Dan Kubb (dkubb) wrote: > Michael, > >>> I am of the opinion that these complex multi-repo problems should >>> not be attempted in v1.0. Do we really have that strong a need for >>> it? >> The whole idea originally was to rewrite poor DM spec suite first, and >> then fix bugs having something to protect your back as you go. >> I think this is extremely important in 1.0, and 1.0 should not happen >> until the effort is finished. > > I know you already posted that you misread Jon's post, but I wanted to > address this just so everyone's clear. > > DM is supposed to support cross-repo associations right now. There is > code that attempts to do it, poorly, and makes all sorts of > assumptions that turn out not to be true when you begin to break it > down. > > The association behavior has already been specced, and work has begun > to make dkubb/dm-core pass those specs. There's still some work to be > done, but the results have been really good and I encourage people to > compare them with sam/dm-core. > > However, Many to Many associations were a sticking point. The code > passed maybe half of the specs, and the choice was to either try to > hack the Relationship, RelationshipChain and ManyToMany classes to > work or rewrite the internals properly. The problem wasn't so much as > just buggy code, it was code that was designed only to do reads. > Writes barely work at all (!), reads weren't as efficient as they > could be and forget about doing anything that spanned more than 3 > models, never mind cross-repo stuff. With those requirements I > decided to rewrite it, and so far it's been really good -- I should > have some working many to many association code in the next week or > so. > > Dan > (dkubb)
Just want to throw in my $.02. My scenario for multiple repository usage is this: identical schema, mapped onto n databases. The schema includes has-many, belongs-to, and habtm relationships, all of which reference tables within a given database (i.e. no cross-repository associations). Currently in sam/dm-core, when I access the far end of a habtm relationship, it always looks in the default repository for the far end. I've hacked a couple of lines to make it look in the near end's repository. That works for me, but I think I remember some others preferring the current behavior. I guess my point is that this behavior should be selectable in dkubb/dm-core. I'm not sure how clear the above paragraph is. If elaboration is required, let me know. Earle --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "DataMapper" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/datamapper?hl=en -~----------~----~----~----~------~----~------~--~---
