Convert json_in and jsonb_in to report errors softly.

This requires a bit of further infrastructure-extension to allow
trapping errors reported by numeric_in and pg_unicode_to_server,
but otherwise it's pretty straightforward.

In the case of jsonb_in, we are only capturing errors reported
during the initial "parse" phase.  The value-construction phase
(JsonbValueToJsonb) can also throw errors if assorted implementation
limits are exceeded.  We should improve that, but it seems like a
separable project.

Andrew Dunstan and Tom Lane

Discussion: https://postgr.es/m/3bac9841-fe07-713d-fa42-606c225567d6@dunslane.net

1 files changed, 4 insertions, 0 deletions

diff --git a/src/test/regress/sql/json_encoding.sql b/src/test/regress/sql/json_encoding.sql
index d7fac69733..f87b3bd4f3 100644
--- a/src/test/regress/sql/json_encoding.sql
+++ b/src/test/regress/sql/json_encoding.sql
@@ -76,3 +76,7 @@ SELECT jsonb '{ "a":  "dollar \u0024 character" }' ->> 'a' as correct_everywhere
 SELECT jsonb '{ "a":  "dollar \\u0024 character" }' ->> 'a' as not_an_escape;
 SELECT jsonb '{ "a":  "null \u0000 escape" }' ->> 'a' as fails;
 SELECT jsonb '{ "a":  "null \\u0000 escape" }' ->> 'a' as not_an_escape;
+
+-- soft error for input-time failure
+
+select pg_input_error_message('{ "a":  "\ud83d\ude04\ud83d\udc36" }', 'jsonb');