BWTs on disk since OSX doesn't implement certain semaphores' functions.
…tead of aborting
… header line of the fasta file.
This option limits the number of walks that are output for complex components.
…threaded to be faster. Is there any reason not to do this?
…at is passed in is a substring of a read.
…n the input scaffold/contig collection
…in the case where it was installed with make install
For very complex graphs with repetitive genomes, Myers' transitive reduction algorithm can take a very long time to complete. Since sga overlap does not output transitive edges, in the general use case Myers' algorithm is not needed. So it is now turned off by default. Second, the small repeat resolution algorithm can also take a long time to run. I'm attempting a fix here where I limit to run on not-too- complex vertices.
…oice a command line argument
…2-bit encoded strings everywhere
…rects the library link ordering and fixes the build in the case --as-needed is used in ld, as in newer versions of gcc.
…dges when --component-walks is specified.
…ead ID in sga preprocess