Commits

Author Commit Message Labels Comments Date
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-55. ExpressionCompiler.shouldCast(Node) was returning true for ASTConst ...Obviousoly not correct.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-54. OgnlRuntime method finders and getChildSource weren't properly handling properties that should only be referencing root expressions (if any)..Such as method parameters / etc..Added logic to correctly set the context type/object state up for each of these.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
-) Fix for OGNL-52. Added a synchronized() block around the method object being invoked to prevent failures with the accessible flag being turned on/off temporarily. -) Added new unit test but failed to verify OGNL-51.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-50. Boolean literals within ASTAnd expressions were failing compilation (stupid ass javassist) unless surrounded with ( ).
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
ASTTest had been quietly failing since the last set of changes. Not good for performance...
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Added new getRootExpressionClass() to compiler interface so that other compiler impls can handle special corner cases (like operating on previously javassist'ed classes)
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Re-factored OgnlContext type / accessor logic to use backing List objects so that the full type chain can be evaluated after the fact.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-46. ExpressionCompiler.getRootExpression() was incorrectly casting directly to the root type instead of checking the context accessor type first. (to be friendly with object inheritance)
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes regression of string concatenations in OGNL-48.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Code cleanup , new test for bug in javassist.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-44. ASTAdd wasn't checking for property/var references before expanding value via native conversion methods.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Resolves OGNL-43. ASTAnd was incorrectly casting to a boolean expression when language grammar states && expressions return value of second expression - not just boolean.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Resolves OGNL-40. ASTAdd was improperly double escaping quoting characters.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Resolves OGNL-39. ASTTest wasn't handling the possibility of different object types being on the return stack of a ternary statement.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Resolves OGNL-41. Needed to add "(" around !org.ognl.StaticClass.method()" block to make javassist stfu.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-37. OgnlRuntime.getMethod() wasn't properly checking types using the new context.getCurrentType(). (need to kill off that old NodeType interface)
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-35. Boolean "&&" expressions were being evaluated during source generation regardless of whether either condition was true. Added check that throws UnsupportedCompilationException from ASTAnd until all conditions are true.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Let's just say it fixes most problems and leave it at that...
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Resolves OGNL-24. ListPropertyAccessor wasn't delegating to superclass for getSource in operations where a list object has been extended ..
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Resolves OGNL-23. String concatenation not working properly in ternary statements.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Resolves OGNL-22. Null strings weren't being handled properly in mathematical expressions.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Upgraded javacc and re-ran grammar generator.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes number of bugs resulting from more ASTConst changes and how the context types are figured out.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
OgnlRuntime was incorrectly returning false for shouldConvertNumericTypes when the values were primitives - probably a result of the astconst changes.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Minor update?
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-14 && OGNL-16. OgnlRuntime.getChildSource was still using the old non "($w)" widening casting semantics provided by javassist.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-18. Array/List/etc accessors weren't properly detecting object vs primitive types when accessing their respective properties because ASTConst was still sticking the old object class type as the current context type. This involved refactoring ASTConst to correctly place the native type for numerics as well as fixing all the problems resulting from this new behavior.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Fixes OGNL-17. ...Too much to list, the summary is that ASTCtor / ASTList hadn't had the same kind of context/ refactoring love that ASTMethod/ property/ etc had ....
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Partially fixes OGNL-17 . It's only partial because someone tried to sneak in two different tickets under one.
jkuh...@d4b077e3-5828-0410-b394-cb2b42183085
Added check in OgnlOps to throw an IllegalArgumentException if a value can't be converted to a requested type. Prevents setter expressions from silently failing when the wrong object type is used.
  1. Prev
  2. Next