Fix bed_index_core() handling of overlapping entries

It needs to skip if the current entry ends before the previous
one, which can happen if the input BED file has entries that
are completely contained in an earlier region.  Failing to
do this could cause last_end to jump backwards, resulting
in some incorrect entries being put in the index.  This could
lead to hits between the end of the innermost and outermost
nested BED entries being missed.

Fixes samtools#2104 ("samtools view -L " incomplete output for BED with
nested targets)

bedidx.c

@@ -110,6 +110,8 @@ static int bed_index_core(bed_reglist_t *regions)
         hts_pos_t beg = a[i].beg >= 0 ? a[i].beg >> LIDX_SHIFT : 0;
         hts_pos_t end = a[i].end >= 0 ? a[i].end >> LIDX_SHIFT : 0;
         hts_pos_t j;
+        if (end < last_end)
+            continue;  // Can happen for a containment
         if (end + 1 >= SIZE_MAX / sizeof(*idx)) { // Ensure no overflow
             errno = ENOMEM;
             free(idx);

test/dat/nested.bed

@@ -0,0 +1,3 @@
+ref2	1	436871168	reg1a
+ref2	8192	8193	reg1b
+ref2	541556270	541556290	reg3

test/dat/nested.expected.sam

@@ -0,0 +1,34 @@
+@HD	VN:1.4	SO:coordinate
+@RG	ID:grp1	DS:Group 1	LB:Library 1	SM:Sample
+@RG	ID:grp2	DS:Group 2	LB:Library 2	SM:Sample
+@RG	ID:grp3	DS:Group 3	LB:Library 3	SM:Sample
+@PG	ID:prog1	PN:emacs	CL:emacs	VN:23.1.1
+@CO	
+@CO	Copyright (C) 2014 Genome Research Ltd.
+@CO	
+@CO	Permission is hereby granted, free of charge, to any person obtaining
+@CO	a copy of this software and associated documentation files (the
+@CO	"Software"), to deal in the Software without restriction, including
+@CO	without limitation the rights to use, copy, modify, merge, publish,
+@CO	distribute, sublicense, and/or sell copies of the Software, and to
+@CO	permit persons to whom the Software is furnished to do so, subject
+@CO	to the following conditions:
+@CO	
+@CO	The above copyright notice and this permission notice shall be
+@CO	included in all copies or substantial portions of the Software.
+@CO	
+@CO	THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+@CO	EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+@CO	MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+@CO	IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+@CO	CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+@CO	TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+@CO	SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+@CO	
+@SQ	SN:ref1	LN:56	M5:08c04d512d4797d9ba2a156c1daba468
+@SQ	SN:ref2	LN:541556283	M5:7c35feac7036c1cdef3bee0cc4b21437
+ref2_grp3_p001	83	ref2	1	99	15M	=	31	45	ATTCTATAGTGTCAC	~~~~~~~~~~~~~~~	RG:Z:grp3	NM:i:0	MD:Z:15
+ref12_grp1_p001	145	ref2	2	50	10M	ref1	36	0	TTCTATAGTG	BBBBBBBBBB	RG:Z:grp1	NM:i:0	MD:Z:10
+ref12_grp2_p001	145	ref2	12	50	10M	ref1	46	0	TCACCTAAAT	BBBBBBBBBB	RG:Z:grp2	NM:i:0	MD:Z:10
+ref2_grp3_p002	147	ref2	436870911	99	15M	=	46	45	CTAAATAGCTTGGCG	}}}}}}}}}}}}}}}	RG:Z:grp3	NM:i:0	MD:Z:15
+ref2_grp3_p002	99	ref2	541556280	99	15M	=	16	-45	CTGTTTCCTGTGTGA	{{{{{{{{{{{{{{{	RG:Z:grp3	NM:i:0	MD:Z:15

test/test.pl

@@ -2368,6 +2368,15 @@ sub test_view
         }
     }
 
+    # -L option with nested regions
+    run_view_test($opts,
+		  msg => "$test: -L with nested regions",
+		  args => ['-h', '-L', "$$opts{path}/dat/nested.bed", '--no-PG',
+			   "$$opts{path}/dat/large_chrom.sam"],
+		  out => sprintf("%s.test%03d.sam", $out, $test),
+		  compare => "$$opts{path}/dat/nested.expected.sam");
+    $test++;
+
     # -T / -t options
     my $sam_no_sq = "$$opts{tmp}/view.001.no_sq.sam";
     filter_sam($sam_no_ur, $sam_no_sq, {no_sq => 1});