Re: silva-1.1-cvs-head: ghost

Marc Petitmermet wrote:
> dear developers
> 
> "This ghost is broken" message
> ------------------------------
> when the original of a silva document is not published a ghost pointing 
> to it says "This ghost is broken" in the preview tab or the public 
> preview. as soon as the original is published, the ghost isn't broken 
> any more and the previews work as expected. i find it rather 
> inconvenient to have to publish a document only to be able to test a 
> ghost especially when the site is online...

A ghost cannot know whether you have a right to access an unpublished 
version of another document. This is why ghosts *only* access the 
published version of a document ever. Opening this up would create a 
security hole, in that one author would be able to check the unpublished 
   version of a document that this author has no right to see, as long 
as the author knows the URL of this document.

I understand that this is inconvenient. The minimum thing we could do is 
make them say 'this ghost is pointing to an object that has no published 
version' or something like that. More complicated would be to verify 
whether the person currently previewing the ghost actually has author 
access to the original anyway, and show it a version in that case.

> author: unknown user
> --------------------
> when creating a ghost of a silva document the author is "unknown user" 
> in the contents tab although the creator and author information is 
> correct in the properties tab of the ghost.

This looks like a bug we should look into. Does this only happen with 
ghosts?

> modified: n/a
> -------------
> the last modified date of a ghost of a silva document is shown as "n/a" 
> in the contents tab. but why is there a modification date in the 
> properties tab? i think that it is correct to also show the modification 
> date in the content tab because an author could change the reference in 
> a ghost and this changes also the modification date, doesn't it? at 
> least in the properties tab...

We should look into that one as well.

Thanks,

Martijn