Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-15151

Set symbol visibility to hidden in pybind11 wrappers

    XMLWordPrintable

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      Hiding symbols in the pybind11 wrapped modules (potentially) reduces the build size and enables importing modules wrapped with different pybind11 versions in the same session.

        Attachments

        1. importCls.py
          2 kB
        2. pb_cmake.zip
          1.88 MB

          Issue Links

          is triggered by

          Improvement - An improvement or enhancement to an existing feature or task. DM-14828 Upgrade pybind11 to 2.2.3

          • Done

          Improvement - An improvement or enhancement to an existing feature or task. DM-15599 Stop using boost persistence in afw

          • Done
          relates to

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-15789 Fix symbol visibility warnings in ndarray pybind11 converters

          • Done

          Improvement - An improvement or enhancement to an existing feature or task. DM-15790 Compile pybind11 with hidden symbol visibility on linux

          • Done

            Activity

            Ascending order - Click to sort in descending order
            14 older comments
            Hide
            Permalink
            rowen Russell Owen added a comment -

            The afw pull request is here: https://github.com/lsst/afw/pull/387

            Show
            rowen Russell Owen added a comment - The afw pull request is here: https://github.com/lsst/afw/pull/387
            Hide
            Permalink
            rowen Russell Owen added a comment - - edited

            I confess that boost persistence is mostly broken by this on macOS (in afw the formatters for image-like classes do still work, though we plan to stop using those shortly). Perhaps I should remove it as part of this ticket. My plan was to do that on DM-15599 (which I plan to start work on shortly) as it's more directly related to that ticket. Still...it might be safer to remove the formatters that are not for image-like objects on this ticket, or export the symbols that might make them work (they aren't tested or used anywhere).

            Also I will say again that afw was changed but the pull request doesn't show up on JIRA. See the link in the previous comment.

            Show
            rowen Russell Owen added a comment - - edited I confess that boost persistence is mostly broken by this on macOS (in afw the formatters for image-like classes do still work, though we plan to stop using those shortly). Perhaps I should remove it as part of this ticket. My plan was to do that on DM-15599 (which I plan to start work on shortly) as it's more directly related to that ticket. Still...it might be safer to remove the formatters that are not for image-like objects on this ticket, or export the symbols that might make them work (they aren't tested or used anywhere). Also I will say again that afw was changed but the pull request doesn't show up on JIRA. See the link in the previous comment.
            Hide
            Permalink
            krzys Krzysztof Findeisen added a comment - - edited

            Review complete (though I didn't look at the question of which classes should or should not have LSST_EXPORT), minor comments on GitHub.

            Show
            krzys Krzysztof Findeisen added a comment - - edited Review complete (though I didn't look at the question of which classes should or should not have LSST_EXPORT ), minor comments on GitHub.
            Hide
            Permalink
            krzys Krzysztof Findeisen added a comment - - edited

            Actually, wait: I just noticed that daf_persistence has a branch but no pull request. I assume that was intentional, but please let me know if it wasn't.

            Show
            krzys Krzysztof Findeisen added a comment - - edited Actually, wait: I just noticed that daf_persistence has a branch but no pull request. I assume that was intentional, but please let me know if it wasn't.
            Hide
            Permalink
            rowen Russell Owen added a comment - - edited

            The existence of a branch for daf_persistence is an accident. It has no changes on it but I can't get rid of it because of lsstsw limitations.

            Show
            rowen Russell Owen added a comment - - edited The existence of a branch for daf_persistence is an accident. It has no changes on it but I can't get rid of it because of lsstsw limitations.

              People

              Assignee:
              rowen Russell Owen
              Reporter:
              pschella Pim Schellart [X] (Inactive)
              Reviewers:
              Krzysztof Findeisen
              Watchers:
              Jim Bosch, John Swinbank, Krzysztof Findeisen, Pim Schellart [X] (Inactive), Russell Owen
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Dates

                Created:
                Updated:
                Resolved:

                  Jenkins

                  No builds found.