2005/9/13, Nico Klasens <[EMAIL PROTECTED]>:
 > [ X] -1 (NO)

I too feel that everywhere MMBase, besides in one DataSource
implementation,  JDBC should not explicitel be used. So I think agree
on half of the proposed removals or so.

But I won't like the complete disappearance of   MultiPool and JDBC because

- It is simply configurable by a database URL. More simple than that
is hardly possible, which gives several advantages
 - We can ship a working mmbase with an hsql url.
 - You can easily try out another app-server you are not familiar with.
 - Running stand-alone remains an option, also for junit tests
 - It is possible to get the database name, because that is in the
connection url...
   This is currently not used, but I would like it as a fall back for
hsql, because
   its connection.getCatalog() simply always returns null (few days
back I asked quest about         this on hsql's dev-list, but no
response yet :-()
  I think it is a bit important to know the actual database name, e.g.
for blob dirs.
  Is there  a way to do this using only DataSource? Is 'getCalalog()'
the right way, and is this   a bug of hsql?

Furthermore is the MultiPool code of course  a distaster but I think
that perhaps now it is reasonably good and at least longly tested, and
it remains to be seen if implementations offered by app-servers are as
good.

But of course I'm all +1 for cleaning up al other mess. I'll change to
+0 with the amendements of Pierre, and to +1 if jdbc.xml with its
basic classes (JDBC, MultiPool and one or two others) simply remain
forever, though of course cleaned, or - of course - if one can
convince me that my points are invalid or irrelevant.

Michiel 

-- 
mihxil'  http://mihxil.komputilo.org/
nl_NL eo_XX en_US
_______________________________________________
Developers mailing list
[email protected]
http://lists.mmbase.org/mailman/listinfo/developers

Reply via email to