From 26f4ae74bcb7ea1d6c6cd254fccbfaf6c0869db4 Mon Sep 17 00:00:00 2001
From: mariano scasso <mariano.scasso@scanoss.com>
Date: Tue, 11 Nov 2025 00:41:45 +0000
Subject: [PATCH 1/5] License query improvement: search also using latest
 component version

---
 inc/debug.h   |  1 +
 inc/scanoss.h |  2 +-
 src/license.c | 28 ++++++++++++++++++++++++----
 src/scan.c    |  4 ++++
 src/scanlog.c | 18 +++++++++++++++++-
 5 files changed, 47 insertions(+), 6 deletions(-)

diff --git a/inc/debug.h b/inc/debug.h
index 12d8a2e..ee2caa7 100644
--- a/inc/debug.h
+++ b/inc/debug.h
@@ -32,6 +32,7 @@ extern bool debug_on; //= false; //set debug mode from main.
 extern bool quiet;
 
 
+bool scanlog_init(void);
 void scanlog(const char *fmt, ...);
 void map_dump(scan_data_t *scan);
 long microseconds_now(void);
diff --git a/inc/scanoss.h b/inc/scanoss.h
index 23bd905..855a99c 100644
--- a/inc/scanoss.h
+++ b/inc/scanoss.h
@@ -39,7 +39,7 @@
 #define WFP_LN 4
 #define WFP_REC_LN 18
 
-#define SCANOSS_VERSION "5.4.18"
+#define SCANOSS_VERSION "5.4.19"
 
 /* Log files */
 #define SCAN_LOG "/tmp/scanoss_scan.log"
diff --git a/src/license.c b/src/license.c
index f33a671..3035e12 100644
--- a/src/license.c
+++ b/src/license.c
@@ -51,7 +51,7 @@
 	 4 = Scancode detection 
 	 5 = Scancode detection at mining time
 	 6 = osslot */
-const char *license_sources[] = {"component_declared", "file_spdx_tag", "file_header", "license_file", "scancode", "scancode-file", "osselot"};
+const char *license_sources[] = {"component_declared", "file_spdx_tag", "file_header", "license_file", "scancode-file", "scancode", "osselot"};
 bool full_license_report = false;
 
 
@@ -337,8 +337,6 @@ bool print_licenses_item(uint8_t *key, uint8_t *subkey, int subkey_ln, uint8_t *
 
 	int src = atoi(source);
 
-	scanlog("Fetched license %s\n", license);
-
 	if (strlen(license) > 2 && (src < (sizeof(license_sources) / sizeof(license_sources[0]))))
 		license_add_to_list(&licenses[src], license);
 
@@ -401,8 +399,30 @@ void print_licenses(component_data_t *comp)
 		scanlog("License for %s@%s license returns %d hits\n", comp->purls[i], comp->version, records);
 
 		if (records)
-			break;
+		{
+			//Look if someone of the prefered liceses ids already has a match
+			for (int i = 0; i < 4; i++)
+			{
+				if (licenses_by_type[i].count > 0)
+				{
+					scanlog("Stop searching for licenses\n");
+					break;
+				}
+			}
+		}
 
+		if (strcmp(comp->version, comp->latest_version) != 0)
+		{
+			/* Calculate purl@latest_version md5 */
+			purl_version_md5(purlversion_md5, comp->purls[i], comp->latest_version);
+
+			records = ldb_fetch_recordset(NULL, oss_license, purlversion_md5, false, print_licenses_item, &licenses_by_type);
+			scanlog("License for %s@%s license returns %d hits\n", comp->purls[i], comp->latest_version, records);
+
+			if (records)
+				break;
+		}
+		/* Unversioned purl license */
 		records = ldb_fetch_recordset(NULL, oss_license, comp->purls_md5[i], false, print_licenses_item, &licenses_by_type);
 		scanlog("License for %s license returns %d hits\n", comp->purls[i], records);
 		
diff --git a/src/scan.c b/src/scan.c
index 0123ab0..84401fa 100644
--- a/src/scan.c
+++ b/src/scan.c
@@ -449,6 +449,10 @@ void ldb_scan(scan_data_t *scan)
 		exit(EXIT_FAILURE);
 	}
 
