While working on other tasks related to error messages in Cite (e.g. T202593) we learned that what should be a purely technical code cleanup will have an effect that might or might not be considered user-facing, depending on how you want to argue.
Here is an example you can test on e.g. https://en.wikipedia.beta.wmflabs.org/wiki/Special:MyPage:
<references> <ref></ref> <ref></ref> <ref name=a></ref> <ref name=a></ref> <ref name=b>b</ref> <ref name=b>b</ref> </references>
This creates 6 error messages: 2 identical "<ref> in this context must have a name", 2 identical "<ref> in this context must have content", and 2 identical "<ref> in this context cannot be unused".
The planned code cleanup in https://gerrit.wikimedia.org/r/982355 would remove the duplicates.
Possible arguments:
- The duplicates are useless. We don't really loose anything when we remove them. The only thing that gets lost is the fact that there is more than one <ref> with the same error.
- However, what can the user do with this information? There is no way to address the duplicates individually. Using an index position (e.g. "the third <ref> from the top") is probably not helpful, especially because it will conflict with how the rest of the references on the page is numbered.
- If we want to help the user clean up duplicates we need another error message that says that. But that wouldn't be a blocker for the Technical-Debt cleanup.
- Maybe it's possible to improve the existing error messages. But that's a separate project that already has separate tickets, and not a blocker for the Technical-Debt cleanup as well.
- …?