Forum:possible bug: can make minia run forever with a bogus FASTA

Question

[adam.ewing]

I tested minia (2.0.2, 64 bit, precompiled binary) on a FASTA file with one entry consisting of a single base i.e.:

``1

test A ```

Which leads to it running forever while outputting status messages. Obviously this isn't a significant issue as nobody in their right mind would feed such a file to minia, but in an automated setting this can happen if something goes wrong upstream. Anyway, thanks for a great assembler!

-Adam

[Dan D]

It's great that you posted this here. Other users of minia might find it useful if their execution of it hangs. I changed the post type to "forum" since you're not really asking a question.

However, if you want a high chance of getting the developer's attention, you will probably have better results if you report this as a bug on the github repository.

[Rayan Chikhi]

No, here is perfect. I get notifications for any post with the "minia" tag. That github repo, however, isn't for Minia itself, but for a docker image that wraps it. The actual git repo of Minia is hosted by our institution (Inria), not Github, for now.

Answer 1

[Rayan Chikhi]

Thanks for the feedback!

We've recently noticed a possibly related bug with tiny k values, due to optimizations in the k-mer counter. Values of k<=4 will be forbidden in Minia version above 2.0.3, which should address this issue as well.