+	// Clean up the log file
+	if (debug_on)
+		scanlog_init();
+
 	scan->matchmap_size = 0;
 	scan->match_type = MATCH_NONE;
 	scan->timer = microseconds_now();
diff --git a/src/scanlog.c b/src/scanlog.c
index 01b73bc..adef2da 100644
--- a/src/scanlog.c
+++ b/src/scanlog.c
@@ -47,9 +47,25 @@ long microseconds_now()
 	return (now.tv_sec*(int)1e6+now.tv_usec);
 }
 
+/**
+ * @brief Initialize the log file as blank
+ * @return true if successful, false otherwise
+ */
+bool scanlog_init()
+{
+	FILE *log = fopen(SCAN_LOG, "w");
+	if (!log)
+	{
+		fprintf(stderr, "Warning: Cannot create/initialize the log file\n");
+		return false;
+	}
+	fclose(log);
+	return true;
+}
+
 /**
  * @brief Print the logs in stderr
- * @param fmt string to be printed  
+ * @param fmt string to be printed
  * @param ... //TODO
  */
 void scanlog(const char *fmt, ...)

From b59d3cf26ca0f64d763dae96624f1fc00d5049fc Mon Sep 17 00:00:00 2001
From: mariano scasso <mariano.scasso@scanoss.com>
Date: Fri, 14 Nov 2025 21:51:05 +0000
Subject: [PATCH 2/5] Change matched percent type. Change final selection on
 snippet matching logic

---
 inc/match.h             |  2 +-
 src/component.c         |  6 ++----
 src/license.c           |  6 +++---
 src/match.c             | 29 ++++++++++++++++-------------
 src/report.c            |  2 +-
 src/snippet_selection.c |  2 +-
 src/util.c              |  2 +-
 7 files changed, 25 insertions(+), 24 deletions(-)

diff --git a/inc/match.h b/inc/match.h
index 31278f3..a807919 100644
--- a/inc/match.h
+++ b/inc/match.h
@@ -16,7 +16,7 @@ typedef struct match_data_t
 	int lines_matched; /*number of matched lines*/
 	char * line_ranges; /*input snippet line ranges */
 	char * oss_ranges; /* kb snippet line ranges */
-	char * matched_percent; /* matched percent */
+	int matched_percent; /* matched percent */
 	int  path_ln; /*file path lenght*/ //TODO check if this is needed.
 	uint8_t file_md5[MD5_LEN]; /* file md5 */
 	char source_md5[MD5_LEN * 2 + 1]; /*matched file md5 in hex format */
diff --git a/src/component.c b/src/component.c
index 40b4eeb..8906313 100644
--- a/src/component.c
+++ b/src/component.c
@@ -209,8 +209,9 @@ static char * look_for_version(char *in)
 void fill_component_path(component_data_t *component, char *file_path)
 {
 	component->file = strdup(look_for_version(file_path));
-	component->path_ln = strlen(file_path);
+	component->path_ln = strlen(component->file);
 	flip_slashes(component->file);
+	component->path_depth = path_depth(component->file);
 }
 
 /**
@@ -298,9 +299,6 @@ bool fill_component(component_data_t *component, uint8_t *url_key, char *file_pa
 	}
 	else
 		component->rank = COMPONENT_DEFAULT_RANK;
-	
-	component->path_depth = path_depth(component->file);
-
 	return true;
 }
 
diff --git a/src/license.c b/src/license.c
index 3035e12..83433f2 100644
--- a/src/license.c
+++ b/src/license.c
@@ -220,7 +220,7 @@ static char *json_from_license(uint32_t *crclist, char *buffer, char *license, i
 	string_clean(license);
 	int len = 0;
 
