Re: [Gramps-devel] Reports in CVS

[Alex R. <sh...@al...>]

Mitchell,=20

On Sat, May 15, 2004 at 09:56:12PM -0700, Mitchell Ferschweiler wrote:
>=20
> Here's a fun little error that happens when you run out of space on
> your HD while importing from a GRAMPS package.
>=20
> GRAMPS has encountered an internal error.
> Please copy the message below and post a bug report
> at http://sourceforge.net/projects/gramps or send an
> email message to gra...@li...
>=20
> GRAMPS : 1.1.0-0.CVS20040512
> Python : 2.3.4 candidate
> GTK : 2.4.1
> PyGTK : 2.3.92
> Traceback (most recent call last):
>   File "./plugins/ReadPkg.py", line 113, in on_ok_clicked
>     importData(self.db,dbname,self.progress)
>   File "/home/mitchell/playarea/gramps/gramps2/src/ReadXML.py", line
> 122, in importData
>     except ExpatError, msg:
> NameError: global name 'ExpatError' is not defined

This should be corrected in CVS now. The real error message should
come up in an error dialog.

> and from the terminal:
>=20
> Traceback (most recent call last):
>   File "./plugins/ReadPkg.py", line 126, in on_ok_clicked
>     self.callback()
>   File "/home/mitchell/playarea/gramps/gramps2/src/gramps_main.py",
> line 937, in import_tool_callback
>     self.people_view.build_tree()
>   File "/home/mitchell/playarea/gramps/gramps2/src/PeopleView.py", line
> 122, in build_tree
>     self.person_model =3D PeopleModel.PeopleModel(self.parent.db)
>   File "/home/mitchell/playarea/gramps/gramps2/src/PeopleModel.py",
> line 88, in __init__
>     self.rebuild_data()
>   File "/home/mitchell/playarea/gramps/gramps2/src/PeopleModel.py",
> line 124, in rebuild_data
>     for person_id in self.db.get_person_keys():
>   File "/home/mitchell/playarea/gramps/gramps2/src/RelLib.py", line
> 2499, in get_person_keys
>     if self.person_map:
> ValueError: __nonzero__ should return >=3D 0

This I haven't touched yet.=20
Don, maybe we should catch this specific exception and react=20
accordingly, rather then delegate it to the pass clause in ReadXML?

> > I almost did it, but then I thought that the present way is cleaner:
> > we have two check buttons, as opposed to three radio buttons. I think
> > as long as the present way is not confusing for the user (and I don't
> > think it is), we'd want to go with that. We can live with not being=20
> > purists and "mimic radio buttons with check boxes." Does this make=20
> > sense?
>=20
> It actually occurred to me that perhaps a combo box would be the best
> for this. It could be labeled something like "Show families as":
> "Individuals"
> "Stack"
> "Node"

That might be a good idea: one control instead of two or three.
I'll see about implementing this.

> > This should be fixed now. Of course, I only tested it with the
>=20
> Hmmm. I still get the same output when I choose display families as
> stacks. It also seems to create some people with no information about
> them.

Most likely, you have the database that was corrupted from previous
runs of the buggy RelGraph. This is what I used to get until I fixed
a couple of issues. It should be fine with the new database,
let me know if it is not.=20

And this is only true if you see my today's commits (check the ChangeLog,=
=20
the RelGraph entry should be the top one at the moment). The anonymous
CVS is a little behind the developer CVS, and "a little behind" might
be up to a day, depending on what kind of things those guys at sf.net
are fixing :-) Their site status page mentions 5 hours, so it's got
to be there tomorrow for sure.=20

Alex

--=20
Alexander Roitman   http://ebner.neuroscience.umn.edu/people/alex.html
Dept. of Neuroscience, Lions Research Building
2001 6th Street SE, Minneapolis, MN  55455
Tel (612) 625-7566   FAX (612) 626-9201