CDR Tickets

I have begun the work on this task. Since we last did any investigation into the question of upgrading XMetaL the vendor has moved on from version 7.0 to version 9.0 (though the companion Developer 9.0 is not yet released). I downloaded and installed an evaluation copy of XMetaL Author 9.0 on my NCI virtual workstation (nci-ocedev03-v-m). I submitted a separate request for CBIIT to open port 2019 between that VM and nciws-d141-v (CDR DEV), but several days have gone by and there's been no movement on this request, so I folded in the tunneling over HTTPS (from another task) into a custom version of CdrClient.exe and Cdr.dll to work around this block. The good news is that the failure we ran into with XMetaL 7.0 (crashing on startup when our DLL was loaded), which the vendor was never able to explain or resolve, is no longer happening with 9.0. So far I have encountered three new bugs introduced by XMetaL 9.0, all connected with toolbars:

• our code to suppress visibility of toolbars for other doctypes is broken

• toolbars are now stacked in a vertical display

• our code to customize the context menu is broken

The vendor knows about the second problem already. I have reported the other two to the vendor, with a repro case for the first bug. That bug, in combination with the second, leaves little room for the display of the document being edited.

I will continue my testing.

Lots of back-and-forth with XMetaL tech support. Derek Read (their lead support person) wrote to say that they had been able to reproduce the behavior I reported using the repro case I sent them. He also said that he was surprised that the code I sent him ever worked in previous versions of XMetaL. He recommended that we re-write our toolbar management code to re-create the toolbars fresh each time the user opens a document or switches between documents. I read through the latest documentation on the event we're using to handle the toolbar visibility, and that documentation explicitly says we should avoid doing what he's telling us to do ("... do not create a new toolbar in this event macro"). I then went back to the documentation for the version of XMetaL that we're currently running to see if the docs had changed since then, and found that there wasn't any documentation for the macro. So then I combed back through my email archives and found that this event was created after the release of XMetaL 4.5 in order to support what we were trying to do with the toolbars, and the only guidance we had for how to use the event macro was the code that the vendor gave us showing how we should accomplish what we were trying to do. So I've written back to Derek to tell him that the code he's so surprised to find working in 4.5 was code he gave us himself. I'm so glad I don't throw away old email. :-)

Waiting for further instructions from the vendor.

Ran into another bug. While I was manually re-positioning the toolbars to work around the vertical stacking bug, the main menu bar disappeared, and I can't get it back. Reported the problem to Derek.

I have opened a ticket for CBIIT to create a machine William can use to test using XMetaL 9.0 to work with the CDR. Will let you know when it's ready, William. There will likely be some parts of your normal test suite which depend on connections to other machines which won't be accessible. We'll try to open up connections to the most important of those, but if we're not successful, we may need to fall back on switching the DEV bastion host from XMetaL 4.5 to XMetaL 9.0 for the rest of the testing. We'll see.

Users should be finishing up testing by the end of this week.

I have attached a Word document that highlights many of the tasks, activities, new features we tested as well some of the problems/bugs we encountered during testing. This document is the result of user testing from both ICRDB and CIAT.