General question about policy ACLs

[Matt Snyder]

Hi all, I have a couple questions about policy ACLs. The first
question is, does this scenario work?
1. I crawl some unsecured web content
2. I add some ACLs with some of those unsecured urls

Are the unsecured documents suddenly made secure by the GSA? Or do
the documents need to be flagged as secure during the crawl, in order
for the ACLs to be applied?

[brianb]

If the URLs are actually made public on the webserver, the GSA will
think that they are public (not really sure why you would want to
block something that was actually public). You may be able to trick
the GSA into thinking that they are secure by adding a fake forms auth
rule and making sure the Make Public checkbox is not checked. The GSA
will then try to execute that bogus rule *before* it crawls and mark
the document as secure even if it is not. This will not work if you
use the Crawler Access settings since these patterns are not used
until after the GSA first tries to crawl them (needed for the initial
auth challenge).

Hope this helps.

Brian

[Matt Snyder]

Got it, good suggestion, I'll try it. My next question is, how do you
use ACLs with content that is fed with Incremental or Full feed types,
and/or database feeds?

[brianb]

Hi Matt,

For now, you would need to use the API here:

http://code.google.com/apis/searchappliance/documentation/62/gdata/acapi_policy_acls.html

or you could do it manually in the UI but that would probably be a
pain. :-(

Brian

[Matt Snyder]

Thanks! What I am wondering is, if I feed in content using a FULL
feed, then I think it starts out as public, right? Or is there a flag
in the feed API to specify the content should be secure?

If the content starts out as public, then how does the GSA know to
make the content secure and to apply ACLs? Like my first question,
basically, except the GSA is not crawling the content that comes in
via a FULL feed, so I don't think the forms auth trick would help.