Allow dropping empty json objects

Resolves: RHEL-39979

liblognorm-2.0.0-rhbz1990737-add-skipempty.patch

@@ -0,0 +1,204 @@
+From eb2299a0897577048205e4d8a331168d82ce09d0 Mon Sep 17 00:00:00 2001
+From: Noriko Hosoi <nhosoi@momo7.localdomain>
+Date: Thu, 26 Jul 2018 17:18:38 -0700
+Subject: [PATCH] Add a parameter skipempty to the json field type.
+
+If skipempty is set as follows, empty json objects are dropped from
+the parsed result.
+  %field_name:json:skipempty%
+
+If any parameter other than "skipempty" is given ("bogus" in this
+example), an error message "invalid flag for JSON parser: bogus"
+is issued.
+---
+ src/parser.c                  | 127 ++++++++++++++++++++++++++++++++++++++++++
+ src/parser.h                  |   2 +-
+ src/pdag.c                    |   2 +-
+ 3 files changed, 129 insertions(+), 2 deletions(-)
+
+diff --git a/src/parser.c b/src/parser.c
+index 77407c6..6736c6f 100644
+--- a/src/parser.c
++++ b/src/parser.c
+@@ -2325,6 +2325,85 @@ PARSER_Parse(v2IPTables)
+ 	return r;
+ }
+ 
++/*
++ * Delete children of the given object if it has children and they are empty.
++ *
++ * return 0 if object is not empty
++ * return 1 if object is empty
++ * return < 0 if error
++ *
++ * Caller should do this:
++ * if (jsonSkipEmpty(obj) > 0) {
++ *     json_object_put(obj);
++ *     obj = NULL;
++ * }
++ * or otherwise not use obj if jsonSkipEmpty returns > 0.
++ */
++static int
++jsonSkipEmpty(struct json_object *__restrict__ json)
++{
++	int rc = 0;
++	struct json_object *val = NULL;
++
++	if(json == NULL) {
++		rc = 1;
++		goto finalize_it;
++	}
++
++	switch (json_object_get_type(json)) {
++	case json_type_string:
++		rc = json_object_get_string_len(json) == 0;
++		break;
++	case json_type_array:
++	{
++		int i;
++		int arrayLen = json_object_array_length(json);
++		for (i = 0 ; i < arrayLen ; ++i) {
++			val = json_object_array_get_idx(json, i);
++			if ((rc = jsonSkipEmpty(val)) > 0) {
++				/* delete the empty item and reset the index and arrayLen */
++				json_object_array_del_idx(json, i--);
++				arrayLen = json_object_array_length(json);
++			} else if (rc < 0) {
++				goto finalize_it;
++			}
++		}
++		rc = json_object_array_length(json) == 0;
++		break;
++	}
++	case json_type_object:
++	{
++		struct json_object_iterator it = json_object_iter_begin(json);
++		struct json_object_iterator itEnd = json_object_iter_end(json);
++		while (!json_object_iter_equal(&it, &itEnd)) {
++			val = json_object_iter_peek_value(&it);
++			if ((rc = jsonSkipEmpty(val)) > 0) {
++				json_object_object_del(json, json_object_iter_peek_name(&it));
++			} else if (rc < 0) {
++				goto finalize_it;
++			}
++			json_object_iter_next(&it);
++		}
++		rc = json_object_object_length(json) == 0;
++	}
++	case json_type_null:
++	case json_type_boolean:
++	case json_type_double:
++	case json_type_int:
++	default: break;
++	}
++finalize_it:
++	return rc;
++}
++
++/* 
++ * Parameters for field type json
++ *   skipempty - skips empty json objects.
++ *             - %field_name:json:skipempty%
++ */
++struct data_JSON {
++	int skipempty;
++};
+ /**
+  * Parse JSON. This parser tries to find JSON data inside a message.
+  * If it finds valid JSON, it will extract it. Extra data after the
+@@ -2340,6 +2419,7 @@ PARSER_Parse(v2IPTables)
+ PARSER_Parse(JSON)
+ 	const size_t i = *offs;
+ 	struct json_tokener *tokener = NULL;
++	struct data_JSON *const data = (struct data_JSON*) pdata;
+ 
+ 	if(npb->str[i] != '{' && npb->str[i] != ']') {
+ 		/* this can't be json, see RFC4627, Sect. 2
+@@ -2368,6 +2448,20 @@ PARSER_Parse(JSON)
+ 	if(value == NULL) {
+ 		json_object_put(json);
+ 	} else {
++		if (data && data->skipempty) {
++			int rc = jsonSkipEmpty(json);
++			if (rc < 0) {
++				json_object_put(json);
++				FAIL(LN_WRONGPARSER);
++			} else if (rc > 0) {
++				/* 
++				 * json value is empty.
++				 * E.g., {"message":""}, {"message":[]}, {"message":{}}
++				 */
++				json_object_put(json);
++				FAIL(0);
++			}
++		}
+ 		*value = json;
+ 	}
+ 
+@@ -2376,7 +2470,40 @@ PARSER_Parse(JSON)
+ 		json_tokener_free(tokener);
+ 	return r;
+ }
++PARSER_Construct(JSON)
++{
++	int r = 0;
++	struct json_object *ed;
++	struct data_JSON *data = NULL;
++	char *flag;
+ 
++	if(json == NULL)
++		goto done;
++
++	if(json_object_object_get_ex(json, "extradata", &ed) == 0) {
++		/* No JSON parameter */
++		goto done;
++	}
++	data = (struct data_JSON*) calloc(1, sizeof(struct data_JSON));
++	flag = json_object_get_string(ed);
++	if (strcasecmp(flag, "skipempty") == 0) {
++		data->skipempty = 1;
++	} else {
++		ln_errprintf(ctx, 0, "invalid flag for JSON parser: %s", flag);
++		r = LN_BADCONFIG;
++		goto done;
++	}
++	*pdata = data;
++done:
++	if(r != 0) {
++		free(data);
++	}
++	return r;
++}
++PARSER_Destruct(JSON)
++{
++	free(pdata);
++}
+ 
+ /* check if a char is valid inside a name of a NameValue list
+  * The set of valid characters may be extended if there is good
+diff --git a/src/parser.h b/src/parser.h
+index 38be62d..5b4a821 100644
+--- a/src/parser.h
++++ b/src/parser.h
+@@ -70,7 +70,7 @@ PARSERDEF_NO_DATA(Time24hr);
+ PARSERDEF_NO_DATA(Duration);
+ PARSERDEF_NO_DATA(IPv4);
+ PARSERDEF_NO_DATA(IPv6);
+-PARSERDEF_NO_DATA(JSON);
++PARSERDEF(JSON);
+ PARSERDEF_NO_DATA(CEESyslog);
+ PARSERDEF_NO_DATA(v2IPTables);
+ PARSERDEF_NO_DATA(CiscoInterfaceSpec);
+diff --git a/src/pdag.c b/src/pdag.c
+index 0768e99..9feb755 100644
+--- a/src/pdag.c
++++ b/src/pdag.c
+@@ -90,7 +90,7 @@ static struct ln_parser_info parser_lookup_table[] = {
+ 	PARSER_ENTRY_NO_DATA("duration", Duration, 16),
+ 	PARSER_ENTRY_NO_DATA("cisco-interface-spec", CiscoInterfaceSpec, 4),
+ 	PARSER_ENTRY_NO_DATA("name-value-list", NameValue, 8),
+-	PARSER_ENTRY_NO_DATA("json", JSON, 4),
++	PARSER_ENTRY("json", JSON, 4),
+ 	PARSER_ENTRY_NO_DATA("cee-syslog", CEESyslog, 4),
+ 	PARSER_ENTRY_NO_DATA("mac48", MAC48, 16),
+ 	PARSER_ENTRY_NO_DATA("cef", CEF, 4),

