Rules for building PCL with bazel.
In your WORKSPACE.bazel file:
load("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
http_archive(
name = "rules_pcl",
url = "https://github.com/kgreenek/rules_pcl/archive/refs/tags/pcl-1.13.1-v1.tar.gz",
sha256 = "06d2ed755fbf9ddd30f8bb3ed73f23ee6155d52e742d4e4a278c9458e9a3e17b",
strip_prefix = "rules_pcl-pcl-1.13.1-v1",
)
load("@rules_pcl//bzl:repositories.bzl", "pcl_repositories")
pcl_repositories()
# NOTE: This must be loaded after the call to pcl_repositories().
load("@rules_pcl//bzl:init_deps.bzl", "pcl_init_deps")
pcl_init_deps()
In your top-level BUILD.bazel file:
load("@rules_pcl//bzl:pcl.bzl", "pcl_config")
pcl_config()
You can customize how pcl is compiled by passing args to the pcl_config() macro, but the defaults will suffice for most. See bzl/pcl.bzl for more info.
To define a target that depends on PCL:
cc_binary(
name = "example",
srcs = ["main.cc"],
deps = ["@pcl//:common"],
)
PCL requires C++14. You can either pass it as an argument to the bazel build command:
bazel build --cxxopt='-std=c++14' //...
Or you can make C++14 the default for all builds by adding this to a file called .bazelrc in your
workspace root:
build --cxxopt='-std=c++14'
Or you can update your toolchain to build with C++14.
All sub-modules are currently supported except for the following:
- cuda
- gpu
- outofcore
- visualization
An example bazel workspace can be found under the //example directory of this repo.