There was no problem in getting the various instances to point to a single Authenticator. I was worried that the UmlsDictionaryFast module was derived from the old one and thus dependent on it. But it’s not.
But there are still some wrinkles I’m working on as there have been other (4.0.0-4.0.1) changes about the places where the UMLS creds are collected and how params are propagated, coupled with “mandatory” statements in descriptors that prevent one from centralizing things. Because this is a patch, I am trying to make as few changes as possible, and this will take me a few more days, sorry. If anyone is in a hurry and has a total understanding of 4.0.0 configuration and wants to do it instead, I can just send you the authenticator module and its unit test which are working. Peter Sent from my iPad > On Nov 27, 2020, at 02:05, Finan, Sean <sean.fi...@childrens.harvard.edu> > wrote: > > Arg. Yeah. We need to deprecate that module ... I frequently forget > that it exists. > > There should not be any problems with having the one class ask permission > from the other. > > There shouldn't be any need to duplicate the working code. > > > ________________________________________ > From: Peter Abramowitsch <pabramowit...@gmail.com> > Sent: Wednesday, November 25, 2020 4:27 PM > To: dev@ctakes.apache.org > Subject: Small hitch in Authentication backport [EXTERNAL] > > * External Email - Caution * > > > Hi All > > It turns out that the code to approve UMLS usage is duplicated and exists > in two different jars. My upgrade was only in the class UmlsUserApprover, > but after some head scratching to explain what I was seeing, I discovered > that there's an earlier version of the same code still buried in > ./ctakes-dictionary-lookup/src/main/java/org/apache/ctakes/dictionary/lookup/ae/UmlsDictionaryLookupAnnotator.java > > It is a problem even in the trunk, but only affects the old UMLS Lookup > annotator > > I'll modify it to pont back to the distinct Approver class - unless doing > that causes a cyclic dependency between projects. > > Does anyone have other thoughts? > > Peter
