rindex yields wrong results on overlapped occurrence

[itchyny]

Describe the bug
s | rindex(t) yields wrong results when t appears in s with overlapped occurrence.

To Reproduce

 % jq -n '"ababa" | rindex("aba")'                      
0

Expected behavior
Considering rindex should yield the last occurrence index, the result should be 2 since "ababa"[2:] is "aba".

Environment (please complete the following information):

• OS and Version: macOS 12.4
• jq version: jq-1.6-159-gcff5336 (cff5336)

Additional context
The filter works well with overlapped occurrence in arrays.

 % jq -n '["a","b","a","b","a"] | rindex(["a","b","a"])'
2

This issue is an instance of the inconsistent behavior of indices.

 % jq -n '"abababa" | indices("aba")'
[
  0,
  4
]
 % jq -n '["a","b","a","b","a","b","a"] | indices(["a","b","a"])'
[
  0,
  2,
  4
]

[pkoppstein]

@emanuele6 - Any chance of fixing the def ofindex/1 while you're at it?

The line in builtin.jq says it all: def index($i): indices($i) | .[0]; # TODO: optimize