|Re: Undo in Multiuser Environment [message #1735884 is a reply to message #1735878]
||Thu, 23 June 2016 12:42
| Ed Willink
Registered: July 2009
I think it is part of the problem of maintaining the history as a
sequence of semantic primitives. Then you undo by do-ing an inverse
semantic change, not by restoring an earlier context. Successfully
do-ing the inverse semantic change must be heavily predicated by a
consistent starting state; you cannot undo a move to something that has
been refactored in the meantime.
On 23/06/2016 13:22, Christian W. Damus wrote:
> Hi, Thorsten,
> That is exactly what I would suggest. It is probably not worth the
> risk missing something in an analysis of the change-description of
> every command on the undo stack to see whether it intersects with an
> incoming delta notification. Besides that the API to access those
> change descriptions doesn't even exist (and some commands may not have
> change descriptions to look at).
> In my experience, I don't think I've ever seen an on-line
> transactional application that had a meaningful "undo" support (except
> for local uncommitted edits in text fields and the like).
> On 2016-06-23 07:16:19 +0000, Thorsten Schlathölter said:
>> Hi, I would like to raise the question on how to deal with undo if
>> papyrus is used in a multi user environment as it is provided with
>> the CDO integration.
>> I cannot think of any other solution than to clear the undo\redo
>> history if delta notifications from other CDOTransactions come in.
>> Are there any ideas?
Powered by FUDForum
. Page generated in 0.02011 seconds