Hi Alexander,

Il giorno ven, 14/01/2011 alle 16.28 +0100, Alexander Wagner ha
scritto: 
> > Great job for using ^M character!
> 
> I admit I was about to suggest to allow also for another
> character in the role definition for easier handling. ;)

That is a good one, although it would make the parsing of firerole rules
a bit more complex (since you would need to properly understand that
additional character when it happens e.g. inside group-name-based
authorizations) and firerole rules are designed to be extremely fast to
be interpreted and applied. I would take this into consideration for
further enhancements.

> If I may, I would suggest to add this to the --help of
> bibdocfile. Probably, also with some example. As a newbe
> you'll most likely stumble upon such things.

Sure, you arere right!

> This results in something like:
> 
> [...]
>     106:95:::doctype=Main
>     106:95:::status=firerole: ALLOW FROM "2011-01-01"
>     ALLOW ALL
> 
>     106:95:::basedir=/opt/invenio/var/data/files/g0/95
>     106:95:::creation date=2011-01-14 15:12:58
> [...]
> 
> If I get it correctly, in the second line "status=" should
> have changed to "firerole=", right? 

Nope. Status is just a generic attribute of a bibdoc. So when the status
is "firerole: foo", then the status is interpreted as firerole rule. If
the status was "DELETED" then the file would have been considered as
deleted...

> Additionally, I still
> have the problem, that the file as such is locked, though
> the clock is clearly beyond 2011-01-01. Thus if I try to
> download the file I'm redirected to the login page and once
> I have become John Doe wiht out any rights it tells me this
> file is locked.

That is strange. I should investigate this and come back to you ASAP.

> I wonder, however, if it wouldn't make sense to rethink fft
> subfield. It seems sensible that firerole definitions for a
> given file are accessible from the MARC editor and thus to
> store them in some MARC field for easy handling. Also for
> showing the embargo time on the details page and easy
> transport of the records between two systems seem to hint at
> this. What do you think?

There has been indeed some internal talking on putting both record-level
and document-level restriction in the MARC although it will take a bit
before it will made into Invenio in a stable way. We should definitively
bring that thread into this mailing list...

Cheers,
Sam
-- 
Samuele Kaplun
Invenio Developer ** <http://invenio-software.org/>

Reply via email to