I've attached a unified diff file containing a new filter
I'd like to include with gramps.
The new filter is called RelationshipPathBetweenBookmarks.
The file also contains a small documentation update in
The reasons for this update are:
1. I couldn't get RelationshipPathBetween to work right.
Sometimes, it would use an ever-increasing amount of
virtual memory, and thrash the disk for as long as I
allowed it to continue. So, I wrote a new filter.
2. With the new filter, you don't have to hard-code
individual ID's into the filter.
3. With the new filter, you can specify any number of
bookmarks, and it will come up with a reasonable collection
of individuals to connect them together.
I have tested this fairly well. It seems to work.
-- Wayne Bergeron
The keenest sorrow is to recognize ourselves
as the sole cause of all our adversities.
Sorry for a very long time to reply.
On Mon, 2006-05-22 at 01:08 -0500, wb wrote:
> I've attached a unified diff file containing a new filter
> I'd like to include with gramps.
> The new filter is called RelationshipPathBetweenBookmarks.
I have committed your patch, thanks a lot!
> 3. With the new filter, you can specify any number of
> bookmarks, and it will come up with a reasonable collection
> of individuals to connect them together.
A question then: if I have many people (more than 2) in
bookmarks and they are not related, what would the filter
match? All ancestors of all bookmarked people?
On Thursday June 22 2006 14:03, Alex Roitman wrote:
> A question then: if I have many people (more than 2) in
> bookmarks and they are not related, what would the filter
> match? All ancestors of all bookmarked people?
Thanks for processing the filter into the codebase.
No, the filter is really pretty smart about edge conditions.
If you find it misbehaving, I'd like to know.
If a bookmarked person is unrelated to other bookmarked
people, he will be returned without any relatives.
So, with enough bookmarks, you can get some interesting
relationship graphs -- unrelated people,
ancestor/descendant chains, multiple cousin relationship
graphs, and so on. All connected, or multiple unconnected
graphs. Graphviz prints them well.
I wasn't able to make up a condition which it could not
process gracefully. Of course, Murphy's Law predicts that
someone will produce such a condition immediately after you
release the software.