Closed euresti closed 6 years ago
Oops. All this time I've been coding against an old version of the spec. Sorry about that!
On Thu, Jun 21, 2018, 4:16 PM Christopher Singley notifications@github.com wrote:
Merged #43 https://github.com/csingley/ofxtools/pull/43.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/csingley/ofxtools/pull/43#event-1694859596, or mute the thread https://github.com/notifications/unsubscribe-auth/ABCCAssYfrsY_kpO6ZIopwvF_O9PGuApks5t_ClTgaJpZM4UyNCf .
The current set of models are doubtless missing attributes, as I wrote them against older versions of the spec, and JPM/WFC keep pushing extra cruft into the standard. Among its many other failings, the ofxtools project lacks a compliance review process. However it's worthwhile downloading a fresh copy of the spec whenever you're in there doing work.
This is all great.
In the future you could me a favor, when adding a new aggregate like
CCCLOSING
, as long as you've got the spec open to go ahead and add class attributes for all the garbage likeREWARDINFO
. The idea is to have an attribute for every subagg/element in the spec.This is the reason I added
models.base.Unsupported
. If a server returns an aggregate containing stuff like this, and the ofxtools model doesn't have a corresponding attribute,Parser.OFXTree
crashes with a most unhelpful error message. If it hitsUnsupported
the user at least knows what's going wrong, where, and what needs to be done to fix it.