#4284 Update Boost to 1.52


Successfully built with fink -m rebuild

1 Attachments


  • Hanspeter Niederstrasser

    Seems fine. I modified the Conflicts and Replaces fields to include all the existing boost header packages (including %n), as that makes it easier to copy the fields between pkgs and makes it less likely that a package might be skipped during updates.

    For the 10.7 tree, can it be updated to use system-python2.7?

  • Hanspeter Niederstrasser

    Actually, it will need to move to system-python2.7 because "-framework Python" picks up the 2.7 version of the Python framework (on 10.7). This also affects boost1.41 and 1.46.1, which can be dealt with at the same time this is checked in.

  • Abdulla Kamar

    Abdulla Kamar - 2013-06-06

    Sounds good. Should I make the changes, or will you be doing that?

  • Hanspeter Niederstrasser

    They're simple string changes, so I can do them all at once here.

  • Hanspeter Niederstrasser

    On a related unrelated note: The non-cmake flavor of boost allows picking and choosing which libraries to build. Is it possible to do the same with boost.cmake, for example, if we wanted to have the python version be built separately?

  • Abdulla Kamar

    Abdulla Kamar - 2013-06-07

    It is possible, you can specify a list of libraries to build. I just assumed the default of building all, using the system libraries, would be sufficient.

  • Hanspeter Niederstrasser

    99% of the time it is. The issue would be for just the boost-python library. Some things don't care which python it uses, others care very much; which is why Fink's non-cmake boost can build boost-python for any python flavor (system, 25, 26, 27...). The flexibility is nice, but it's also why it's stuck at 1.35: no one has had the time to deal with boost's weird build system.

    This isn't something that will affect this submission unless you want to try it. Just thinking about future possibilities.

  • Hanspeter Niederstrasser

    Sorry about the weekend delay. Was waiting for wgscott to get back to see if rdkit could use this boost (it can't, so it's no longer a blocker). But he did ask why you were using github to host the source vs grabbing it from the official upstream source.

    Apparently he gets this error during download:

    Unable to locally verify the issuer's authority.
    To connect to github.com insecurely, use `--no-check-certificate'.
  • Abdulla Kamar

    Abdulla Kamar - 2013-06-12

    I'm using github for 2 reasons:
    - The upstream version uses Jam, and this version uses CMake
    - This is a forked version of what's available at https://github.com/pocb/boost/tree/cmake-1.52.0, I needed to do this to make it compile correctly (I kept getting linker failures with the original)

  • Hanspeter Niederstrasser

    • labels: --> New Package
  • Hanspeter Niederstrasser

    By 'upstream', I meant the upstream boost.cmake providers. I understand that the boost.jam and boost.cmake groups are different.

    I compared your tarball to the one at https://github.com/pocb/boost/tree/cmake-1.52.0 and your changes seem pretty straightforward. Fink policy is normally to use the official (from pocb in this case) tarball and then use PatchFile for changes. If this is really the extent of the changes you needed for it to properly build, then please in the future, use a .patch rather than hosting a modified tarball.

    I'll check this in as is now so that this can go live.

    For the record: changes in CVS from this submission:

    1. Add C/R for all boosts, including the new 1.53
    2. Update to system-python2.7

    Thanks for the new pkg.

  • Hanspeter Niederstrasser

    • status: open --> closed-accepted
    • assigned_to: Hanspeter Niederstrasser
    • Group: Undergoing_Validation --> Added_to_Fink
  • Hanspeter Niederstrasser

    Now actually checked in.


Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

No, thanks