Handles the differences between full tarball, lite tarball, and cloned sources.
Removes the skip_unused and sysroot flags from _extraction.py and downloads.py.
When downloading the official Chromium tar file, axel is no longer able
to parallelize the connections nor detect the filesize. It is likely
Google added a specific rule for axel since curl and Chromium work
normally. While it is possible to circumvent this rule, Google might
have imposed it to reduce spikes of heavy load on their servers. We
should respect this and just use curl instead, which only uses one
connection and has the ability to resume interrupted downloads.
Downloading resumption is done by downloading files with ".partial"
extensions, then renaming them after they're done.
The urllib downloader should probably be replaced with requests, but the
CLI downloaders are better anyway. Also, support for aria2 and wget is
missing.
Fixes#996
The downloader does not handle network interruptions well (e.g. a
connection can hang indefinitely and the downloader will not terminate).
This message makes it easier to manually download it using a more
advanced downloader like axel, which supports parallel downloading and
download resumption.