-	if (strlen(license) < 2)
+	if (!*license || strlen(license) < 2)
 		return buffer;
 	/* Calculate CRC to avoid duplicates */
 	uint32_t CRC = string_crc32c(license);
@@ -337,7 +337,7 @@ bool print_licenses_item(uint8_t *key, uint8_t *subkey, int subkey_ln, uint8_t *
 
 	int src = atoi(source);
 
-	if (strlen(license) > 2 && (src < (sizeof(license_sources) / sizeof(license_sources[0]))))
+	if (src < (sizeof(license_sources) / sizeof(license_sources[0])))
 		license_add_to_list(&licenses[src], license);
 
 	free(source);
@@ -431,7 +431,7 @@ void print_licenses(component_data_t *comp)
 	}
 
 	/* Open licenses structure */
-	char * result = calloc(MAX_FIELD_LN * 100, 1);
+	char * result = calloc(MAX_FIELD_LN * 1024, 1);
 	char * buffer = result;
 	int len = 0;
 
diff --git a/src/match.c b/src/match.c
index 1b1a41d..f50e962 100644
--- a/src/match.c
+++ b/src/match.c
@@ -68,7 +68,6 @@ void match_data_free(match_data_t *data)
     free_and_null((void **)&data->snippet_ids);
     free_and_null((void **)&data->line_ranges);
     free_and_null((void **)&data->oss_ranges);
-    free_and_null((void **)&data->matched_percent);
     free_and_null((void **)&data->crytography_text);
     free_and_null((void **)&data->quality_text);
     component_list_destroy(&data->component_list);
@@ -91,7 +90,7 @@ match_data_t * match_data_copy(match_data_t * in)
     out->type = in->type;
     out->line_ranges = strdup(in->line_ranges);
     out->oss_ranges = strdup(in->oss_ranges);
-    out->matched_percent = strdup(in->matched_percent);
+    out->matched_percent = in->matched_percent;
     out->snippet_ids = strdup(in->snippet_ids);
     strcpy(out->source_md5, in->source_md5);
     return out;
@@ -339,7 +338,7 @@ static bool component_hint_date_comparation(component_data_t *a, component_data_
 	if (b->rank < COMPONENT_DEFAULT_RANK || a->rank < COMPONENT_DEFAULT_RANK)
 	{
 		//shorter path lenght are prefered
-		if (b->rank < a->rank &&b->path_depth < a->path_depth/2)
+		if (b->rank < a->rank && b->path_depth < a->path_depth/2)
 			return true;
 		else if (b->rank > a->rank && a->path_depth < b->path_depth/2)
 			return false;
@@ -450,8 +449,8 @@ bool add_component_from_urlid(component_list_t *component_list, uint8_t *url_id,
 	if (!new_comp)
 		return false;
 		
-	fill_component_path(new_comp, path);
 	/* Create a new component and fill it from the url record */
+	fill_component_path(new_comp, path);
 
 	new_comp->file_md5_ref = component_list->match_ref->file_md5;
 	/* If the component is valid add it to the component list */
@@ -543,13 +542,13 @@ bool load_matches(match_data_t *match)
 	{
 		asprintf(&match->line_ranges, "n/a");
 		asprintf(&match->oss_ranges, "n/a");
-		asprintf(&match->matched_percent, "%d functions matched", match->hits);
+		match->matched_percent = -1;
 	}
 	else if (match->type == MATCH_FILE)
 	{
 		asprintf(&match->line_ranges, "all");
 		asprintf(&match->oss_ranges, "all");
-		asprintf(&match->matched_percent, "100%%");
+		match->matched_percent = 100;
 	}
 
 	uint32_t records = 0;
@@ -745,7 +744,7 @@ void match_select_best(scan_data_t *scan)
 			}
 
 			//If best match has 20% more of hits do nothing.
-			if (best_match->hits >= match->hits * 1.2)
+			if (best_match->hits >= match->hits * 1.2 && best_match_component->path_depth < match_component->path_depth)
 				continue;
 
