#954 Merged
Repository
ngoldbaum
Branch
yt-3.0
Repository
yt_analysis
Branch
yt-3.0
Author
  1. Nathan Goldbaum
Reviewers
Description

For more information, see this e-mail thread.

Unfortunately the diff is pretty enormous and bitbucket doesn't render it. To partially fix that, I'm serving a clone of my repository on my website. The full commit diff is here:

http://ngoldbaum.net/hg/hgweb.cgi/yt/rev/67507b4f8da9

I've tried to eliminate all uses of pf, pfs, and pf.h, replacing them with ds, datasets, or ds.index where appropriate. I've also replaced parameter_file with dataset, preserving it only when it refers specifically to a file that contains parameters for a dataset.

This passes the unit tests on my laptop. We should probably wait until we have answer testing set up for all of the frontends and the PR backlog has cleared a bit to merge this in.

Comments (18)

  1. MattT

    So, this has languished long enough. If you merge it and bring it up to date, then I think it can go in after these outstanding pull requests, and before any other ones:

    • 1001
    • 1016
    • 993
    • 996
    • 955 (maybe)
    • 1015
    • 997
    • 1009

    Sound good?

    1. Nathan Goldbaum author

      Yes, thanks for the concrete advice.

      I will merge once again within the next day or two.

      1. Nathan Goldbaum author

        Only 1001 and 1016 are left to merge. I will try my best to merge this PR with tip tonight.

        Additionally, I will create a compatibility layer for data_source.pf and include that in a separate PR that depends on this one.

    1. Nathan Goldbaum author

      I'm going to merge once more to catch any issues created this morning. If the tests pass at that point, this is ready to go in. Let's not merge any more PRs until then.