Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-21989

Package a new version of CFITSIO ( 3.47 / 3470 )

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: DM
    • Labels:
      None
    • Sprint:
      Arch 2019-10-21, Arch 2019-11-04, Arch 2019-11-11, Arch 2019-11-18, Arch 2019-11-25, Arch 2019-12-02, Arch 2019-12-09, Arch 2019-12-16
    • Team:
      Architecture

      Description

      Package a new version of CFITSIO. This will definitely cause some tests to fail and FITS serialization code in afw (and other packages) may need to be updated, especially metadata packages.

        Attachments

          Issue Links

            Activity

            Hide
            bvan Brian Van Klaveren added a comment -

            This was solved in conda packaging and various other tickets.

            Show
            bvan Brian Van Klaveren added a comment - This was solved in conda packaging and various other tickets.
            Hide
            bvan Brian Van Klaveren added a comment -

            DM-24376 was merged, and I think RFC-686 will be adopted. Soon after RFC-686 is adopted, I think we can merge this.

            A build with DM-24376 and this ticket was performed this morning:
            https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/31615/pipeline

            Show
            bvan Brian Van Klaveren added a comment - DM-24376 was merged, and I think RFC-686 will be adopted. Soon after RFC-686 is adopted, I think we can merge this. A build with DM-24376 and this ticket was performed this morning: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/31615/pipeline
            Hide
            swinbank John Swinbank added a comment -

            I see all that, but at the same time this seems like a breaking and confusing, albeit minor, interface change. It may be that it's sufficiently minor that nobody will care, but I'd be more comfortable if we could reach consensus on that — at least with a discussion at the CCB, and possibly though an RFC — before merging.

            Show
            swinbank John Swinbank added a comment - I see all that, but at the same time this seems like a breaking and confusing, albeit minor, interface change. It may be that it's sufficiently minor that nobody will care, but I'd be more comfortable if we could reach consensus on that — at least with a discussion at the CCB, and possibly though an RFC — before merging.
            Hide
            tjenness Tim Jenness added a comment -

            We can't keep using the old cfitsio for ever so we options are only allow upper case in PropertyList, refuse to write PropertyList entries that have lower cased keys, or somehow try to encode the real key in the FITS file. The current motivation for this ticket is switching to conda-forge since our eups cfitsio is really old.

            Show
            tjenness Tim Jenness added a comment - We can't keep using the old cfitsio for ever so we options are only allow upper case in PropertyList, refuse to write PropertyList entries that have lower cased keys, or somehow try to encode the real key in the FITS file. The current motivation for this ticket is switching to conda-forge since our eups cfitsio is really old.
            Hide
            swinbank John Swinbank added a comment - - edited

            I don't want to get into a philosophical debate... but it seems like a substantial change to go from “PropertyList transparently round-trips through FITS” to “You can save a PropertyList to FITS, but when you load it again you won't get back exactly what you stored”. Maybe that's fine, but it feels like it's RFC worthy...?

            Show
            swinbank John Swinbank added a comment - - edited I don't want to get into a philosophical debate... but it seems like a substantial change to go from “ PropertyList transparently round-trips through FITS” to “You can save a PropertyList to FITS, but when you load it again you won't get back exactly what you stored”. Maybe that's fine, but it feels like it's RFC worthy...?

              People

              • Assignee:
                bvan Brian Van Klaveren
                Reporter:
                bvan Brian Van Klaveren
                Reviewers:
                John Swinbank
                Watchers:
                Brian Van Klaveren, John Swinbank, Kian-Tat Lim, Tim Jenness
              • Votes:
                0 Vote for this issue
                Watchers:
                4 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved:

                  Summary Panel