[Dbix-class] DBIx::Class 0.08009 for testing

[Marc Mims]

* Matt S Trout <dbix-class at trout.me.uk> [080122 04:00]:
> On Mon, Jan 21, 2008 at 10:49:43AM -0800, Marc Mims wrote:
> > * Michael Higgins <linux at evolone.org> [080121 10:40]:
> > > On Sun, 20 Jan 2008 13:44:50 +0000 (GMT)
> > > Jess Robinson <castaway at desert-island.me.uk> wrote:
> > > 
> > > > 
> > > > http://desert-island.me.uk:8888/perldists/DBIx-Class-0.08009.tar.gz
> > > > 
> > > > Changes:
> > > >          - Added Storage::DBI subclass for MSSQL over ODBC.
> > > > 
> > > > Please poke this tarball some, so we can put it on CPAN!
> > > 
> > > Doing my best to be helpful here.
> > > 
> > > There was another addition to this (mentioned above) that I assumed
> > > would be added in this next release, so I guess I should jump in.
> > > 
> > > This:
> > > 
> > > sub build_datetime_parser {
> > >   my $self = shift;
> > >   my $type = "DateTime::Format::Strptime";
> > >   eval "use ${type}";
> > >   $self->throw_exception("Couldn't load ${type}: $@") if $@; 
> > >   return $type->new( pattern => '%Y-%m-%d %T' );
> > > }
> > > 
> > > ... is missing. (Perhaps Marc Mims is the one to contact, as he added
> > > the module, but hasn't, it would seem, committed this other key piece.)
> > 
> > Hi, Michael.
> > 
> > I'm trying to find time to finish work on the DateTime issue.  I know
> > the build_datetime_parser you quoted above is working for you, but I'm
> > not sure it's sufficient for all cases.  I believe it needs to handled
> > dates as well, not just timestamps.
> > 
> > I've done a bit of research and it seems the timestamp format you get
> > from MSSQL through ODBC may be a standard ODBC format, not an MSSQL
> > format.  I see a similar format, differing only in number of digits in
> > the fractional seconds, for DB2 over ODBC.  That leads me to believe
> > that build_datetime_parser might be best in ::DBI::ODBC with both
> > ::DBI::ODBC::DB2_400_SQL and ::DBI::ODBC::MSSQL derived from it.
> 
> Ok, I need you to ask yourself three questions:
> 
> (1) Is this broken now?
> (2) Will committing what you have so far be less broken?
> (3) If you do, will there be anything that appears to work that will change
> behaviour in a surprising fashion when you commit the complete version?
> 
> If your answers are yes, yes, and no, commit it and document that it's only
> a partial solution. Better half-works than not at all provided it won't cause
> an impediment to getting to "works 100%".

Yes. Yes. No. Committed with an appropriate note in the documentation.

Fear of commitment?  (Don't tell my spouse!)

	-Marc