AutoSuggest: normalize away diacritics similar to the way search does
Bug #943560 reported by
Lebbeous Fogle-Weekley
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Evergreen |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
For Evergreen master:
AutoSuggest: suggest "josé" when the user types "jose"
"josé" will serve as my example of a term with diacritics in it. I
don't mean that there is special handling just for the word "josé" per se.
Before now, the user could type "josé" and see suggestions containing
exactly that. You could not type "jose" and expect to see "josé". Now
you can.
In other words, typing "jose" and "josé" should now produce the exact same
set of suggestions. Only the version you actually typed will be
*highlighted* in the suggestions, however.
Changed in evergreen: | |
milestone: | 2.2.0alpha2 → 2.2.0alpha3 |
Changed in evergreen: | |
status: | Fix Committed → Fix Released |
To post a comment you must log in.
Wrapped the upgrade script and pushed to master, with an additional commit to address the "bouyancy" typo (which introduced the need for some additional changes to the upgrade script).
Tested the upgrade on a system running 0678 with Chromium and Firefox, and the concerto test dataset, with good results for "bartok", "fur", "Frederic", "Gyorgy". Nice work! Also tested the same with a clean install of 0679 and a reload of the database schema and concerto test dataset.
One note: I didn't update the 2.1-2.2 version upgrade script because the current version of that currently exists in a different branch, but whenever that comes to master and/or we branch rel_2_2, we'll need to ensure that it contains the fixes for "buoyant" / "buoyancy".