Consider eliminating FCL's dependence on boost

[sherm1]

For discussion -- boost is a very heavyweight dependency and could be an impediment to FCL adoption. It can also cause problems when FCL is integrated into a larger program that already has a boost dependency, possibly involving a different version of boost (Matlab comes with a version linked in, which caused Drake to have to drop boost).

I'm wondering if the need for boost could be eliminated in favor of C++11 use instead? Are there boost features in use that C++11 can't provide?

[mamoll]

I'd be in favor of that. I just started the transition to C++11 in OMPL (https://github.com/ompl/ompl/tree/c++11). After grep-ing through FCL's code, it looks fairly straightforward.

One thing that took me a while to figure out was that C++03 and C++11 code are not ABI compatible (usually things just work, though). The FCL library I had installed was compiled linked against Boost and both were built on C++03 mode. After compiling OMPL in C++11 code I got some odd crashes, possibly related to some constexpr's in Boost. So eliminating Boost as a dependency and forcing C++11 could potentially make it easier to adopt FCL.

[scpeters]

Requiring c++11 is a good way to remove many uses of boost, which I have been spearheading for urdfdom and some other repositories. Here's a list of boost header files included in the FCL master branch:

$ grep -rnIi include.*boost . | sed -e 's@.*:#include@@' | tr -d '\r' | sort | uniq
 <boost/array.hpp>
 <boost/assert.hpp>
 <boost/assign/list_of.hpp>
 <boost/bind.hpp>
 <boost/cstdint.hpp>
 <boost/date_time/posix_time/posix_time.hpp>
 <boost/dynamic_bitset.hpp>
 <boost/filesystem.hpp>
 <boost/foreach.hpp>
 <boost/function.hpp>
 <boost/iterator/zip_iterator.hpp>
 <boost/math/constants/constants.hpp>
 <boost/math/special_functions/erf.hpp>
 <boost/math/special_functions/fpclassify.hpp>
 <boost/noncopyable.hpp>
 <boost/random/lagged_fibonacci.hpp>
 <boost/random/mersenne_twister.hpp>
 <boost/random/normal_distribution.hpp>
 <boost/random/uniform_int.hpp>
 <boost/random/uniform_real.hpp>
 <boost/random/variate_generator.hpp>
 <boost/shared_array.hpp>
 <boost/shared_ptr.hpp>
 <boost/test/unit_test.hpp>
 <boost/thread.hpp>
 <boost/thread/mutex.hpp>
 <boost/timer.hpp>
 <boost/unordered_map.hpp>
 <boost/unordered_set.hpp>
 <boost/weak_ptr.hpp>

One that jumps out at me is boost filesystem, which isn't available in c++11, though it may be included in some later standards.

[mamoll]

boost filesystem is only used in the unit tests. The boost unit testing framework would also be a problem. Perhaps limiting boost to unit tests is acceptable (since end users wouldn't have to install boost to use fcl)?

[jslee02]

It seems filesystem will be included in C++17. Also, zip_iterator actually isn't used anywhere.

[isucan]

I am also fully in favor of dropping the boost dependency. I think FCL's
use of is is rather light, so it should be quite doable.
On Mar 22, 2016 6:07 PM, "Jeongseok Lee" notifications@github.com wrote:

It seems filesystem will be included in C++17
https://en.wikipedia.org/wiki/C%2B%2B17. Also, zip_iterator actually
isn't used nowhere.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub
#100 (comment)

[scpeters]

@sherm1 we could make a checklist based on that list of header files that I made and then check them off as they are removed

[sherm1]

we could make a checklist based on that list of header files that I made and then check them off as they are removed

Great idea, Steve! Then we could do small PRs that knock them out one at a time and check them off until the last one is gone.

[sherm1]

@mamoll: One thing that took me a while to figure out was that C++03 and C++11 code are not ABI compatible

Yeah, there is no going back. You have to commit to C++11 and then everyone who links with you does also. But ... this is now a 5 year old standard so it doesn't seem too soon to switch!

[sherm1]

I made issue #102 with Steve's boost checklist in it.

[jslee02]

All the boost dependencies are removed by #103, #104, #105, #108, #140, #146, #147, and #148!

Closing this issue.