liblognorm.spec

@@ -2,7 +2,7 @@
 
 Name:		liblognorm
 Version:	2.0.6
-Release:	13%{?dist}
+Release:	14%{?dist}
 Summary:	Fast samples-based log normalization library
 License:	LGPL-2.1-or-later AND Apache-2.0
 URL:		http://www.liblognorm.com
@@ -20,6 +20,7 @@ BuildRequires:	libtool
 Patch0: liblognorm-2.0.6-rhbz2105934-sphinx5.patch
 Patch1: liblognorm-configure-glitch.patch
 Patch2: liblognorm-2.0.6-rhbz2128320.patch
+Patch3: liblognorm-2.0.0-rhbz1990737-add-skipempty.patch
 
 %description
 Briefly described, liblognorm is a tool to normalize log data.
@@ -64,6 +65,7 @@ log files.
 %patch -P 0 -p1 -b .sphinx5
 %patch -P 1 -p1 -b .configure-glitch
 %patch -P 2 -p1 -b .pcre2
+%patch -P 3 -p1
 
 %build
 # Prevent rebuild of the configure script.
@@ -102,6 +104,10 @@ rm %{buildroot}%{htmldir}/{objects.inv,.buildinfo}
 
 
 %changelog
+* Tue Aug 06 2024 Attila Lakatos <alakatos@redhat.com> - 2.0.6-14
+- Allow dropping empty json objects
+  Resolves: RHEL-39979
+
 * Mon Jun 24 2024 Troy Dawson <tdawson@redhat.com> - 2.0.6-13
 - Bump release for June 2024 mass rebuild