Re: Fwd: gasql questions

[Chris Wiegand <cwiegand urgentmail com>]

CVS that works would be nice. :) It's 7am in Colorado and the anoncvs
server still doesn't like me. I should just blame Canada (after all, it's
where the server appears to be :). Yes, I'm noting this all down, and I
think I'm close to having figured out in my head how the cnc string
parsing works, so that I can add these features. Are these strings also
going to be supported in other providers (only makes sense for those which
login to a server somehow). I'm also close to figuring out code for
auto-connecting to the default port, then two relatively standard places
for the socket if that fails, if no port and socket were defined. Once I
get some sort of CVS (anon or no) to get my computer up to date, I'll
start hacking that code.

On Tue, 28 Mar 2000, Rodrigo Moya wrote:

> >     Seems easier to just introduce two more keys "USERNAME" and "PASSWORD"
> > which are checked.  The DSNs aren't standardized anyways, correct?  That fix
> > is pretty easy and can be incorporated into the code near the area I pointed
> > out below.
> >
> 
> The DSN field is to pass any needed information to the database, so you might add
> USERNAME, PASSWORD and whatever you want, as long as this is understood by the
> MySQL provider.
> 
> >
> > > >     Unfortunatly, it looks like gda_open_connection is broken.  It's
> > > >pulling the username and password from the Gda_MYSQL_Connection structure,
> > > >and it *appears* that the username and password fields are never initialized
> > > >anywhere.  A *completetly untested* fix is thus:
> > > >
> > > >gda-mysql-server/gda-mysql-connection.c, line 218, insert:
> > > >
> > > >  cnc->user = g_strdup(user);
> > > >  cnc->password = g_strdup(passwd);
> > > >
> > > >    Which will allow you to specify the username/password at the login
> > > >dialog.  But there is still a question, the docs for gda_connection_open say:
> > > >
> > > >"""
> > > > If user or pwd is not NULL, it will overwrite the appropriate entry in the
> > > > DSN passed as par2. Entries in the DSN have the form = seperated
> > > > from the database name . Currently the DSN is not parsed.
> > > >"""
> > > >
> 
> Yes. Definitely, we should add the USERNAME and PASSWORD fields to the DSN, and do
> what is said in the docs, that is, to use the user and passwd parameters, or if not
> given, use the USER and PASSWORD in the DSN entry.
> 
> Chris, are you changing all this? (BTW, I should ask for a CVS account for you)
> ���{.n�+�������z&{�[�X���ު笷�'�g���0����{.n�+���j�ay+���-
>