[Bug-wget] How to exclude link?

[Peng Yu]

Suppose I need to excluded links that start with xyz with the
following option. However,

  wget --mirror \
    --limit-rate=400k \
    --no-parent \
    --no-host-directories \
    --wait=5 \
    --random-wait \
    --convert-links \
    --cut-dirs=2 \
    --reject "xyz*" \
    --directory-prefix=directory_prefix \
    -o logs/directory_prefix.log \
    some.web.com/xyz_want_to_download

Wget still download the files that match the pattern 'xyz*'. But it
deletes them after they are downloaded. See the output below from
wget.

Removing xyz_link_name since it should be rejected.

I'm wondering what should be the way to not to download the link
starting with 'xyz*' at all.

Also I need to keep the start link some.web.com/xyz_want_to_download.
However, wget delete them as well. Is there a way to keep it?

-- 
Regards,
Peng