Re: Splitting a rev list into 2 sets

Francis Moreau wrote:

I don't fully understand your query, because almost anything is possible with rev-list:

  $ git rev-list foo..bar master # reachable from master, bar, not foo

What I _suspect_ you're asking is for help when you can't construct this "foo..bar master" programmatically (or when you cannot express your criterion as arguments to rev-list). You want an initial commit set, and filter it at various points in your program using various criteria, right? In that case, I'd suggest something like this:

    # Returns a list of commits given a committish that `rev-list`
    # accepts.
    def self.list_commits(committish)
        commits = []
        revlist = execute("git", "rev-list", "--reverse", "--date-order",
                          "--simplify-merges", committish).chomp.split("\n")

        # do it in batches of 1000 commits
        while revlist
            these_revs = revlist.first(1000).join("\n")
            this_chunk = execute({ :in => these_revs }, "git",
                               "cat-file", "--batch")

            # parse_cat_file parses the chunk and updates @commit_index
            parse_cat_file(this_chunk) { |struct| commits << struct }

            revlist = revlist[1000 .. revlist.length - 1]
        end
        return commits
    end

    # Filters a list of commits with the precondition that it exists
    # in the committish.  :sha1 is used to uniquely identify a commit.
    def self.filter_commits(commits, committish)
        revlist = execute("git", "rev-list", "--simplify-merges",
                                 committish).split("\n")
        allowed_commits = revlist.map { |sha1| @commit_index[sha1.hex] }
        return commits & allowed_commits
    end

In essence, I use '&' to filter and it's extremely fast. The trick is to shell out to git sparingly, store the data you get in a sensible manner, and build fast custom filters based on what you want. Here are a few more examples:

    # Filters a list of commits with the precondition that it is a
    # first-parent commit in a given committish.
    def self.filter_fp_commits(commits, committish)
        revlist = execute("git", "rev-list", "--first-parent",
                          "--simplify-merges", committish).split("\n")
        allowed_commits = revlist.map { |sha1| @commit_index[sha1.hex] }
        return commits & allowed_commits
    end

    # Slice a list of commits using a start_hex and end_hex, which
    # may both be nil.
    def self.slice_commits(commits, start_commit, end_commit)
        start_idx = commits.index(start_commit)
        end_idx = commits.index(end_commit)
        start_idx = 0 if start_idx.nil?
        end_idx = commits.size - 1 if end_idx.nil?
        return commits[start_idx..end_idx]
    end

    def self.filter_commits_tree_path(commits, path)
        commit_chunk = (commits.map { |commit| commit.sha1 }).join("\n")
        commit_chunk = "#{commit_chunk}\n"
        diff_tree_chunk = execute({ :in => commit_chunk }, "git", "diff-tree", \
                                  "-m", "-r", "-s", "--stdin", path)
        matching_sha1s = diff_tree_chunk.split("\n")
        allowed_commits = matching_sha1s.map { |sha1| @commit_index[sha1.hex] }
        return commits & allowed_commits
    end

Did that help?