Honor GCT header when parsing dataset rows

[ACastanza]

Currently GSEA Desktop will parse the GCT header and then throw an error if there are additional blank rows at the end of the gct file. I think that the correct behavior here would be to honor the header and exclude blank rows that occur after the end of the specified number of genes perhaps providing a warning message instead.

[davideby]

Yeah, I tend to agree. GCT was designed in a time of memory-constrained computers where it was more important to know how many items it should expect than to be more user-friendly. We can clearly change the focus there.

Skipping blank rows is simple thing.

I'm just not totally clear what to do if there are more "real" rows than specified. I would say that should still be an error.

Honestly we should just steer people to the TXT format since GCT doesn't bring much to the table at this point.