geoffj-FUG
commented
2 months ago Anne
We are certainly getting there. I have done a pretty comprehensive test. One bug which baffled me for a while: If I propagate when the POB recurs the validation stops at the next identical appearance if it is the next warning. I downloaded instances and found that the flags on the next entry were: False False True prior to validation of the first entry True True True after the validation of the first entry (correct). Then I found the issue: Row 369 had the first instance. Validate and select Next Warning. Row 370 had no entries in teh Warning explanation box but had a Current warning button. I selected this and it returned to row 369. Next warning took me to 370. Row 370 had been filled with the correct alternative POB but the current/next warning buttons had not updated.
We need some additional testing: If the verbatim POB is in the prevalidation data already and the piece was prevalidated then propagation should not be offered at all. When we remove the option to prevalidate (when everybody is comfortable with it) then this test will just be whether the verbatim POB is in the prevalidation data.
Have we considered that if the piece is ERY or WRY or NRY then all Yorkshire entries should be able to be enterer into prevalidation by the validator. Ditto for IOW/HAM and the Channel Islands.
I have now successfully validated and incorporated a piece using prevalidation in test.
Geoff
I have added the content for this story to Google Docs. Most of the code already exists. It can be found at: https://docs.google.com/document/d/1G76xWlt2a2jV91flHyk0JrizskVpqv4SKkSSA8Nx9Pk/edit
A cost benefit analysis is at the end of the document.
Geoff