On Thu, 16 Oct 2014 10:14:53 -0500
Kent Tenney <[email protected]> wrote:

> Could a plugin offer a get_gnx() function?

You mean plugin specific gnxs?  Or let a plugin override gnx form?

Cheers -Terry

> On Thu, Oct 16, 2014 at 10:06 AM, Edward K. Ream
> <[email protected]> wrote:
> > On Thu, Oct 16, 2014 at 8:19 AM, 'Terry Brown' via leo-editor
> >
> >> It's not clear to me that uuids are needed.  #35 is a bug because
> >> Leo is creating clashing gnxs *on load*, which it shouldn't.  Bob's
> >> always incrementing time stamp actually fixed the problem in hand,
> >> the case when things were done within one second.  But we decided
> >> to tackle the more general problem of files from other places in
> >> other timezones.  I think the max. index from post-scan fixes
> >> that.  With max. index from post-scan, there's no known way for
> >> Leo to corrupt a Leo file with clashing gnxs on load.
> >
> > I agree, but Bob reports that the post scan does not fix the
> > problem. Possible causes:
> >
> > - The post scan is buggy.
> > - Caching bypasses the post scan.
> > - (My guess) The hidden root vnode's gnx is the culprit.
> > - (Gasp) Something else.
> >
> > I'll investigate today.
> >
> >> So then the remaining questions relate to two separate files with
> >> clashing gnxs.  So for the following *cross outline* copy / pastes:
> >>
> >>  - I assume simple copy / paste between files assigns new
> >>    gnxs anyway, so no problem there.
> >
> > Correct.
> >
> >>  - Copy-retaining clones, not exactly sure on the intended behavior
> >>    there, but whatever it is I'm sure it can be covered by the
> >> creation of new gnxs as with the simple copy / paste case.
> >
> > paste-retaining-clones *means* not changing gnxs.  I'm willing
> > assume, for now, that gnx clashes in this case are rare enough to
> > ignore.
> >
> >> So if we're not accidentally creating clashing gnxs on load (*bug*
> >> #35), and moving content between files where there could be
> >> clashing is handled by the appropriate API functions, why do we
> >> need uuids?
> >
> > Suppose Tony sends you an outline containing a node *he* created
> > that clashes (same id, same timestamp) with a node *you* created.
> > paste-retaining clones will cause data loss: either your node wins
> > or Tony's node does.  The only way to prevent such clashes is to use
> > uuids.
> >
> > In short, we agree there is no need to rush into uuid's, but I
> > think a --uuid option should be provided.  Another thread will
> > discuss the design of this option.
> >
> > Edward
> >
> > --
> > You received this message because you are subscribed to the Google
> > Groups "leo-editor" group. To unsubscribe from this group and stop
> > receiving emails from it, send an email to
> > [email protected]. To post to this group,
> > send email to [email protected]. Visit this group at
> > http://groups.google.com/group/leo-editor. For more options, visit
> > https://groups.google.com/d/optout.
> 

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/d/optout.

Reply via email to