJsonScanner and disabled by default. So the impact of upgrade is limited.
ORC NextStripeReader now only support read columns by indices, but it is hard to get column indices for complex types.
We patch ORC adapter to support read columns by column names.
Co-authored-by: cambyzju <zhuxiaoli01@baidu.com>
There are some issues with the compile flag `-Wno-unused-but-set-variable` for clang.
1. `-Wno-unused-but-set-variable` should be set when building source by clang-15 on Linux. (#13000#13016)
2. On macOS Monterey, Apple Clang 13 may treat it as a unknown warning option and the compilation process may interrupt.
This PR introduces a better way to make this compile flag more portable.
1. Test whether the compiler recognizes this flag.
2. Add this flag if the compiler recognizes it.
Overwrite the environment variable PATH to avoid using binutils from Homebrew to build third parties which may cause compilation errors.
Error: building for macOS-x86_64 but attempting to link with file built for unknown-unsupported file format
# Proposed changes
This PR fixed lots of issues when building from source on macOS with Apple M1 chip.
## ATTENTION
The job for supporting macOS with Apple M1 chip is too big and there are lots of unresolved issues during runtime:
1. Some errors with memory tracker occur when BE (RELEASE) starts.
2. Some UT cases fail.
...
Temporarily, the following changes are made on macOS to start BE successfully.
1. Disable memory tracker.
2. Use tcmalloc instead of jemalloc.
This PR kicks off the job. Guys who are interested in this job can continue to fix these runtime issues.
## Use case
```shell
./build.sh -j 8 --be --clean
cd output/be/bin
ulimit -n 60000
./start_be.sh --daemon
```
## Something else
It takes around _**10+**_ minutes to build BE (with prebuilt third-parties) on macOS with M1 chip. We will improve the development experience on macOS greatly when we finish the adaptation job.
Strip debug info of most of thridparty dependencies' static lib.
If can significantly reduce the size of thirdparty libs: 3.4G -> 1.6G
And the doris_be binary size will be reduced: 1.5G -> 868M (clang build)
And after compress, the BE binary is only 195M with debug info!
