Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-9858

investigate lsst_build using shallow git clones

    XMLWordPrintable

    Details

    • Type: Story
    • Status: To Do
    • Resolution: Unresolved
    • Fix Version/s: None
    • Component/s: Continuous Integration
    • Labels:
      None

      Description

      Using shallow git clones may improve the performance of the prepare subcommand.

        Attachments

          Issue Links

            Activity

            Hide
            tjenness Tim Jenness added a comment -

            I've reassigned this ticket to Architecture. Brian Van Klaveren given your recent work on lsst_build can you please comment on the status of this ticket?

            Show
            tjenness Tim Jenness added a comment - I've reassigned this ticket to Architecture. Brian Van Klaveren given your recent work on lsst_build can you please comment on the status of this ticket?
            Hide
            bvan Brian Van Klaveren added a comment -

            This is definitely possible after RFC-714 (it was not possible before) - it's only about 30 minutes of work

            The only questions I would care about in implementing this today would be:

            1. What should the default checkout/prepare behavior(s) be (sometimes they are different in jenkins vs. developer workflows)
            2. What are the command-line flags

            (2) is leads to a bigger question because prepare has a lot of flags already, and we also tend to hide CLI complexity behind rebuild rather in lsstsw than using a config file or something for lsst_build

            Show
            bvan Brian Van Klaveren added a comment - This is definitely possible after RFC-714 (it was not possible before) - it's only about 30 minutes of work The only questions I would care about in implementing this today would be: 1. What should the default checkout/prepare behavior(s) be (sometimes they are different in jenkins vs. developer workflows) 2. What are the command-line flags (2) is leads to a bigger question because prepare has a lot of flags already, and we also tend to hide CLI complexity behind rebuild rather in lsstsw than using a config file or something for lsst_build
            Hide
            tjenness Tim Jenness added a comment -

            I'd be inclined to say that the default should stay the same but jenkins can use a command-line switch (it's much easier to change jenkins).

            Show
            tjenness Tim Jenness added a comment - I'd be inclined to say that the default should stay the same but jenkins can use a command-line switch (it's much easier to change jenkins).
            Hide
            bvan Brian Van Klaveren added a comment -

            I tried a naive approach to this - just using "--depth=1" in the git clone commands. Because we do a few git operations (A clone and a fetch with some hacking for git-lfs credential workaround) the naive approach wasn't any faster. I will have to see what we can do to improve the situation.

            Show
            bvan Brian Van Klaveren added a comment - I tried a naive approach to this - just using "--depth=1" in the git clone commands. Because we do a few git operations (A clone and a fetch with some hacking for git-lfs credential workaround) the naive approach wasn't any faster. I will have to see what we can do to improve the situation.

              People

              Assignee:
              bvan Brian Van Klaveren
              Reporter:
              jhoblitt Joshua Hoblitt
              Watchers:
              Brian Van Klaveren, Joshua Hoblitt, Tim Jenness
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

                Dates

                Created:
                Updated:

                  Jenkins

                  No builds found.