 			//if cantidate has 10% more of hits do not consider dates and switch
@@ -753,13 +752,18 @@ void match_select_best(scan_data_t *scan)
 			{
 				scanlog("Replacing best match due to big hits difference\n");
 				scan->matches_list_array[i]->best_match = item->match;
+				continue;
 			} 
 			// if the hit numbers are close, select the oldest.
-			else if (abs(scan->matches_list_array[i]->best_match->hits - item->match->hits) <= 2 &&
-					find_oldest_match(scan->matches_list_array[i]->best_match, item->match))
+			int diff = abs(best_match->matched_percent - match->matched_percent);
+
+			if (diff <= 10)
 			{
-				scanlog("Replacing best match for an older version with equal hits\n");
-				scan->matches_list_array[i]->best_match = item->match;
+				if (component_date_comparation(best_match_component, match_component))
+				{
+					scanlog("Replacing best match for an older version with similar hits\n");
+					scan->matches_list_array[i]->best_match = item->match;
+				}
 			}
 		}
 	}
@@ -792,10 +796,9 @@ void match_select_best(scan_data_t *scan)
 						{
 							free(scan->matches_list_array[i]->best_match->line_ranges);
 							free(scan->matches_list_array[i]->best_match->oss_ranges);
-							free(scan->matches_list_array[i]->best_match->matched_percent);
 							scan->matches_list_array[i]->best_match->line_ranges = r.local;
 							scan->matches_list_array[i]->best_match->oss_ranges = r.remote;
-							scan->matches_list_array[i]->best_match->matched_percent = r.matched;
+							scan->matches_list_array[i]->best_match->matched_percent = atoi(r.matched);
 
 							max_hits = scan->matches_list_array[i]->best_match->hits;
 							index = i;
diff --git a/src/report.c b/src/report.c
index 5a9e3b8..190f8d5 100644
--- a/src/report.c
+++ b/src/report.c
@@ -321,7 +321,7 @@ bool print_json_match(struct match_data_t * match)
 	printf("\"id\": \"%s\"", matchtypes[match->type]);	
 	printf(",\"lines\": \"%s\"", match->line_ranges);
 	printf(",\"oss_lines\": \"%s\"", match->oss_ranges);
-	printf(",\"matched\": \"%s\"", match->matched_percent);
+	printf(",\"matched\": \"%d%%\"", match->matched_percent);
 	
 	if ((engine_flags & ENABLE_SNIPPET_IDS) && match->type == MATCH_SNIPPET)
 	{
diff --git a/src/snippet_selection.c b/src/snippet_selection.c
index 388705a..adeb307 100644
--- a/src/snippet_selection.c
+++ b/src/snippet_selection.c
@@ -151,7 +151,7 @@ void biggest_snippet(scan_data_t *scan)
 				matched_percent = 99;
 			if (matched_percent < 1)
 				matched_percent = 1;
-			asprintf(&match_new->matched_percent, "%u%%", matched_percent);
+			match_new->matched_percent = matched_percent;
 			match_new->lines_matched = matched_lines;
 			//match_new->hits = hits;
 
diff --git a/src/util.c b/src/util.c
index 820f216..2c6d509 100644
--- a/src/util.c
+++ b/src/util.c
@@ -380,7 +380,7 @@ int path_is_third_party(const char* path)
         "contrib",           // Contributed/third-party code
         "plugin",            // Plugins (often third-party)
 
-        "lib", "components", "modules", "ext",
+        "utils","lib", "components", "modules", "ext",
         "test", "fixtures", "examples",
         "files", "assets", "runtime",
         "subprojects", "managed", "local_packages", "published",

From a9b110dd25eaa13f4571a8ab00767beeb69bb1e9 Mon Sep 17 00:00:00 2001
From: mariano scasso <mariano.scasso@scanoss.com>
Date: Fri, 14 Nov 2025 22:25:30 +0000
Subject: [PATCH 3/5] fix mutiple licenses parsing bug

