On Jun 28, 2012, at 1:27 PM, Alexey Tourbin wrote:

> 
> The downside is, of course, that when a dependency R subset P is
> broken, it is not easy to find out which P symbols were deleted or
> renamed (or which R symbols are missing).  But this is largely a
> developer's, or should I say a hacker's, problem.
> 

Calculating set union/disjunction (for Bloom filters, surely almost as
easy with set:versions) for P/R "sets"
        union           P | R
        intersection    P & R
is natural/rapid/easy even if the symbols are unknown.

FWIW: the "arbitrary tags" implemented @rpm5.org have a similar
one-way lossy mapping where the string representation of the
tag name cannot be unmapped from the 30 bit (truncated SHA1) tag number.

The lack of mapping invertibility matters little to "batch oriented" installer 
usage cases
in RPM libraries even if "enquiring minds" wish concise/complete/informative 
error
messages for diagnosis of problems.

Error messages (and symbols) are overrated and less useful with moderately
sized token dictionaries imho.

I hope to get around to generating set:versions this weekend. I was almost there
10 days ago, but got side-tracked by 1 week's worth of cable modem issues which
has now led to postponed backup/upgrade chores … almost finished.

73 de Jeff______________________________________________________________________
RPM Package Manager                                    http://rpm5.org
Developer Communication List                        rpm-devel@rpm5.org

Reply via email to