instabilty in head

[ghost]

I work for the US DOE Joint Genome Institute, where we assemble 100's of metagenomes and metatranscriptomes. One of our analysts checked out the repo, from head, a few weeks ago, and had no problems running MEGAHIT. Since then myself, and other analysts have checked out from head, and we have had problems running MEGAHIT. Would it be possible to tag a stable revision, or provide a stable build as a download? We would really like to use MEGAHIT for our assemblies, but are concerned about the stability of the master branch.

[voutcn]

First thank you for using our software.

I would conduct several tests before commit changes to the master branch. I did not encountered any problem on my CentOS server and MacBook 2013. I think in most case it works well. But I am sorry that it is not an easy job for me to test MEGAHIT across many platforms and data.

Would you please let me know what problem did you encounter by using the head? Compilation error, or run time error? It's better if you could send me some log files. If MEGAHIT ran well a few weeks ago in your environment, please download and use v0.1.2 from the release page before I fix the problem.

[ajtritt]

I apologize, I did not see the releases page, I just followed the directions on the Github homepage. The releases page provides me with what I need. Thanks for pointing that out.

Here is an example error message, that I was consistently getting. I was using commit deab846

[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[ERROR] Found a read of length 271 > max read length = 270
[Mon Dec 1 19:56:50 2014] Building graph for k = 21
[Mon Dec 1 20:08:14 2014] Assembling contigs from SdBG for k = 21
megahit_assemble: succinct_dbg.cpp:356: void SuccinctDBG::LoadFromFile(const char*): Assertion `word_read == word_needed_w' failed.
Error occurs when assembling contigs for k = 21
[Exit code -6]

To me, it doesn't seem to be an issue with reads being longer than the max read length, since I have assemblies that have completed using reads that are longer than the specified max read length.

I have switched to using release 0.1.2 and no longer have this problem.

[voutcn]

Are you using a 32-bit system? I've replace all fopen64 to fopen as the former is NOT a standard function in C. This probably caused the error. MEGAHIT is designed (and merely has been tested) for 64-bit system. I am sorry that I did not mention this in the README.