---
 src/license.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/license.c b/src/license.c
index 83433f2..2651123 100644
--- a/src/license.c
+++ b/src/license.c
@@ -272,7 +272,7 @@ static char *split_in_json_array(uint32_t *crclist, char *buffer, char *license,
 
 	} while (next_lic);
 
-	return buffer;
+	return r;  // Return the updated buffer pointer, not the original
 }
 
 char *  license_to_json(uint32_t *crclist, char *buffer, char *license, int src, bool *first_record)

From 87f4b6a181a4115f5d4c4c512edb47a539d30841 Mon Sep 17 00:00:00 2001
From: mariano scasso <mariano.scasso@scanoss.com>
Date: Wed, 19 Nov 2025 11:24:43 +0000
Subject: [PATCH 4/5] third party filter update. Best component selection logic
 tune up for ranked urls. Performance optimization

---
 inc/component.h |  1 +
 inc/scanoss.h   |  2 +-
 inc/util.h      |  2 +-
 src/license.c   |  1 +
 src/limits.c    | 11 +-----
 src/match.c     | 90 +++++++++++++++++++++++++++++++++++++++----------
 src/util.c      | 16 ++++++---
 7 files changed, 89 insertions(+), 34 deletions(-)

diff --git a/inc/component.h b/inc/component.h
index 3a3c122..5c73f46 100644
--- a/inc/component.h
+++ b/inc/component.h
@@ -59,6 +59,7 @@ typedef struct component_data_t
 	int health_stats[3]; /* health stats: forks, watchers, contributors */
 	int rank; /* purl ranking - optional*/
 	int path_depth; /* depth of the matched file path*/
+	int third_party_rank; /* Saves third party ranking*/
 } component_data_t;
 
 component_data_t * component_init(void);
diff --git a/inc/scanoss.h b/inc/scanoss.h
index 855a99c..e32c313 100644
--- a/inc/scanoss.h
+++ b/inc/scanoss.h
@@ -31,7 +31,7 @@
 #include "limits.h"
 
 #define MAX_FILE_PATH 1024
-#define FETCH_MAX_FILES 20000
+#define FETCH_MAX_FILES 12000
 #define MIN_FILE_SIZE 256 // files below this size will be ignored
 #define CRC_LIST_LEN 1024 // list of crc checksums to avoid metadata duplicates
 #define SNIPPET_LINE_TOLERANCE 10
diff --git a/inc/util.h b/inc/util.h
index 6445c0e..c0f89a2 100644
--- a/inc/util.h
+++ b/inc/util.h
@@ -63,7 +63,7 @@ char * str_cat_realloc(char **a, char * b);
 
 void free_and_null(void ** pr);
 
-int path_is_third_party(const char* path);
+int path_is_third_party(component_data_t *comp);
 
 /* Counts the number of '/' characters in a path string */
 int path_depth(char* path);
diff --git a/src/license.c b/src/license.c
index 2651123..13a8592 100644
--- a/src/license.c
+++ b/src/license.c
@@ -336,6 +336,7 @@ bool print_licenses_item(uint8_t *key, uint8_t *subkey, int subkey_ln, uint8_t *
 	free(CSV);
 
 	int src = atoi(source);
+	scanlog("Fetched License %s - source ID %d\n", license, src);
 
 	if (src < (sizeof(license_sources) / sizeof(license_sources[0])))
 		license_add_to_list(&licenses[src], license);
diff --git a/src/limits.c b/src/limits.c
index 3e27b69..8471794 100644
--- a/src/limits.c
+++ b/src/limits.c
@@ -9,17 +9,8 @@
   * @see https://github.com/scanoss/engine/blob/master/src/limits.c
   */
 
