Home
Home
 
Bug Reports And Fixes

The following is a list of major bugs discovered in SIMMAP (Minor bugs, such as aesthetic problems are not reported as they don't affect the outcomes). Bugs are classified as Minor, Important, and Critical. As with any software the results should be scrutinized before publication and checked if a bug has been reported that might have an effect on your analysis.

Please report bugs to:


Unresolved Bugs

  • An issue exists when writing large files, > 1GB. The current suggested work around is to perform multiple analyses and combine the results.

    Fixed Bugs

  • 04 August 2008: (Minor) [Affects Beta 2.4, build 04082008-1.0-B2.4]
    Problem reading tree files from BEAST. This bug does not effect any analyses as the problem caused a full program crash. These problems have been fixed in build 20022009-1.0-B2.4.

  • 04 August 2008: (Minor) [Affects Beta 2.4, build 26062008-1.0-B2.4]
    Minor issues in the size of some the windows caused cliiping of their contents. These problems have been fixed in build 04082008-1.0-B2.4.

  • 26 June 2008: (Important) [Affects all Beta 2 versions]
    A bug in which the calculation of the conditional likelihoods for very large trees when performing molecular analyses would cause an underflow. This is not consider critical as the results produced are NAs rather than real values. However, large analyses (>200 species) should be checked.

  • 26 June 2008: (Minor) [Affects all Beta 2 versions run on Mac OS X 10.5]
    Apple's 10.5 version of the OS on Intel machines caused numerous issues with the interface. All of these issue pertained to displaying the interface and had no effect on the analyses. Fixed the interface issues associated with the previous version of SIMMAP running on machines operating the new version of the Mac OS (10.5). The primary issue pertained to the built in help documents. These are still available but will be displayed by the user's default web browser.

  • 26 June 2008: (Important) [Affects all Beta 2 versions]
    Problems when pruning taxa that would result in a program crash. Fixed the pruning taxa bug. The pruning taxa option now permanently removes both the taxa from trees and from the data matrix in memory. The original files are not effected.

  • 26 June 2008: (Important) [Affects all Beta 2 versions]
    Fixed the bug that would cause a crash after all of the files (molecular data) are closed and new files (particularly the parameter file) are opened.

  • 12 September 2006: (Critical) [Affects all Beta 2 versions]
    A bug was identified by a user in which when saving the summary statistics from the morphology statistics window the individual transformation values are incorrectly printed - some are miss-aligned with the headers and some are missing. The total number of transformations is reported correctly as are all of the values in the window display. You will need to repeat all analyses that have used this data. The bug does not affect these values if they were saved to a file during the analysis. This bug DOES NOT AFFECT molecular analyses or data/results from any other type of analysis.

    To check if your results are affected by this bug simply sum the individual transformation values. The value should be equal to the total number of transformations.

    Fixed in Beta 2.3 (Build 12092006-1.0-B2.3).

  • 06 September 2006: (Important) [Affects all Beta 2 versions]
    A bug causing the long runs of calculating ancestral states to crash at the end of the run.

    Fixed (likely) in Beta 2.3 (Build 06092006-1.0-B2.3).

  • 06 September 2006: (Important) [Affects Beta 2.2 version only]
    A bug was introduced during the 2.2 update that caused predictive analyses to crash.

    Fixed in Beta 2.3 (Build 06092006-1.0-B2.3).

  • 23 August 2006: (Minor) [Affects all Beta 2 versions]
    A bug causing a failure to report ancestral states for some constraints. The bug seems to have been intermitten and dependent on the constraint. Constraints which were not present in all trees seemed to be the only ones effected. The problem was that SIMMAP would fail to report any results for these constraints. This has been fixed such that any selected constraint that appears in at least one tree now successfully reports. This bug is considered minor because it has no effect on the quality of the results but involved a failure to report any information in the user interface.

    Fixed in Beta 2.2 (Build 23082006-1.0-B2.2).

  • 23 August 2006: (Minor) [Affects all Beta 2 versions]
    A bug causing the link to the Submit a Bug web page to direct the user to the old web site. This has been fixed to point to the new web site.

    Fixed in Beta 2.2 (Build 23082006-1.0-B2.2).

