Seg fault on the following file

[earonesty]

Current version seg-faults on this read.. not sure why

@hwi:1:X:1:2211:11960:84344 3:N:0:
NNATNNTA
+

In order to get the seg fault, you have to compress and then uncompress the file.

[earonesty]

Not just on it's own... in a combination with other reads... for example this (poor performing) barcode set:

@hwi:1:X:1:2211:11960:84344 3:N:0:
NNATNNTA
+
########
@hwi:1:X:1:2211:11865:84347 3:N:0:
NNAGNNGC
+
########
@hwi:1:X:1:2211:11830:84349 3:N:0:
NNATNNTA
+
########
@hwi:1:X:1:2211:11979:84355 3:N:0:
NNTANNCT
+
########
@hwi:1:X:1:2211:11822:84366 3:N:0:
NNAANNTT
+
########
@hwi:1:X:1:2211:11949:84370 3:N:0:
NNTANNCT
+
########
@hwi:1:X:1:2211:11969:84371 3:N:0:
NNTANNCT
+
########
@hwi:1:X:1:2211:11845:84380 3:N:0:
NNTANNCT
+
########
@hwi:1:X:1:2211:11989:84381 3:N:0:
NNAGNNGC
+
########
@hwi:1:X:1:2211:11790:84386 3:N:0:
NNTANNCT
+
########

[earonesty]

I think the issue is the ##### quality scores. if you set even a single reads quality higher, it works.

[earonesty]

In order to get the seg fault, you have to compress and then uncompress the file!

[lrog]

Thanks for detailed bug report!

The issue was in the special case of the qualities encoding, having only '#' symbols in quality stream and some 'N' symbols in DNA stream.

I committed the fix in 464d5f3 to the quality-fix branch.

The fix passes the case you presented. I run the tests on the selected datasets from Illumina, 454 and ColorSpace and it also passed.

Could you please check if the fix also works in your full dataset? Then I'll merge the fix to master.

[earonesty]

It does. Yes.

[lrog]

Fix merged to master. Thanks!