segfault in trio graph output stage

[ASLeonard]

Hi,

I get a segfault near the end when assembling with v0.19.5-r592. This sample is a mix of Sequel IIe and Revio, but I've assembled several other genomes with just SIIe, just Revio, and a mix, so this appears to be limited to this one sample (and previously worked for only assembling the Revio data for this sample).

Here are the last few lines of the log file. I checked the written asm.bp.p_ctg.gfa file and it all looks as expected for size and contiguity, so this looks to be localised to the graph unzipping.

Writing asm.bp.p_ctg.gfa to disk... 
[M::reduce_hamming_error_adv::2.880] # inserted edges: 74780, # fixed bubbles: 370
[M::adjust_utg_by_trio] primary contig coverage range: [44, infinity]
[M::recall_arcs] # transitive arcs::2066
[M::recall_arcs] # new arcs::364565172, # old arcs::203094
[M::clean_trio_untig_graph] # adjusted arcs::0
[M::adjust_utg_by_trio] primary contig coverage range: [44, infinity]
[M::recall_arcs] # transitive arcs::2080
[M::recall_arcs] # new arcs::368426000, # old arcs::203648
[M::clean_trio_untig_graph] # adjusted arcs::2
####ERROR1-fill: i: 77, v>>1: 10424839, v&1: 1, w>>1: 2768540, w&1: 0
####ERROR1-fill: i: 77, v>>1: 10424839, v&1: 1, w>>1: 2768540, w&1: 0
####ERROR1-fill: i: 77, v>>1: 10424839, v&1: 1, w>>1: 2768540, w&1: 0
####ERROR1-fill: i: 77, v>>1: 10424839, v&1: 1, w>>1: 2768540, w&1: 0
####ERROR1-fill: i: 77, v>>1: 10424839, v&1: 1, w>>1: 2768540, w&1: 0
####ERROR1-fill: i: 77, v>>1: 10424839, v&1: 1, w>>1: 2768540, w&1: 0
[M::output_trio_graph_joint] dedup_base::6270217, miss_base::0
Writing asm.bp.hap1.p_ctg.gfa to disk... 
Writing asm.bp.hap2.p_ctg.gfa to disk... 
ERROR
ERROR
####ERROR1: v>>1: 10424839, v&1: 1, w>>1: 2768540, w&1: 0, r_seq: 10538420
ERROR
/cluster/apps/sfos/bin/bash: line 1: 61191 Segmentation fault

The data isn't public, but I can probably host the data and share a link with you by email if that helps.

Best,
Alex

[chhylp123]

Could you pelase have a try with the current github HEAD with bin files? Thanks in advance.

[ASLeonard]

This was with the current HEAD (d91fc50)