  • 19 May 2006: (Minor) [Affects all Beta 2 versions]
    A bug causing SIMMAP to crash when reading trees generated by BayesPhylogenies. This is considered minor because ir prevents the program from crashing and therefore could not result in reporting erroneous results. The following excerpts (or something similar) would have appeared in the CrashReporter log if you have experienced this problem:

    Exception: EXC_BAD_ACCESS (0x0001)
    Codes: KERN_INVALID_ADDRESS (0x0001) at 0xbf7ffff8

    Thread 0 Crashed:
    0 SMMP 0x00013334 getDownPassSequenceForTree + 0 (SMMPController.m:7553)

    OR

    Thread 0 Crashed:
    0 SMMP 0x000299f4 setUpStructureRepresentationOfTreeString + 444 (SMMPController.m:8147)

    Fixed in Beta 2.1 (Build 19052006-1.0-B2.1).

  • 09 April 2006: (Critical) [Affects all Beta 2 versions]
    A bug causing SIMMAP to miss-assign character states at the tips of the tree effecting all character analyses. The bug effects all tree files but is exacerbated when the indexing of the translate block starts at an index of 0 rather than 1. This has been fixed allowing indexing from zero but SIMMAP now checks that the translation is only numeric values. SIMMAP does not support other forms of translate blocks in the tree file. The bug seems to effect morphological data sets only but may have effects on molecular data. ALL analyses should be checked. Fixed in Beta 2.0.8 (Build 09042006-1.0-B2.0.8).

  • 30 March 2006: (Minor) [Affects Beta 2 versions (most builds)]
    Calculation of the frequency of a constraint changes on each update as a factor of the number of trees. Fixed as of Build 30032006-1.0-B2.0.7.

  • 18 March 2006: (Critical) [Affects Beta 2 versions (all builds)]
    Calculation of character correlations with some data sets (primarily morphological in my tests) and under some sampling designs don't evaluate the expectation of the statistics properly. Specifically, the expectation is not evaluated over the trees when sampling the predictive (null) distribution. Results generated with these versions and possibly previous versions should be checked against Beta 2.0.6+.

  • 14 March 2006: (Minor) [Affects Beta 2 versions 2.0.3, 2.0.3.1, and 2.0.4 (all builds)]
    When saving raw character histories for morphological data SIMMAP fails to write the tree histories. The output file would only contain the begin trees and translate block. This bug did not effect molecular data. This bug has been fixed as of Build 14032006-1.0-B2.0.5 and later versions and builds. (Thanks to Kim Roelants for reporting this bug.)

  • 13 March 2006: (Important) [Affects Beta 2 version, Build 24022006-1.0-B2.0.4]
    A number of new features were being added to the morphology priors but were not deactivated for this update. The features are fixed Beta prior, equal Beta prior, fixed rate prior, and no rate prior. While these options seemed avalailble they had no effect on the analysis which would use the prior values from the Beta and Gamma distributions. These features have been added and as of Build 11032006-1.0-B2.0.5 are available and active.


    The following fixes are current for Build 31012006-1.0-B2.0.4 and all later builds and versions.

  • BUG & FIX:
    Fixed a bug in reporting the individual character history results for morphological data. The values reproted were not written to the file in the correct order (i.e., column) and in some cases were reproted as zero. This has been fixed. This bug should have only effected the files written by versions > than Beta 2.0.0 although it isn't clear exactly when this bug appeared. Saving summary data from the statistics window was not effected nor was molecular data. (Thanks to Stefan Ekman for reporting this bug.)

  • BUG & FIX:
    Fixed a bug that caused a program crash or hang when taxon names in the tree file consisted of numbers only or numbers in the first couple of character positions. This was a problem in a sort function and should have not affected other features.

  • BUG & FIX:
    Fixed a bug that caused a program to crash when more than 400 species or so are being used. Was a problem with underflow. This has been corrected for both molecular and morphological characters.

  • BUG & FIX:
    Made reading trees in the Newick format more robust. For example, some rooted trees failed to read. Tree reading should be completely complaint with bifurcating trees in Newick format. You can not simply make the branch lengths zero to trick SIMMAP into thinking the tree is truly bifurcating. While this might work when the states at each node happen to be the same the program will go into and infinite loop when the states are different.

  • BUG & FIX:
    Fixed a bug that led to a program crash when ancestral states were requested but a character map hadn't been simulated. Part of this fix is that by default character one is selected initially.

  • BUG & FIX:
    Some outgroup configurations, while monophyletic, wouldn't root. This has been fixed along with adding new options for rooting trees.

  • BUG & FIX:
    When requesting an internal node's states for a particular character be fixed they will revert to being uncertain (i.e., unfixed). This has been fixed and was perhaps the most serious bug in the last version. It didn't effect character histories just ancestral states option. In addition to the bug fix, ancestral states has been re-worked.

  •     
    About Search System Requirements License Acknowledgements Contact
     

    Page Last Updated: 6 August 2008