Skip to content

Conversation

@Flamefire
Copy link

@pdimov I created a PR for discussing proposed changes which is easier than the mailing list.

Basicall what I found:

  • BUG: _Boost_INCLUDEDIR/LIBRARYDIR are not unset for interface targets
    • Moved down to where it was used and removed right after
  • lot's of manual unset for "temporary" variables.
    • Used function which adds a scope. To set "persistent" variables one needs to explicitely use set(... PARENT_SCOPE). I believe this makes it much more clear what is used where and passed back
  • Code duplication in generated files and lot's of (mostly) static CMake code embedded in the bjam files making maintenance harder
    • Factor out the search for variants into own function in own file.
    • Interface could probably be even further reduced to 1 parameter if the 2nd is always boost_<name>, but I wasn't entirely sure. Could you try to validate the behaviour especially for the (versioned) python stuff? I think there is just boost_python and a Boost_PYTHON_VERSION variable, isn't it?

On interface targets _Boost_INCLUDEDIR/LIBRARYDIR were not `unset`
Limit scope of them and remove _Boost_CMAKEDIR
Better encapsulation of common behaviour
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant