X-Git-Url: http://woldlab.caltech.edu/gitweb/?p=samtools.git;a=blobdiff_plain;f=bam_index.c;h=9610a2656031657b796abc68bf32c1b6f1564745;hp=f60a6f874b106f03c655d1207dddf21c4df7fc48;hb=0f906dafb2ad22371a753557562ef95c3034480d;hpb=6a0c6f060a60789b48e10a72b1381f6e54599302

diff --git a/bam_index.c b/bam_index.c
index f60a6f8..9610a26 100644
--- a/bam_index.c
+++ b/bam_index.c
@@ -172,19 +172,23 @@ bam_index_t *bam_index_core(bamFile fp)
 
 	save_bin = save_tid = last_tid = last_bin = 0xffffffffu;
 	save_off = last_off = bam_tell(fp); last_coor = 0xffffffffu;
-    n_mapped = n_unmapped = n_no_coor = off_end = 0;
+	n_mapped = n_unmapped = n_no_coor = off_end = 0;
 	off_beg = off_end = bam_tell(fp);
 	while ((ret = bam_read1(fp, b)) >= 0) {
 		if (c->tid < 0) ++n_no_coor;
-		if (last_tid != c->tid) { // change of chromosomes
+		if (last_tid < c->tid || (last_tid >= 0 && c->tid < 0)) { // change of chromosomes
 			last_tid = c->tid;
 			last_bin = 0xffffffffu;
-		} else if (last_coor > c->pos) {
+		} else if ((uint32_t)last_tid > (uint32_t)c->tid) {
+			fprintf(stderr, "[bam_index_core] the alignment is not sorted (%s): %d-th chr > %d-th chr\n",
+					bam1_qname(b), last_tid+1, c->tid+1);
+			return NULL;
+		} else if ((int32_t)c->tid >= 0 && last_coor > c->pos) {
 			fprintf(stderr, "[bam_index_core] the alignment is not sorted (%s): %u > %u in %d-th chr\n",
 					bam1_qname(b), last_coor, c->pos, c->tid+1);
-			exit(1);
+			return NULL;
 		}
-		if (c->tid >= 0) insert_offset2(&idx->index2[b->core.tid], b, last_off);
+		if (c->tid >= 0 && !(c->flag & BAM_FUNMAP)) insert_offset2(&idx->index2[b->core.tid], b, last_off);
 		if (c->bin != last_bin) { // then possibly write the binning index
 			if (save_bin != 0xffffffffu) // save_bin==0xffffffffu only happens to the first record
 				insert_offset(idx->index[save_tid], save_bin, save_off, last_off);
@@ -203,7 +207,7 @@ bam_index_t *bam_index_core(bamFile fp)
 		if (bam_tell(fp) <= last_off) {
 			fprintf(stderr, "[bam_index_core] bug in BGZF/RAZF: %llx < %llx\n",
 					(unsigned long long)bam_tell(fp), (unsigned long long)last_off);
-			exit(1);
+			return NULL;
 		}
 		if (c->flag & BAM_FUNMAP) ++n_unmapped;
 		else ++n_mapped;
@@ -212,13 +216,20 @@ bam_index_t *bam_index_core(bamFile fp)
 	}
 	if (save_tid >= 0) {
 		insert_offset(idx->index[save_tid], save_bin, save_off, bam_tell(fp));
-		insert_offset(idx->index[save_tid], BAM_MAX_BIN, off_beg, off_end);
+		insert_offset(idx->index[save_tid], BAM_MAX_BIN, off_beg, bam_tell(fp));
 		insert_offset(idx->index[save_tid], BAM_MAX_BIN, n_mapped, n_unmapped);
 	}
 	merge_chunks(idx);
 	fill_missing(idx);
-	if (ret >= 0)
-		while ((ret = bam_read1(fp, b)) >= 0) ++n_no_coor;
+	if (ret >= 0) {
+		while ((ret = bam_read1(fp, b)) >= 0) {
+			++n_no_coor;
+			if (c->tid >= 0 && n_no_coor) {
+				fprintf(stderr, "[bam_index_core] the alignment is not sorted: reads without coordinates prior to reads with coordinates.\n");
+				return NULL;
+			}
+		}
+	}
 	if (ret < -1) fprintf(stderr, "[bam_index_core] truncated file? Continue anyway. (%d)\n", ret);
 	free(b->data); free(b);
 	idx->n_no_coor = n_no_coor;
@@ -466,6 +477,10 @@ int bam_index_build2(const char *fn, const char *_fnidx)
 	}
 	idx = bam_index_core(fp);
 	bam_close(fp);
+	if(idx == 0) {
+		fprintf(stderr, "[bam_index_build2] fail to index the BAM file.\n");
+		return -1;
+	}
 	if (_fnidx == 0) {
 		fnidx = (char*)calloc(strlen(fn) + 5, 1);
 		strcpy(fnidx, fn); strcat(fnidx, ".bai");
@@ -698,6 +713,7 @@ int bam_fetch(bamFile fp, const bam_index_t *idx, int tid, int beg, int end, voi
 	b = bam_init1();
 	iter = bam_iter_query(idx, tid, beg, end);
 	while ((ret = bam_iter_read(fp, iter, b)) >= 0) func(b, data);
+	bam_iter_destroy(iter);
 	bam_destroy1(b);
 	return (ret == -1)? 0 : ret;
 }