Thread: Smb to get involved
hello pgsql-hackers, I'm very very very... new here, but... Well, I see you're all stressed by the new release comming up soon, but may be one of you guys has got an ear and a couple of words for a smb who really wants to get involved. me: I think, I've got enough programming experience to get involved, if smb. really want's to take a look at what I can, please go to http://www.pbit.org/me_e.html I used to work with SQL databases in many projects, and PostgreSQL was the better one :-) why I'm here: well, I don't think, that I really can become an internals-guru after I read some of your messages here - it would take a long time - and I think, you guys don't need anyone new in the internals-kitchen. But I checked out your TODOs and I found an exotic issue, which sound interesting to me: to make a PostgreSQL database to work like an Oracle database to clients - it's not small and is not a bugfix, so I don't have to know about the deepest internals of PostgreSQL to implement it. my questions: is there smb. already working on it? Is it smth. this database realy needs? Did you guys worked out any usable docs about what's the ultimate way to implement such a feature (I just see a listener-idea in the list)? Or is it the right time to discuss smth. like that? Or is there an other "major" feature waiting for it's implementation and being more urgent? rgds Pavlo
Hi Pavlo, welcome aboard! Like any good free software collaborative project, PostgreSQL is always happy to have new contributors. Just be prepared for public, honest, _productive_ criticism of your code. If you've got an itch to make an Oracle compatability layer, scratch away: noone here will try to tell you what you _should_ be working on. Do note that being 'just like Oracle' is not a major goal for the project, but making it easy to port or write software against both databases is. If you decide to take a look at this, I'd suggest talking to the OpenACS people, and in particular Don Baccus <dhogaza@pacifier.com> who pops up here occasionally: they've probably ported more code from Oracle to PostgreSQL than anybody, and have a good idea about the niggly little details you'd need to address. Ross On Mon, Dec 24, 2001 at 05:26:34PM +0100, Pavlo Baron wrote: > hello pgsql-hackers, > > I'm very very very... new here, but... > Well, I see you're all stressed by the new release comming up soon, but may > be one of you guys has got an ear and a couple of words for a smb who really > wants to get involved. > > me: I think, I've got enough programming experience to get involved, if smb. > really want's to take a look at what I can, please go to > http://www.pbit.org/me_e.html > I used to work with SQL databases in many projects, and PostgreSQL was the > better one :-) > > why I'm here: well, I don't think, that I really can become an > internals-guru after I read some of your messages here - it would take a > long time - and I think, you guys don't need anyone new in the > internals-kitchen. But I checked out your TODOs and I found an exotic issue, > which sound interesting to me: to make a PostgreSQL database to work like an > Oracle database to clients - it's not small and is not a bugfix, so I don't > have to know about the deepest internals of PostgreSQL to implement it. > > my questions: is there smb. already working on it? Is it smth. this database > realy needs? Did you guys worked out any usable docs about what's the > ultimate way to implement such a feature (I just see a listener-idea in the > list)? Or is it the right time to discuss smth. like that? > Or is there an other "major" feature waiting for it's implementation and > being more urgent? > > rgds > Pavlo > > > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster
Ross J. Reedstrom writes: > Hi Pavlo, welcome aboard! :-) > Like any good free software collaborative > project, PostgreSQL is always happy to have new contributors. Just be > prepared for public, honest, _productive_ criticism of your code. oh yeah, I've already comunicated with some core gurus (Tom, Bruce, Thomas) between years about providing a fix for a TODO-item, but it seems to me, the thing I implemented would never match their expectations about how to fix smth. lying deep in the basic code. I thought, nobody ever replies to my first, Oracle-related email, so I used the moment to take a look at the deepest basics. Then, I nerved those guys a little ,) Now, I really don't think the codebase to be the right thing to me - in my honest opinion - there is a great amount of code written by smb. else and there are sufficient codebase experts having and being able to fix all those bugs. But this playing around did really help me to understand some basic concepts of the parser. > If > you've got an itch to make an Oracle compatability layer, scratch away: > noone here will try to tell you what you _should_ be working on. Do > note that being 'just like Oracle' is not a major goal for the project, > but making it easy to port or write software against both databases is. I never wanted to make PostgreSQL=Oracle, oh no. My original idea was to get involved by providing a translation service or a part of it to overlay one of it with the other one. I found some code and an Oracle-specific TODO in the source-tree thinking a bit different from the idea with a Net8 listener from the main TODO - I found this one terrible - maybe because I couldn't find any free documentation on how to implement the Net8 layer. I think, a complete SQL-translater for both databases would be the first step. But there are very very many features with Oracle missing in PG, so the translater Oracle2PG would get a little bit tricky to shallow some things. Is it interesting to discuss this theme now? (Smb. said, it's the wrong moment now for such discussions.) I didn't want to appear where everybody got stressed with the release completion. rgds Pavlo Baron