Segmentation Fault

[bobbyr...@gmail.com]

Hi There, me again!

I have attempted to start another bali-phy run with the same alphabet/s-model/I-model (have not changed anything since the previous run, except with a new FASTA file with different sequences). 

I am able to use the alignment-info command to get an overview however, once i try to start the bali-phy run i get a segmentation fault and the program exits. 

Is there any way to fix this? Is it perhaps something wrong with my fasta file (as far as I can tell, it is in the same format as any of the other fasta files i have used for Bali-phy - which have all worked fine)?

All the best,

Ben

[Benjamin Redelings]

Hi Ben,

Sorry to take so long to get back to you!

A segmentation fault always indicates a mistake in the program - it should at least give an error message if something is wrong.

I should be able to fix it if I can get the segmentation fault myself. Would I be right in assuming that you are using version 3.5.0 on windows? Does it default every time, or did it just happen once? Can you copy and paste the exact command line that to used?

If you can also email me (privately) the fasta file, that might help.

-BenRI

--
You received this message because you are subscribed to the Google Groups "bali-phy-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bali-phy-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bali-phy-users/143743ae-9432-44ee-92a2-de01467ec958o%40googlegroups.com.

[Aaron Lewis]

Hello,

I have also encountered this problem, running the precompiled 64-bit Linux binary for version 3.5.0. The seg fault happens after the output directory and Main.hs file is created; no error logs or other messages. As with the original post, the alignment can be read and manipulated by the alignment-* tools without issue. Any tips? Happy to provide the input. (Hopefully the answer isn't 'upgrade to 3.6.0', because our computing cluster's libc and gcc are too far out of date.)  

Cheers

Aaron

[Benjamin Redelings]

Hi Aaron,

Thanks for the bug report!  Yes, I will need the input in order to debug this.  Can you send me the input files and the command line that you used? 

I sympathize about clusters with ancient libc versions.  Sometimes you can work around this by compiling on a different computer that has an old libc but a newer gcc.  What version of libc does your cluster have?

-BenRI

To view this discussion on the web visit https://groups.google.com/d/msgid/bali-phy-users/30fa4c8c-741b-4598-8a88-46ee4354c38cn%40googlegroups.com.

[Benjamin Redelings]

Hi Aaron,

Thanks for the files!  Yeah, I can reproduce the segfault, so I will fix it for the next version.  Since the workaround is pretty easy, I may either make a 3.6.1 or wait for 4.0

BTW. you can probably install bali-phy and a number of other packages on your cluster using linuxbrew.  Perhaps I should mention this on the installation page.

-BenRI

[Aaron Lewis]

Thanks for the linuxbrew tip. As workarounds go, for anyone else that finds this: the seg fault was caused by duplicated sequence names, and finding and removing or renaming dupes is easy with seqkit.