shibboleth and wget/curl

[Peter Schober]

* Nathaniel Watson <nathankw at stanford.edu> [2015-03-04 02:03]:
> The scenario I have in mind is where a single HTML file has many
> URLs that I'd like to follow with wget, instead of click on them one
> by one to download them ...

For a one-time thing with manual messing about you could copy the
cookies into a curl cookiejar. To bootstrap the creation of one:
$ curl -L -c /tmp/cookiejar https://wiki.shibboleth.net/confluence

Next duplicate the last line in the file /tmp/cookiejar and adapt to
your SPs specifics (fqdn, path, cookie name, cookie value), leaving
all the tab chars in place as field seperators.

Then call the URL again using the modified cookiejar and get access:
$ curl -b /tmp/cookiejar https://your.example.org/secure/

The SAML SP might have additional protection in place, though, much
like the Shibboleth SP has (consistentAddress, checkAddress), so curl
should be called from the same machine you initiated the session with
before to be safe.

Finally, a more general and user friendly way to achieve that is
probably using a browser extension. A quick search for "download all
files browser extension" got me plenty of results that claim to offer
easy ways to do just that (e.g. DownThemAll for Firefox, Download
Master for Chrome, etc.). I haven't used/tested any of those myself.
-peter