Move source repository to https://github.com/sys-bio/antimony/
Disallow non-ASCII characters when otherwise not a part of a string or anything else.
Update local python installation rules so that it will create explicit 'wheel' files.
Change installation defaults: turn on libsbml compression, since that's on by default in libsbml, and do the local Python install instead of the global one by default.
Fix warnings from mac build, most of which involved the fact that 'varConstraint' wasn't in a bunch of 'switch' statements. This actually fixed a bug that I hadn't realized was there, when flattening models with constraints.
Updates to make antimony compile and install on unix.
Can't use '_strdup' on linux/gcc.
Update python binding and CMake system.
Add API test.
Install libantimony file with python bindings.
Create new Antimony python bindings based on ctypes instead of swig. Remove the old python stuff.
Various fixes:
Update installer files.
Remove warnings, and fix one warning that was actually a bug.
Various fixes:
Solved via https://stackoverflow.com/questions/39689162/qt-project-in-visual-studio-2015-unresolved-external-symbol-wwinmain and https://stackoverflow.com/questions/18067870/what-option-do-i-set-to-use-main-instead-of-wmain namely: add 'mainCRTStartup' to Visual Studio/QTAntimony/Properties/Linker/Advanced/Entry point Still need to figure out how to do this in CMake, so others don't run into the issue.
QTAntimony won't compile in Release mode
Remove Boolean errors from setLevelAndVersion, since we already were removing them from the base version.
Update to allow compilation on latest experimental release of libsbml, instead of latest SVN only.
Fix warnings.
Update to v2.12; update documentation.
More memory leak fixes (for error conditions).
Fix memory leak, plus mismatched free/delete.
Fix wrong way to get the module name.
Change includes to be correct format with correct slash for unix.
Initialize member variables that need values.
Add the ability to read/write 'distrib' style annotations (mean, stdev, etc.) Many tests added as well. Removed ability to define units in curly brackets, since that interfered with libsbml's vector reading for arrays, but I don't think anyone was using it, anyway.
Add test for unary negative of parentheticals, which had been broken.
Antimony now saves function names, so they're preserved across the translations.
removed BECOMESMOD token, which prevented parsing of unary negation + parentheses; -( reaction divider can still be parsed as separate tokens
Clean up parsing of module/function names, and make function names actually work.
Error out if functionname.something = 3 is used, if 'something' isn't 'sboTerm'.
Error out if modulename.something = 3 is used, if 'something' isn't 'sboTerm'.
Adjust error messages; update tests.
* Check the type of reactants as we add them.
Disallow synchronizing an SBO term with something.
* Add restriction that variables without mathematical meaning cannot be used in equations.
More memory leak fixes on erroring.
If a submodel has both an sbo term and a variable named 'sboTerm', try other names to set the term.
Fix memory leaks for errors.
Add missing files.
A model's sbo term may fall after 'end', so replace that, too.
Add new files.
Revert this--broke tests.
* Consolidate getting/setting annotation information.
Add reaction/SBO test.
Add function/sbo test.
Various fixes for setting SBO terms, plus added tests for everything:
Fix some issues with setting the SBO term, including making it caseless (when it's clearly not anything else), and moving the sboTermWrapper class to its own files.
Fix naming modules so it actually works. Add tests for this and for incorrect CV terms.
Update version number for tests.
Revert commenting the tests.
Add tests for Kyle's new CV Terms syntax.
Support for python 3.7 (pip packages)
pass second string arg to new module
Bump version number to 2.11.0
Remove some shift/reduce conflicts by centralizing inequalities to be similar to mathThings, and pushing some parsing to the lexer.
Change parsing so that right-side-only reactions work! Add tests for it and everything.
Protect against future oddnesses with metaid differences; make test more similar across types.
More leaks fixes in tests.
More leaks fixes in tests.
More leaks fixes in tests.
More leaks fixes in tests.
More leaks fixes in tests.
* Fix more leaks, UnitDef's in particular. Cleaner architecture where there used to be leaks.
Fix leaks in test code.
Add required includes.
Pull in libsbml's new 'WITH_CHECK' system for linux, so the right libraries are included.
Fix memory leak of ASTNode.
Update libcheck bit of CMakeLists to match libsbml. Still doesn't solve the linking issue, but hey.
Memory leak plugging: delete the SBML document after we're done with it.
Fix memory leak when updating the level/version of an SBML model.
Fix for unix compiler, so we don't take address of temp variable.
Change how default units are stored so they don't leak like a sieve.
Fix module copy constructors so they copy the new 'm_displayname' member variable. Fix tests so they match.
revert test-breaking changes
Fix memory leak: delete SBML document after reading it.
Fix memory leak of ASTNodes.
Fix memory leak of converter.
Fix memory leaks of 'AllElements' lists.
Fix memory leak when we get an error string.
Fix memory leak when we get an error string.
sync api ver to 2.10.0
pass display name to module ctor
clean up
use comma separators for rdf resources
fix object parsing
use new-style annot parser to add cv terms
fix eof return
add first impl of composite annot parser, will clean up (21 errors in tests)
merge cv term work; all tests passing
add metaids to distrib test files
perform metaid-less comparisons in tests
add names to sbml tests
don't add built-in units to flattening
add name attr to fixname_test
elide metaids from comparison of flattened model
update expected syntax error
add git diff utility function
add elideMetaIds and elideMetaIdsFromSBMLstring SBML utility functions to allow testing for compliance sans metaids