Versions 4.4.*
Version 4.4.0
The new version of the library is issued 2017/08/27:
- Correction of a bug in the Result class “multi iterator”. This correction should fix the
problem in “SGI_SR” predefined criterion.
- In COM component: correction of a bug in the conversion of complex vector to variant.
- Correction for the reading from XDB files of complex stresses in CBEAM elements (MP
format).
- In “DataBase” class of the component, the “NbrAbbreviations”, “NbrGroups”,
“NbrCoordSys”, “NbrElements”, “NbrNodes”“get*” methods are replaced by property
“getters”.
- Several bug fixes.
- The most significant innovation is that FeResPost is also distributed as a Python library.
Version 4.4.1
The new version of the library is issued 2017/10/22:
- Modification of the “WriteGmsh” method in “DataBase” class. A check of the Result
values is done and infinite of NaN values are substituted with MAXFLOAT and
MINFLOAT respectively.
- Correction of string conversion in PYTHON version > 3. C/C++ strings are now
converted to Unicode strings, and no longer to bytes.
- The problems related to the different operators in the Python extension “Result” class
have been fixed. It is now possible to use the “+”, “*”, “*”, “/” and “**” operators.
The corresponding “opAdd”, “opSub”, “opMul”, “opDiv”, “opPow” methods have been
removed. (This means you will have to modify your Python scripts if you have used these
methods.)
- Correction of the “extractLayers” and “extractSubLayers” methods in “Result” class.
The methods return the “NONE” layers or cub-layers, and corrections are done in COM
component and .NET assembly.
Version 4.4.2
The new version of the library is issued 2018/01/01:
- In ruby extension, addition of a “coerce” method in the “Result” class for the
management of operators with “Result” object as second argument. (See section I.7.)
The standard classes are no longer modified by “modifStdClasses.rb”.
- Addition of “opPow” method in “Post” Module.
- In the .NET assembly, for compatibility with other FeResPost libraries, one has defined
the “opAdd”, “opSub”, “opMul”, “opDiv” and “opPow” methods in the “Post” Module.
- Correction of a bug in the reading of Results from XDB files produced with option
“DBCFACT=4” (BBBT accessor). The reading failed when the table fitted a single page
(output of a small amount of data).
- Correction of a bug in method “deriveTensorToThreeScals” of “Result” class.
- Addition of method “eigenQR” to “Result” class.
- In Windows binaries, python extensions are now compiled with GNU compiler.
Version 4.4.3
The new version of the library is issued 2018/04/29:
- Correction of a bug in the “insertLaminate” of the “ClaDb” class. The method no longer
returns an exception when laminate properties calculation fails.
- Correction of a bug in the reading of text files on UNIX and LINUX systems. The DOS
“end-of-line characters” were not well supported. This bug fix concerns the reading of
Nastran BDF files, Samcef DAT files, composite databases from EDF or NDF files and
groups from Patran session files.
- correction of a bug in the generation of GMSH “skeletons”. (Segmentation fault occured
if one node used in element definition was missing.)
- Modification of “LaminateAnalysis.xls” example. Now the material and laminate thermal
and moisture properties are also calculated in the corresponding spreadsheets. Also, one
improved the calculation of minimum or maximum composite criteria.
- A few corrections in the User Manual. For example, some remarks have been added to
the description of CBUSH element forces and moments read from Nastran results.
- Addition of a few checks on “NULL” arguments for “claDb” and "ClaLoad" arguments.
- Correction of a major bug in Python complement. (The bug prevented numerical
operations of Result class objects with other type’s arguments.) This bug concerns the
version 2.* of Python.
- Several other bugs have been fixed in the Python complement.
- Addition of an object oriented post-processing in the Python examples (section V.1.4).
- Several errors have been correction in the Python examples.
Version 4.4.4
The new version of the library is issued 2018/08/05:
- Correction of a bug in reading of CBUSH element complex stresses from XDB file.
- Addition of method “importAttachmentResults” in “NastranDb” class.
- Renaming the complex nodal Results read from XDB. The (RI) or (MP) is now at the
end of Result name, and no longer at the beginning. Examples and manual are updated
accordingly.
- Correction of a bug in the “deriveScalToScal” method in “Result” class. (The method
was completely wrong when applied to complex results.)
- Correction of a bug in “deriveVectorToScal” method in “Result” class. The format of
Result object is always set to Real for “abs” and “sq” derivation, even if the argument is
a Complex vectorial Result.
- Correction of a bug in the “max” and “min” dyadic derivation methods of “Result” class.
Results were not correction when keys belonged to only one of the Results, and when the
Results were tensorial or vectorial.
- Modification of the “compare” method behaviour for “deriveDyadic” in “Result” class.
Now the method can also be used with vectorial or tensorial Results. Then, the
comparison is done on a component-per-component basis.
- Modification of the “writeGmsh” method in generic “DataBase” class. The method now
also accepts the “Elements” as output location for the Results.
- Correction of a bug in the “writeGmsh” method in generic “DataBase” class. The
function should no longer crash when one attempts to output empty or nil Results in
GMSH files.
- Correction of a bug in the reading of composite stresses/strains from Nastran XDB and
OP2 files. When inter-laminar results are read, a rotation of the shear components is
now done to obtain results in ply axes. (Nastran produces inter-laminar shear results in
laminate axes.)
- Correction of a bug in the reading of Nastran XDB result files produced with option
“DBCFACT=4”. Results directly written in the indexation data pages, isntead of in
separate data blocks are now read correctly. This is the case, for example, for the
spring scalar forces. (We suspects Nastran does not create separate data pages when
corresponding data are small and can be stored in a number of words <=2.)
- A few other improvements of the XDB access with option “DBCFACT=4”.
- The laminate interlaminar shear criteria are no longer calculated at bottom layer of
laminates bottom ply.
- Addition of a test on the number of components of values inserted into “Result” objects.
- Correction of a bug in the reading of nonlinear stresses and strains from Nastran XDB
files.
- Addition of several methods in the “Result” class: “cloneNoValues”,
“insertResultValues” and “removeKeysAndValues”.
- Production of compiled extensions for ruby version 2.5.* and Python version 3.7.*.
(However, I have problems to run the FeResPost extension with versions 2.4.* and 2.5.*
of ruby on windows OS.)
- A few corrections in User Manual.