-
-int consecutive_score = 4000;   /** Maximumm sUsed for snippet selection */
-
-/* During snippet scanning, when a wfp (with more than consecutive_threshold wfps) produces a score higher 
-   than consecutive_score by consecutive_hits in a row, the scan will skip consecutive_jump snippets */
-int consecutive_hits = 4;       
-int consecutive_jump = 5;      
-int consecutive_threshold = 50; 
-
 int range_tolerance = 5;  /** A maximum number of non-matched lines tolerated inside a matching range */
 int min_match_lines = 10; /** Minimum number of lines matched for a match range to be acepted */
 int min_match_hits  = 4;  /** Minimum number of snippet ID hits to produce a snippet match*/
 
-const int max_vulnerabilities = 50; /** Show only the first N vulnerabilities */
+const int max_vulnerabilities = 50; /** Show only the first N vulnerabilities */
\ No newline at end of file
diff --git a/src/match.c b/src/match.c
index f50e962..d0d1f07 100644
--- a/src/match.c
+++ b/src/match.c
@@ -312,37 +312,41 @@ static bool component_hint_date_comparation(component_data_t *a, component_data_
 			}
 		}
 		else if (a->path_rank < PATH_LEVEL_COMP_REF / 3 + 1)
+		{
+			scanlog("%s rejected, %s wins by path rank %d\n", b->purls[0], a->purls[0], a->path_rank);
 			return false;
+		}
 	}
 
 	if (!*b->release_date)
+	{
+		scanlog("%s rejected due to empty release date\n", b->purls[0]);
 		return false;
+	}
 	if (!*a->release_date)
+	{
+		scanlog("%s accepted, %s has empty release date\n", b->purls[0], a->purls[0]);
 		return true;
+	}
 
 	// Third-party path evaluation
-	int tp_a = path_is_third_party(a->file);
-	int tp_b = path_is_third_party(b->file);
+	int tp_a = path_is_third_party(a);
+	int tp_b = path_is_third_party(b);
 
-	if (tp_a > tp_b)
+	if (tp_a - tp_b > 4)
 	{
-		scanlog("Component rejected by third party path filter (%s=%d > %s=%d)\n", a->purls[0], tp_a, b->purls[0], tp_b);
+		scanlog("Component rejected by third party path filter (%s=%d=%s > %s=%d=%s)\n", a->purls[0], tp_a,a->file, b->purls[0], tp_b, b->file);
 		return false;
 	}
-	else if (tp_a < tp_b)
+	else if (tp_b - tp_a > 4)
 	{
 		scanlog("Component accepted by third party path filter (%s=%d < %s=%d)\n",  a->purls[0], tp_a,  b->purls[0], tp_b);
 		return true;
 	}
+
 	//when the url ranking is enabled
 	if (b->rank < COMPONENT_DEFAULT_RANK || a->rank < COMPONENT_DEFAULT_RANK)
-	{
-		//shorter path lenght are prefered
-		if (b->rank < a->rank && b->path_depth < a->path_depth/2)
-			return true;
-		else if (b->rank > a->rank && a->path_depth < b->path_depth/2)
-			return false;
-		
+	{		
 		bool good_purl_a = binary_file_to_purl(a);
 		bool good_purl_b = binary_file_to_purl(b);
 		if (good_purl_b && !good_purl_a)
@@ -355,11 +359,35 @@ static bool component_hint_date_comparation(component_data_t *a, component_data_
 			scanlog("Component %s rejected by binary purl match\n", b->purls[0]);
 			return false;
 		}
+
+		if (b->rank >= COMPONENT_RANK_SELECTION_MAX && a->rank < COMPONENT_RANK_SELECTION_MAX)
+		{
+			scanlog("%s rejected by rank threshold %d >= %d\n", b->purls[0], b->rank, COMPONENT_RANK_SELECTION_MAX);
+			return false;
+		}
 	
