Issue with json inside json when it contains \n

[donhardman]

Bug Description:

It seems that at some point, we changed our JSON encoding/decoding scheme, which led to different behavior. Previously, we showed \\n, but now we have \n, which causes issues with JSON-in-JSON encoding.

mysql> create table mre(field string);
mysql> insert into mre (field) values ('{"value":"Hello\nWorld"}');
mysql> select * from mre;
+---------------------+-------------------------+
| id                  | field                   |
+---------------------+-------------------------+
| 1227966080878641153 | {"value":"Hello
World"} |
+---------------------+-------------------------+
> ^C
root@dev:/workdir# curl "0:9308/sql?mode=raw" -d "select * from mre;"
[{
"columns":[{"id":{"type":"long long"}},{"field":{"type":"string"}}],
"data":[
{"id":1227966080878641153,"field":"{\"value\":\"Hello\nWorld\"}"}
],
"total":1,
"error":"",
"warning":""
}]root@dev:/workdir#

Expected:

➜  ~  php -r 'var_dump(json_encode(["field" => json_encode(["value" => "hello\nworld"])]));'
string(41) "{"field":"{\"value\":\"hello\\nworld\"}"}"

We can see that \n becomes \\n as it's properly escaped

Manticore Search Version:

Latest dev version

Operating System Version:

Ubuntu Jammy

Have you tried the latest development version?

None

Internal Checklist:

^{To be completed by the assignee. Check off tasks that have been completed or are not applicable.}

• Implementation completed
• Tests developed
• Documentation updated
• Documentation reviewed
• Changelog updated

[tomatolog]

I checked your example and do not see the issue.

insert into mre (field) values ('{"value":"Hello\nWorld"}');

inserts string value and either

mysql> select * from mre;
+---------------------+-------------------------+
| id                  | field                   |
+---------------------+-------------------------+
| 1227966080878641153 | {"value":"Hello
World"} |
+---------------------+-------------------------+

root@dev:/workdir# curl "0:9308/sql?mode=raw" -d "select * from mre;"
[{
"columns":[{"id":{"type":"long long"}},{"field":{"type":"string"}}],
"data":[
{"id":1227966080878641153,"field":"{\"value\":\"Hello\nWorld\"}"}
],

returns field string attribute with the all control chars escaped as it should, ie match the original data.

Your example

php -r 'var_dump(json_encode(["field" => json_encode(["value" => "hello\nworld"])]));'

has json_encode twice - now quite sure why daemon should encode plain string second time