update to fail on duplicate LN tags, with different LN values

samtools · Jan 27, 2025 · 51253f1 · 51253f1
1 parent c6c1d19
commit 51253f1
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 3 deletions.
diff --git a/header.c b/header.c
@@ -145,7 +145,7 @@ static int sam_hrecs_update_hashes(sam_hrecs_t *hrecs,
         const char *name = NULL;
         const char *altnames = NULL;
         hts_pos_t len = -1;
-        int r;
+        int r, dup = 0;
         khint_t k;
 
         while (tag) {
@@ -154,7 +154,12 @@ static int sam_hrecs_update_hashes(sam_hrecs_t *hrecs,
                 name = tag->str+3;
             } else if (tag->str[0] == 'L' && tag->str[1] == 'N') {
                 assert(tag->len >= 3);
+                hts_pos_t tmp = len;
                 len = strtoll(tag->str+3, NULL, 10);
+                if (tmp != -1 && tmp != len) {
+                    //LN again with different value, discard
+                    dup = 1;
+                }
             } else if (tag->str[0] == 'A' && tag->str[1] == 'N') {
                 assert(tag->len >= 3);
                 altnames = tag->str+3;
@@ -173,6 +178,12 @@ static int sam_hrecs_update_hashes(sam_hrecs_t *hrecs,
             return -1; // LN should be present, according to spec.
         }
 
+        if (dup) {
+            hts_log_error("Header includes @SQ line \"%s\" with duplicate LN: tag", \
+                          name);
+            return -1;
+        }
+
         // Seen already?
         k = kh_get(m_s2i, hrecs->ref_hash, name);
         if (k < kh_end(hrecs->ref_hash)) {

diff --git a/sam.c b/sam.c
@@ -1911,6 +1911,12 @@ static sam_hdr_t *sam_hdr_sanitise(sam_hdr_t *h) {
         cp[h->l_text] = '\0';
     }
 
+     if (sam_hdr_fill_hrecs(h) < 0) {
+        //failed in parsing / validation
+        sam_hdr_destroy(h);
+        return NULL;
+    }
+
     return h;
 }
 
@@ -1964,8 +1970,16 @@ static sam_hdr_t *sam_hdr_create(htsFile* fp) {
                     strncpy(sn, q, r - q);
                     q = r;
                 } else {
-                    if (strncmp(q, "LN:", 3) == 0)
-                        ln = strtoll(q + 3, (char**)&q, 10);
+                    if (strncmp(q, "LN:", 3) == 0) {
+                        hts_pos_t tmp = strtoll(q + 3, (char**)&q, 10);
+                        if (ln != -1 && tmp != ln) {
+                            //duplicate LN tag with different value
+                            hts_log_error("111Header includes @SQ line \"%s\" with \
+duplicate LN: tag", sn);
+                            goto error;
+                        }
+                        ln = tmp;
+                    }
                 }
 
                 while (*q != '\t' && *q != '\n' && *q != '\0')