SQLObject

On Thu, Sep 28, 2006 at 02:04:00PM -0300, Loic Jeannin wrote:
> http://www.tahorg.net/stuff/sqlobject-mysql-bug

   I think this should be solved with a proper style.

Oleg.
-- 
     Oleg Broytmann            http://phd.pp.ru/            phd@...
           Programmers don't die, they just GOSUB without RETURN.

On Thursday 28 September 2006 20:16, Oleg Broytmann wrote:
> On Thu, Sep 28, 2006 at 02:04:00PM -0300, Loic Jeannin wrote:
> > http://www.tahorg.net/stuff/sqlobject-mysql-bug
>
>    I think this should be solved with a proper style.

I think the original's poster patch is the right thing to do. I've come 
across this issue myself when testing sqlobject over an existing 
database. I never used fromDatabase other than for occasional tests, so 
it didn't bother me to the point of trying to fix it, but what the patch 
that was posted do make sense. In fact it makes a lot more sense to me to 
set the dbName of the column to the one obtained from the database, 
instead of relying on a "proper style" which should convert from a name 
to the ... same name, wasting time to perform a useless operation.

I really fail to see any reason to use a style here, when simply setting 
dbName to the field name as obtained from the database will do.
Using a style only makes sense when we describe the sqlobject side of the 
things and we need it to generate database field names according to a 
style we use, not when the field names are already known from the db.

>
> Oleg.

-- 
Dan

On Thu, Sep 28, 2006 at 09:16:49PM +0300, Dan Pascu wrote:
> I think the original's poster patch is the right thing to do.

   Better to say "it could have more sense". Unfortunataly, one-line patch
near .style.dbColumnToPythonAttr() looked for me like a quick and dirty
hack, like a shortcut over proper SQLObject path. I could be wrong, of
course, but that was my impression.
   Those who would want to make sense with the patch - welcome. You have
already reviewed the patch. Now it is time to test it and submit tests. I
could extend the patch to the other backends, but I certainly need a set of
good tests.

Oleg.
-- 
     Oleg Broytmann            http://phd.pp.ru/            phd@...
           Programmers don't die, they just GOSUB without RETURN.

On Thursday 28 September 2006 22:00, Oleg Broytmann wrote:
> On Thu, Sep 28, 2006 at 09:16:49PM +0300, Dan Pascu wrote:
> > I think the original's poster patch is the right thing to do.
>
>    Better to say "it could have more sense". Unfortunataly, one-line
> patch near .style.dbColumnToPythonAttr() looked for me like a quick and
> dirty hack, like a shortcut over proper SQLObject path. I could be
> wrong, of course, but that was my impression.

You can twist the expressions all you like, but IMO if fromDatabase is 
used, the database field names should come from the database not from a 
name style mapper.

-- 
Dan

On Thu, Sep 28, 2006 at 10:44:18PM +0300, Dan Pascu wrote:
> if fromDatabase is 
> used, the database field names should come from the database not from a 
> name style mapper.

   I agree. The only think I ask is tests. Please apply the patch, test it
and report. Please create some tests and submit them.

Oleg.
-- 
     Oleg Broytmann            http://phd.pp.ru/            phd@...
           Programmers don't die, they just GOSUB without RETURN.

On at 11:00:01PM +0400, Oleg Broytmann wrote:
> On Thu, Sep 28, 2006 at 09:16:49PM +0300, Dan Pascu wrote:
> > I think the original's poster patch is the right thing to do.
> 
>    Better to say "it could have more sense". Unfortunataly, one-line patch
> near .style.dbColumnToPythonAttr() looked for me like a quick and dirty
> hack, like a shortcut over proper SQLObject path. I could be wrong, of
> course, but that was my impression.

You've got it wrong, the one line patch is in columnFromDatabase().

>    Those who would want to make sense with the patch - welcome. You have
> already reviewed the patch. Now it is time to test it and submit tests. I
> could extend the patch to the other backends, but I certainly need a set of
> good tests.

On Thu, Sep 28, 2006 at 06:07:41PM -0300, Loic Jeannin wrote:
> You've got it wrong, the one line patch is in columnFromDatabase().

   I know. I meant the patch touches the line next to call
.style.dbColumnToPythonAttr()
   But there is no point arguing. I agreed that the *idea* of the patch is
good. It just needs more work. Someone should step in and
1) expand the patch to the other database;
2) add tests.
   I can do 1 (but not right now). But if nobody would do 2 RSN - the patch
will wait in my spool of patches for a long time...

Oleg.
-- 
     Oleg Broytmann            http://phd.pp.ru/            phd@...
           Programmers don't die, they just GOSUB without RETURN.

On Thu, Sep 28, 2006 at 09:16:49PM +0300, Dan Pascu wrote:
> On Thursday 28 September 2006 20:16, Oleg Broytmann wrote:
> > On Thu, Sep 28, 2006 at 02:04:00PM -0300, Loic Jeannin wrote:
> > > http://www.tahorg.net/stuff/sqlobject-mysql-bug
> >
> >    I think this should be solved with a proper style.
> 
> I think the original's poster patch is the right thing to do.

   I extended the original patch to all databases, except for Postgres,
that already had it, and SQLite that doesn't support fromDatabase at all. I
also added a minor patch to test_auto.py. Revisions 2016-2017.

Oleg.
-- 
     Oleg Broytmann            http://phd.pp.ru/            phd@...
           Programmers don't die, they just GOSUB without RETURN.