Re: [Jabref-users] Groups using not possible

[Haines B. <ha...@hi...>]

On Sun, Jun 12, 2016 at 02:22:30PM +0200, Bernhard Kleine wrote:
> Am 12.06.2016 um 14:18 schrieb Haines Brown:
> > I've complained several times in another thread, so far no response,
> > that JabRef 3.4 destroyed the @Comment{jabref-meta: groupstree: section
> > of my .bib databases. The group tree is present, but has lost all
> > keywords.
> >
> > I don't see how this could have happened through my blundering. I can
> > recover that section from all my databases from a backup when still
> > running JabRef 9.10 two weeks ago. I did this with a database, but a
> > major problem remains. If I create a new group, it is not added to the
> > groupstree, but is recorded somewhere else. Also, creating a link from
> > an entry to a group creates a link to all groups with the same name.
> >
> > Please!!! Someone tell me what's going on. Has a new group mechanism
> > been created? Where is that information being stored? Was it supposed to
> > recover groupstree data from the .bib databases?
> >
> > I'll have to create links from entries to groups by entering the
> > information into the databases by hand, but I worry that JabRef 3.4 will
> > again destroy that information.
> >
> > Haines Brown

> I know that there were changes in the group function. But I do not know
> whether your problem is related. My groups at least were maintained when
> I updated jabref to 3.4.
> Regards
> Bernhard

I really would like to know what this new groups mechanism is. CHANGELOG
suggests that with JabRef 3.3 the data held in the jabref-meta:
groupstree: sections of bibliographic databases was transferred to
something called "BibEntry". I find it succeeded for some databases, but
not all. 

I see that JabRef now generates a host of files under
/usr/local/share/JabRef/jabref-master/... and apparently overwriting
them are user's preferences in ~/Xjabref-master/.... Unfortunately I was
unable to locate a BibEntry there. There are MSBibEntry.java files, but
these hold no specific group information.

So my questions are:

a) What and where is this "BibEntry"? It apparently is not the name of a
file.

b) Was it intended that this BibEntry preserve the original groups
hierarchy? It does not seem to distinguish groups organized in
hierarchic fashion. Even where I've restored the group tree
functionality, groups with same name but situated in different places in
the tree are not distinguished. Is BibEntry just a flat file database?

c) How could all group data be lost when I moved up from JabRef 9.10 to
JabRef 3.4? Is there something I should have done to prevent it? While I
can recover old group hierarchies data, I don't want to loose recent
additions again.

Haines Brown