 		//lower rank selection logic
-		if (b->rank < COMPONENT_RANK_SELECTION_MAX && b->path_depth <= a->path_depth)
+		if (b->rank <= COMPONENT_RANK_SELECTION_MAX)
 		{
 			scanlog("path lenght: %s - %d vs  %s - %d\n", b->file, b->path_depth, a->file, a->path_depth);
+						//shorter path lenght are prefered
+		if (b->path_depth < a->path_depth/2)
+		{
+			scanlog("%s accepted by shorter path depth %d vs %d\n", b->purls[0], b->path_depth, a->path_depth);
+			return true;
+		}
+		else if (a->path_depth < b->path_depth/2)
+		{
+			scanlog("%s rejected by longer path depth %d vs %d\n", b->purls[0], b->path_depth, a->path_depth);
+			return false;
+		}
+
+			if(b->path_depth > a->path_depth+1)
+			{
+				scanlog("%s rejected by deeper path in rank selection %d > %d\n", b->purls[0], b->path_depth, a->path_depth);
+				return false;
+			}
+
 			if (b->rank < a->rank)
 			{
 				scanlog("%s wins %s by rank %d/%d\n", b->purls[0],  a->purls[0], b->rank, a->rank);
@@ -377,9 +405,14 @@ static bool component_hint_date_comparation(component_data_t *a, component_data_
 	{	
 		if (purl_source_check(a) > purl_source_check(b))
 		{
-			scanlog("Component prefered by source\n");
+			scanlog("%s accepted over %s by source check\n", b->purls[0], a->purls[0]);
 			return true;
 		}
+		else if (purl_source_check(b) > purl_source_check(a))
+		{
+			scanlog("%s rejected by source check\n", b->purls[0]);
+			return false;
+		}
 
 		//Look for available health information
 		print_health(a);
@@ -404,6 +437,11 @@ static bool component_hint_date_comparation(component_data_t *a, component_data_
 			scanlog("Component %s prefered over %s by vendor+component=purl\n", b->purls[0], a->purls[0]);
 			return true;
 		}
+		else if (purl_vendor_component_check(a) && !purl_vendor_component_check(b))
+		{
+			scanlog("Component %s rejected, %s wins by vendor+component=purl\n", b->purls[0], a->purls[0]);
+			return false;
+		}
 		
 		if (!a->purls_md5[0] && a->purls[0])
 		{
@@ -421,23 +459,39 @@ static bool component_hint_date_comparation(component_data_t *a, component_data_
 		
 		if ((!a->age && b->age) || b->age > a->age)
 		{
-			scanlog("Component %s prefered over %s by purl date\n", b->purls[0], a->purls[0]);
+			scanlog("Component %s prefered over %s by purl date (age: %ld vs %ld)\n", b->purls[0], a->purls[0], b->age, a->age);
 			return true;
 		}
+		else if ((!b->age && a->age) || a->age > b->age)
+		{
+			scanlog("Component %s rejected by purl date (age: %ld vs %ld)\n", b->purls[0], b->age, a->age);
+			return false;
+		}
 
 		if (b->age == a->age && !strcmp(a->component, b->component) && strcmp(a->version, b->version) > 0)
 		{
 			scanlog("Component %s prefered over %s by version\n", b->purls[0], a->purls[0]);
 			return true;
 		}
+		else if (b->age == a->age && !strcmp(a->component, b->component) && strcmp(b->version, a->version) > 0)
+		{
+			scanlog("Component %s rejected by version comparison\n", b->purls[0]);
+			return false;
+		}
 	}
 	/*select the oldest release date */
 	if (strcmp(b->release_date, a->release_date) < 0)
 	{
-		scanlog("Component %s prefered over %s by release date\n", b->purls[0], a->purls[0]);
+		scanlog("Component %s (rank %d) prefered over %s (rank %d) by release date\n", b->purls[0],b->rank, a->purls[0], a->rank);
 		return true;
 	}
+	else if (strcmp(b->release_date, a->release_date) > 0)
+	{
+		scanlog("Component %s (rank %d) rejected, %s (rank %d) wins by older release date\n", b->purls[0], b->rank, a->purls[0], a->rank);
+		return false;
+	}
 
+	scanlog("Component %s rejected, no criteria matched\n", b->purls[0]);
 	return false;
 }
 
@@ -718,7 +772,7 @@ void match_select_best(scan_data_t *scan)
 			if (match_component == best_match_component)
 				continue;
 
