segfault when making a workspace including a long read datastore

[jonwright99]

To reproduce;

1. Make a PE datastore
   sdg-datastore make -1 PE_R1.fastq -2 PE_R2.fastq -t paired -d 1 -n pe -o pe.ds

2. Make a PacBio datastore
   sdg-datastore make -L pb_reads.fastq -t long -n pacbio -o pacbio.ds

3. Make ws with PE reads only
   sdg-workspace make -g k200.contigs_raw.gfa -p pe.ds.prseq -o pe_ws
   ** runs fine

4. Make a ws with PE and PB reads
   sdg-workspace make -g k200.contigs_raw.gfa -L pacbio.ds.loseq -p pe.ds.prseq -o pe_pb_ws

Step 4 gives a segfault although stdout suggests the ws has been built correctly;

/hpc-home/wrightj/github/bioinfologics/sdg/build/sdg-workspace make -g k200.contigs_raw.gfa -L pacbio.ds.loseq -p pe.ds.prseq -o pb_pe_ws

2020-05-29 08:10:24: Graph fasta filesname: k200.contigs_raw.fasta
2020-05-29 08:10:24: Loading sequences from k200.contigs_raw.fasta
2020-05-29 08:10:28: 1024254 nodes loaded (246228 canonised).
2020-05-29 08:10:34: 1024254 nodes after connecting with 1348742 links.
2020-05-29 08:10:34: PairedReadsDatastore open: pe.ds.prseq  max read length: 251 Total reads: 168907918
2020-05-29 08:10:34: LongReadsDatastore open: pacbio.ds.loseq Total reads: 10000000
2020-05-29 08:10:34: Dumping workspace to pe_pb_ws.sdgws
2020-05-29 08:10:36: Dumping long read mappings
2020-05-29 08:10:36: Done!