* 'affect' a bit(j) in the output if two inputs, identical but for bit(i), will differ at output
   * bit(j) about half the time
   *
   * <p>Funneling is pretty simple to detect. The key idea is to find example keys which
   * unequivocally demonstrate that funneling cannot be occurring. This is done bit-by-bit. For each
   * input bit(i) and output bit(j), two pairs of keys must be found with all bits identical except

    without any of its other classes -- but is also available in a second
    "version" that omits the class to avoid conflicts with the copy in Guava
    itself. The idea is:

    - If users want only ListenableFuture, they depend on listenablefuture-1.0.

    - If users want all of Guava, they depend on guava, which, as of Guava
    27.0, depends on

   * 'affect' a bit(j) in the output if two inputs, identical but for bit(i), will differ at output
   * bit(j) about half the time
   *
   * <p>Funneling is pretty simple to detect. The key idea is to find example keys which
   * unequivocally demonstrate that funneling cannot be occurring. This is done bit-by-bit. For each
   * input bit(i) and output bit(j), two pairs of keys must be found with all bits identical except

            like to get an error. Currently we do, but if we add the extra <inherits> lines, we
            won't.

            I have one idea for a better approach, but it's painful, and I haven't tested it: We
            could postprocess Collect.gwt.xml to add <skip> lines for all the files that should be