-			if (path_is_third_party(match_component->file) < path_is_third_party(best_match_component->file) || !strcmp(match_component->release_date, "9999-99-99"))
+			if (path_is_third_party(match_component) < path_is_third_party(best_match_component) || !strcmp(match_component->release_date, "9999-99-99"))
 				continue;
 
 			scanlog("%s - %s - %d - %d VS %s - %s - %d - %d\n",
@@ -736,7 +790,7 @@ void match_select_best(scan_data_t *scan)
 
 			//If the best match is not good or is not identified be prefer the candidate.
 			if ((!best_match_component->identified && match_component->identified) ||
-				(path_is_third_party(best_match_component->file) < path_is_third_party(match_component->file)))
+				(path_is_third_party(best_match_component) < path_is_third_party(match_component)))
 			{
 				scanlog("Replacing best match for a prefered component\n");
 				scan->matches_list_array[i]->best_match = item->match;
diff --git a/src/util.c b/src/util.c
index 2c6d509..8033fe5 100644
--- a/src/util.c
+++ b/src/util.c
@@ -332,9 +332,17 @@ void free_and_null(void ** pr)
     }
 }
 
-int path_is_third_party(const char* path)
+int path_is_third_party(component_data_t *comp)
 {
-    const char* patterns[] = {
+    if (comp->third_party_rank > 0)
+		return comp->third_party_rank;
+	
+	if (!comp->file)
+		return 0;
+	
+	char * path = comp->file;
+
+	const char* patterns[] = {
         // Explicit third-party naming
         "third_party",       // Covers third_party, ThirdParty, third-party via strcasestr
         "thirdparty",        // Alternative spelling
@@ -381,10 +389,10 @@ int path_is_third_party(const char* path)
         "plugin",            // Plugins (often third-party)
 
         "utils","lib", "components", "modules", "ext",
-        "test", "fixtures", "examples",
+        "fixtures", "examples",
         "files", "assets", "runtime",
         "subprojects", "managed", "local_packages", "published",
-        "driver", "libresources", "offloading"
+        "driver", "libresources", "offloading","documentation", "test"
     };
 
     const int numPatterns = sizeof(patterns) / sizeof(patterns[0]);

From dcf5c9e93eb31aad8eb5bcc21131510721b800e8 Mon Sep 17 00:00:00 2001
From: mariano scasso <mariano.scasso@scanoss.com>
Date: Wed, 19 Nov 2025 17:04:39 +0000
Subject: [PATCH 5/5] remove snippets above the min_match_lines limit

---
 src/match.c             | 2 +-
 src/snippet_selection.c | 3 +++
 2 files changed, 4 insertions(+), 1 deletion(-)

diff --git a/src/match.c b/src/match.c
index d0d1f07..6b0770f 100644
--- a/src/match.c
+++ b/src/match.c
@@ -798,7 +798,7 @@ void match_select_best(scan_data_t *scan)
 			}
 
 			//If best match has 20% more of hits do nothing.
-			if (best_match->hits >= match->hits * 1.2 && best_match_component->path_depth < match_component->path_depth)
+			if (best_match->hits >= match->hits * 1.2 && best_match_component->path_depth <= match_component->path_depth)
 				continue;
 
 			//if cantidate has 10% more of hits do not consider dates and switch
diff --git a/src/snippet_selection.c b/src/snippet_selection.c
index adeb307..0d29172 100644
--- a/src/snippet_selection.c
+++ b/src/snippet_selection.c
@@ -262,6 +262,9 @@ int ranges_assemble(matchmap_range *ranges, char *line_ranges, char *oss_ranges)
 				strcat(line_ranges, ",");
 			if (*oss_ranges)
 				strcat(oss_ranges, ",");
+			//discard snippets below the limit of detection
+			if (to - from < min_match_lines)
+				continue;
 
 			/* Add from-to values */
 			sprintf(line_ranges + strlen(line_ranges), "%d-%d", from, to);