[GRASS5] Bugs Status Report

[Thierry Laronde]

On Wed, Oct 08, 2003 at 09:52:01AM +0200, Radim Blazek wrote:
> On Tuesday 07 October 2003 21:17, Thierry Laronde wrote:
> > drop for pg.in.dbf [seems more simple to work on unified DB access via
> > DBMI])?
> 
> JFYI, db.copy can copy a table from one db to another. 
> db.copy from_driver=dbf from_database=./   from_table=pocus \
>           to_driver=pg    to_database=test   to_table=pocus
> should work like pg.in.dbf.

Thanks for the tip. 

I have a request for some advice.

pg.in.dbf has several bugs (the one I have already reported; plus a bug
when the `DB' GRASS variables are set in conjunction with the `PG' ones;
plus the fact that what is proposed is to dump in ascii with a comma as
a field separator...whether or not there are already commas in the
fields' data; plus others in the bug tracking system), so: is it 
better to fix these, and when pg.in.dbf works correctly put everything 
in Attic and concentrate on db.*? Or to leave things as it is _but_ 
remove pg.in.dbf?

More generally, what is the policy for DB accesses in the future? Only
support generic access via DBMI (unixODBC), or allow too accesses
without unixODBC and maintain PostgreSQL direct access, or only support
generic db.* accesses but develop a kind of DBMI stub for direct PG
access (I think this would be definitively better to have a common 
interface)?

Cheers,
-- 
Thierry Laronde (Alceste) <tlaronde at polynum.org>
Key fingerprint = 0FF7 E906 FBAF FE95 FD89  250D 52B1 AE95 6006 F40C