- subprocess.call is much more common and easier to use
- use shlex to split to take care of quoted spaces

This is a copy and paste of videood. We should
factorize it in the future

Since the filename generation is based on the video.ext
field, we have to fill it before to check if the file
is downloaded.

Also do a little code factorisation

For windows compatibility

autopep8 -a -r -i --select E711,E712,W601,W602,W603,W604,W690,E304,E401,E502 .
Also includes some manual reindentations (many are left after the print()
changes).
Manually checked, some modernizations not commited here.

python modernize.py --no-six -f libmodernize.fixes.fix_print -w

With manual fixes as the import was put always on top.

Diff checked manually.

autopep8 -a -ir -j2 --select=E711 .

This required the "agressive" option because it can change code
behavior.

Remove useless trailing \
Remove trailing spaces
Add missing empty lines

autopep8 -ir -j2 --select=E301,E302,E502,W291,W293,W391 .

Diff quickly checked.

autopep8 -ir -j2 --select=E303,E125
Diff checked manually.

And fix configuration example since it is case sensitive

Signed-off-by: Fabien Grumelard <fabix-symlink@grumi.net>

- when a video is downloaded, create immediately a symlink
- if a symlink is removed, empty the downloaded file to free space on disk

Adding -e to rtmpdump.

So instead of:
    rtmpdump -r <url> -o <file>
we will have
    rtmpdump -e -r <url> -o <file>

If <file> doesn't exist, it will be created. So the -e switch doesn't
change anything.

If <file> alredy exists, then rtmpdump will try to complete a previous
download.

That way, if videoob download <id> fails, running the command again
will complete the download without starting again from scratch.