FAMD -  Bugs and Known Issues



If you have spotted a bug, please let me know. I would greatly appreciate if you could let me have as detailed a description as possible. Please let me know your FAMD version/date, operating system version, processor type, a description of what happened and what error messages (if any) you observed. If the bug is reproducible, I would appreciate if you could let me have a copy of the input file you were working on, and what operations you were performing. (Starting from FAMD 1.03dev, you can use the "log to file" function to document what FAMD did). 


Known Bugs. Version numbers indicate the version where a bug was first found. Once fixed in a particular FAMD version, it should also be fixed in subsequent releases.


FAMD 1.01 - Labelling may get confused if data was loaded without locus/individual names and items are deleted from matrix. Status: To my knowledge, this doesn't happen in FAMD 1.02. From v. 1.03 on, this will be fixed once and for all. 

FAMD 1.02 - Crashes if there are certain combinations of delimiter characters in the input file. Status: The source (a filtering routine) has been improved and will is incorporated in FAMD 1.03dev03.

FAMD 1.02 - Crashes if you try to declare an entire line (or more) of the input file as a remark ("{ ...}"). Status: Source of the problem is known, but not yet fixed.

FAMD 1.02 - UPGMA would fail in some medium-large data sets/program trapped in an endless loop. Status: The routine for this (FindFirstLinkedSet) has been reimplemented. The bug was fixed in FAMD 1.02b

FAMD 1.03dev - When dealing with certain large data sets, UPGMA traps the program in another endless loop. Status: This bug has been located and should be should be fixed in FAMD 1.03. [The routine joining individuals in UPGMA only searched through the first FFFFh individuals]. 

FAMD 1.03dev - When dealing with certain large data sets, performing UPGMA results in an exception error ('Invalid Pointer Operation') either during UPGMA itself, or on exiting the program. Status: This has been fixed in FAMD 1.108.

FAMD 1.108 - When calculating Shannon's index, different values for the variance and standard deviation are displayed on the screen than are written to the analysis file. IMPORTANT:  The values in the analysis file are the correct ones! (The value on screen are square roots of the correct values). Status: Fixed in FAMD 1.20.

FAMD 1.20 - FAMD fails to display an error message if a filter (e.g. Group of individuals) is invalid before an analysis. Status: Fixed in FAMD 1.21.

FAMD 1.23 - FAMD fails to export hindex files. "Colour" button in PCoA viewer not displayed in selected colour. Status: Fixed in FAMD 1.25.

FAMD 1.25 - LocusSet selection did not work. Status: Fixed in FAMD 1.26.

FAMD 1.26 - Repeated calculation of 2-level AMOVAs could corrupt memory and lead to incorrect results. Status: Fixed in FAMD 1.27.

FAMD 1.29 - Number of CPU cores displayed incorrectly. Status: Fixed in FAMD 1.30.

FAMD 1.30 - Would not start ML hybrid index calculation when groups were not specified in a specific way. Would not calculate AMOVA if loci were previously removed. Status: Fixed in FAMD 1.31.

(c) Copyright 2005-2013 by pms.