From: Max I. <ma...@ma...> - 2002-10-08 06:55:22
|
Hi, I'd like to try MiddleKit but I am to use PostgreSQL DB, which seems unsupported. How complex would be to add the PostgreSQL support or may be it is already done in CVS? -- Bst rgrds, M.A.X.: Mechanical Artificial Xenomorph. |
From: Ian B. <ia...@co...> - 2002-10-09 00:48:27
|
On Tue, 2002-10-08 at 01:53, Max Ischenko wrote: > Hi, > > I'd like to try MiddleKit but I am to use PostgreSQL DB, which seems > unsupported. > > How complex would be to add the PostgreSQL support or may be it is > already done in CVS? There's a patch on sourceforge -- I have no idea what it looks like, or what kind of state it is in. Many people have started to work on this, but no one has really finished as far as I can tell. I was hoping one of them would pipe up -- at least to say what problem they encountered, or if they have some progress they could share. Maybe it's harder than it seems, or maybe it just doesn't quite get done, I dunno. Ian |
From: Gregory B. <gr...@wi...> - 2002-10-09 01:26:36
|
Ian Bicking wrote: > On Tue, 2002-10-08 at 01:53, Max Ischenko wrote: > >>Hi, >> >>I'd like to try MiddleKit but I am to use PostgreSQL DB, which seems >>unsupported. I did this almost two years ago and submitted the patch back to Chuck, but it never got integrated, probably because the regression testing system needed to be heavily modified to support testing a second database platform, and I didn't have the time to do that. At least one other person has completed the same work again, not knowing that I had done it once. Saying that I "did" this, though is kind of loaded. There was (and probably still is, correct me if I'm wrong) a requirement in MiddleKit for a method to quickly look up an object uniquely by its OID. OIDs in Postgres are not unique, and there is no way to quickly look one up. The port that I did hacked around this by doing a query on every single table looking for a specific row, which was obviously too slow of a method for heavy real-world use. For this reason I am currently using PyDO/Skunkweb. Greg |
From: Ian B. <ia...@co...> - 2002-10-09 01:39:28
|
On Tue, 2002-10-08 at 20:26, Gregory Brauer wrote: > For this reason I am currently using PyDO/Skunkweb. AFAIK, neither PyDO or MiddleKit are tied very closely to their respective frameworks, no? Ian |
From: Gregory B. <gr...@wi...> - 2002-10-09 16:31:18
|
Ian Bicking wrote: > On Tue, 2002-10-08 at 20:26, Gregory Brauer wrote: > >>For this reason I am currently using PyDO/Skunkweb. > > > AFAIK, neither PyDO or MiddleKit are tied very closely to their > respective frameworks, no? That's true. The only real ties are in the coding styles. I use Skunkweb because I got hooked on STML. Greg |
From: Ernesto R. <ae...@re...> - 2002-10-12 00:08:44
|
I did work for PostgreSQL, using pypgsql, although I actually don't know its state. There were some problems, one of them being the unique Object Id, and the other the support of enum data types which PostgresQL doesn't have but could emulate. On the other way it's very compatible to use the database to generate enums, although this makes it easier to do load balancing. In some of the OO papers, it's proposed to do a high/low counter generation, that is, there would be a light weight process to serve high numbers and each webware process would generate it's own low values, e.g. 32 buts for high and other 32 bits low, so each Webware process will request a new high value when it gets out of low values (after 2^32 times). This would make the Object Id code absolutely DB independent. On the other hand I would like to see going MiddleKit towards DB-API 2.0, or the new persistence framework of the Python Crew (when it is out), something more standard and not such mySQL specific, because there are so many wonderful databases out there (Firebird, SQL lite, MetaKit, etc.) Best regards, Erny ----- Original Message ----- From: "Ian Bicking" <ia...@co...> To: "Max Ischenko" <ma...@ma...> Cc: <web...@li...> Sent: Wednesday, October 09, 2002 2:49 AM Subject: Re: [Webware-discuss] MiddleKit port to PostgreSQL > On Tue, 2002-10-08 at 01:53, Max Ischenko wrote: > > Hi, > > > > I'd like to try MiddleKit but I am to use PostgreSQL DB, which seems > > unsupported. > > > > How complex would be to add the PostgreSQL support or may be it is > > already done in CVS? > > There's a patch on sourceforge -- I have no idea what it looks like, or > what kind of state it is in. Many people have started to work on this, > but no one has really finished as far as I can tell. I was hoping one > of them would pipe up -- at least to say what problem they encountered, > or if they have some progress they could share. Maybe it's harder than > it seems, or maybe it just doesn't quite get done, I dunno. > > Ian > > > > > ------------------------------------------------------- > This sf.net email is sponsored by:ThinkGeek > Welcome to geek heaven. > http://thinkgeek.com/sf > _______________________________________________ > Webware-discuss mailing list > Web...@li... > https://lists.sourceforge.net/lists/listinfo/webware-discuss |
From: Gregory B. <gr...@wi...> - 2002-10-12 00:54:25
|
Ernesto Revilla wrote: > the new persistence framework of the Python Crew I'd be curious to know more about this. Is there any place I can get info? Greg |
From: Gerhard <hae...@gm...> - 2002-10-12 01:00:19
|
* Gregory Brauer <gr...@wi...> [2002-10-11 17:54 -0700]: > Ernesto Revilla wrote: > >the new persistence framework of the Python Crew > > I'd be curious to know more about this. Is there any place I can get > info? http://www.python.org/sigs/persistence-sig/ -- Gerhard |