• Peter Wu's avatar
    Always use html2text.py for FAQ, improve output · 68698db8
    Peter Wu authored
    A recent commit broke compilation with Python 3. The original author of
    html2text.py is deceased and the fork has increased the number of files
    for this "simple" helper.
    
    The html2text.py script in this patch was rewritten and its output
    matches with lynx (except for a few newlines around lists). This means
    that indentation has been added for headings, paragraphs and lists.
    Also, since it was written from scratch, a new license could be chosen
    that matches Wireshark.
    
    Since now the in-tree html2text.py script provides nicer output, remove
    detection of the alternative programs (elinks, links). lynx/w3m is
    somehow still necessary for asciidoc though.
    
    (I also looked into reusing html2text.py for the release notes to
    replace asciidoc, but the --format=html output produces different output
    (HTML adds a ToC and section numbers). For now still require lynx for
    release notes)
    
    Tested with Python 2.6.6, 2.7.9, 3.2.6 and 3.4.3 under LC_ALL=C and
    LC_ALL=en_US.UTF-8 on Linux. Tested reading from stdin and file, writing
    to file, pipe and tty. Tested with cmake (Ninja) and autotools on Arch
    Linux x86_64. Test:
    
        # For each $PATH per python version, execute (with varying LC_ALL)
        help/faq.py -b | tools/html2text.py /dev/stdin | md5sum
        help/faq.py -b | tools/html2text.py | md5sum
        help/faq.py -b | tools/html2text.py
        help/faq.py -b | tools/html2text.py >/dev/null
    
    Change-Id: I6409450a3e6c8b010ca082251f9db7358b0cc2fd
    Reviewed-on: https://code.wireshark.org/review/7779
    Petri-Dish: Peter Wu <peter@lekensteyn.nl>
    Tested-by: 's avatarPetri Dish Buildbot <buildbot-no-reply@wireshark.org>
    Reviewed-by: 's avatarAnders Broman <a.broman58@gmail.com>
    68698db8