issue deprecation warning when assertion is made without context #2513
Conversation
… without a context
… Schematron, for no apparent reason other than it is mildly easier to read when consistent.
|
P.S. Since after this almost all Schematron was expressed with a namespace prefix (e.g., |
* Made deprecation message a warning, not an error. * Used 2025-03-15 as the end of deprecation period. * Added validUntil= (and a description) to the constraint specification that does the work, so that this shows up in Appendix G. * Added test for this in Test/detest.odd. * NOTE: The test in detest.odd will not actually fire, apparently because it is a warning not an error. So it will start to fire when we change it to an error on 2025-03-15.
We (the NA subgoup) seemed to think our test was passing without updating expected_results because it was a warning, not an error. But CI server disagrees. Hmmm …
request European subgroup)
|
@martindholmes the tests are passing now, could you please have a look at it? |
There was a problem hiding this comment.
This looks good to me. But I really wish we could keep PRs to a single set of changes; this seems to be changing namespace mechanisms and whitespace as well as handling the @context problem. When there's so much change in one batch, it's really hard to review the PR. Could that be a recommendation somewhere?
|
Thanks @martindholmes for the quick review. You are right, different changes should not be tackled all at once. We are trying to stick to our policy better. |
Work on #2510.
Since this does not use the
@validUntildeprecation mechanism, does it need to be added to Appendix G by hand?I chose 2024-06-30 as end-of-deprecation date on my own accord. But since there is basically no difference between deprecation and error, here (although I suppose we could add role=warning), should we just go straight to error?
Anyway, it passes all tests on my local system both in & out of docker.
Reviewers: You will definitely want to ignore whitespace if you try comparing code. (Other than the code of
<constraintSpec>, it is mostly a kinda pointless activity, IMHO.)