fuzzystrmatch: fix crash and compatibility

[jordanlewis]

Closes #82640

Release note (bug fix): fix behavior of soundex function when passed
certain unicode inputs. Previously, certain unicode inputs could result
in crashes, errors or incorrect outputs.

[cockroach-teamcity]

This change is 

[jordanlewis]

cc @mgoddard

[rafiss]

why did this one change? it looks like PG returns 000 for it (binary d0303030)

[rafiss]

PG returns just Z i think? (binary 5ac5c3c5)

[rafiss]

PG returns K00 (binary 4bd03030)

[jordanlewis]

The differences you wrote don't materialize on the system I was testing on... what version are you testing on?

[rafiss]

rafiss=# select version();
                                                      version
-------------------------------------------------------------------------------------------------------------------
 PostgreSQL 14.2 on x86_64-apple-darwin21.3.0, compiled by Apple clang version 13.0.0 (clang-1300.0.29.30), 64-bit
(1 row)

rafiss=# select soundex('zażółćx');
 soundex
---------
 Z
(1 row)

rafiss=# select soundex('🌞');
 soundex
---------
 000
(1 row)

rafiss=# select soundex('K😋');
 soundex
---------
 K00
(1 row)

(and i captured the results using wireshark)

[jordanlewis]

Here's what I get:

jordan=> select version();
                                                              version
-----------------------------------------------------------------------------------------------------------------------------------
 PostgreSQL 14.3 (Ubuntu 14.3-1.pgdg20.04+1) on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0, 64-bit
(1 row)

jordan=> select soundex('🌞');
 soundex
---------

(1 row)

jordan=> select soundex('K😋');
 soundex
---------
 K000
(1 row)

jordan=>

It sounds like this functionality just doesn't work consistently for UTF8:

[jordanlewis]

BTW, I get this same output even in Postgres 12 (I upgraded to Postgres 14 to see if it would repro on your version)

[rafiss]

ah i see thanks for looking. lgtm then

[rafiss]

bors r+

[craig]

Build succeeded:

• GitHub CI (Cockroach)