CDR Tickets

This turns out to be a bug in XMetaL, which generates cached versions of the compiled DTDs, and is using a flawed algorithm to determine when the cached version must be recompiled.

I have implemented and installed a workaround for the bug on DEV. Until the workaround has been installed on the other tiers, it won't help with a user who has just switched to another tier than DEV, but it does ensure that when you switch to DEV from another tier you get the right versions of the DTDs.

Added Volker and Alan to the ticket as watchers.

Just a reminder that ~MBeckwit or ~JutheR will work with Erika to determine whether we can slip in a tiny CDR patch in December for this and the related Compact attribute issues..

The modifications for this ticket have been deployed on DEV and QA, and the deployment script to be given to CBIIT for STAGE and PROD has been tested on the lower tiers. User testing will involve confirming that switching between tiers (in particular, between tiers which do and do not have this patch applied) using XMetaL works smoothly.

I have installed the schema change dropping the Compact attribute from the list elements, and run the global change to drop the attribute from the documents on QA. CDR62707 and CDR62758 failed the global, because they were locked (checked out to volker and dyerv respectively). You should be able to test loading summaries on QA as long as you avoid those two (or if the lockers edit the documents themselves to remove the Compact attribute).

I propose to apply the schema change on STAGE and run the global change, and then repeat the process on production. I can't change the DTD without CBIIT's cooperation, so XMetaL would let you put the Compact attribute back in, though the CDR server won't accept the document as valid. But no one would do that, because they've been instructed not to, right? :-)

Any reason I shouldn't go forward with this proposal? It will eliminate the need to wipe out the unwanted RLX caches while we're waiting for the December patch to make its way through the process, and it will ensure that when CBIIT does install the patch it will be using the correct version of the common schema to generate the new DTDs.

I think this is a good approach. The sooner we can prevent users to use the Compact attribute the better.

Just to be 100% clear: we wouldn't be preventing them from saving documents with the obsolete attribute; they just wouldn't be able to create valid versions of such documents.

I think this sounds fine, too. William, can you please remind everyone at CIAT not to use the compact attribute, just to be sure? I am doing the same on our end. Thanks.

I guess it would be a good idea to add ~oseipokuw to the ticket if I'm going to ask him a question 🙂

Thanks, Robin! Users know not to use the attribute anymore.

I was in and out of QA and PROD a couple of times today without any problems.

I have switched between QA & PROD too and opened summary docs without a problem.

Verified on PROD.