aleofreddi / timel Goto Github PK

Describe the bug
Detect duplicate signatures as soon as they are registered to the function registry.

Actual behavior
FunctionRegistry registry = ...; registry.add(function); registry.add(function);
does not detect the duplicated.

Expected behavior
The code above should throw IllegalArgumentException on the second add.

Runtime
Any

Additional context
/

Describe the bug
TimEL is not able to parse hexadecimal constants, like 0x1.

Actual behavior

java.lang.NumberFormatException: For input string: "0x1"
	at java.base/java.lang.NumberFormatException.forInputString(Unknown Source)
	at java.base/java.lang.Integer.parseInt(Unknown Source)
	at java.base/java.lang.Integer.parseInt(Unknown Source)
	at net.vleo.timel.impl.parser.ParserTreeAdapter.visitTerminal(ParserTreeAdapter.java:255)
	at net.vleo.timel.impl.parser.ParserTreeAdapter.visitTerminal(ParserTreeAdapter.java:47)

Expected behavior
0x1 (as well 0X1) is interpreted correctly.

Runtime
Any

Additional context
/

Is your feature request related to a problem? Please describe.
Currently one can't escape characters in a string.

Describe the solution you'd like
The grammar and the parser should be evolved to support

fragment
EscapeSequence
    :   SimpleEscapeSequence // Standard C-like escapes
    |   OctalEscapeSequence
    |   HexadecimalEscapeSequence
    |   UniversalCharacterName

Describe alternatives you've considered
/

Additional context
/

Describe the bug
uniform and the current implementation of variables don't play well together.

The problem arises due to the temporally-unbound nature of uniform, because the result varies semantically (not only by resolution) on the evaluation time frame.

Actual behavior
For the interval 07/06/2019 10:09 PM → 07/07/2019 12:00 PM, evaluate the following:

a = extract("DAY_OF_WEEK", "Europe/Rome") == 7 ? uniform(1.0) : uniform(0); 

scale(
    a, 
    every(1, "DAY_OF_MONTH", "Europe/Rome")
);

The expected result would be 0.5, actual 0.75! That one is due to the uniform(1.0) being interpreted differently during the first evaluation and the second one.

Expected behavior
Such inconsistency should not happen, how it is still unclear.

Runtime
Any

Additional context
Maybe forcing uniform to be time-bounded will do? To be investigated.

Is your feature request related to a problem? Please describe.
TimEL should support true and false constants.

Describe the solution you'd like
True and false should be constants so that things like
when(a > 10, 10, b > 5, 5, 1==1, 0)
can be rewritten as
when(a > 10, 10, b > 5, 5, true, 0)

Describe alternatives you've considered
/

Additional context
/

Describe the bug
TimEL lacks unary operator- implementation

Actual behavior
Parsing "-1" leads to:
Parse error: net.vleo.timel.ParseException: Cannot resolve function -(Integer)

Expected behavior
"-1" is parsed correctly

Runtime
/

Additional context
/

Trunc and extract were handy functions available in version 0.2.

These allow better interaction with timestamp values.

Describe the bug
TimEL doesn't support unary operator+ so far.

Actual behavior
Evaluating +1 causes a Parse error at line 1:0: Failed to lookup function +

Expected behavior
Evaluating +1 returns 1.

Runtime
/
Additional context
/

Describe the bug
Compile version(1) or any other invalid function call that doesn't match a valid signature

Actual behavior

Aug 28 20:54:57 timel-demo java[23887]: java.lang.NullPointerException: null
Aug 28 20:54:57 timel-demo java[23887]: at net.vleo.timel.function.FunctionRegistry.lookup(FunctionRegistry.java:145) ~[timel-core-0.9.3-SNAPSHOT.jar!/:na]
Aug 28 20:54:57 timel-demo java[23887]: at net.vleo.timel.impl.intermediate.SyntaxTreeAdapter.visit(SyntaxTreeAdapter.java:97) ~[timel-core-0.9.3-SNAPSHOT.jar!/:na]
Aug 28 20:54:57 timel-demo java[23887]: at net.vleo.timel.impl.intermediate.SyntaxTreeAdapter.visit(SyntaxTreeAdapter.java:49) ~[timel-core-0.9.3-SNAPSHOT.jar!/:na]
Aug 28 20:54:57 timel-demo java[23887]: at net.vleo.timel.impl.parser.tree.FunctionCall.accept(FunctionCall.java:48) ~[timel-core-0.9.3-SNAPSHOT.jar!/:na]
Aug 28 20:54:57 timel-demo java[23887]: at net.vleo.timel.impl.intermediate.SyntaxTreeAdapter.lambda$visit$1(SyntaxTreeAdapter.java:119) ~[timel-core-0.9.3-SNAPSHOT.jar!/:na]
Aug 28 20:54:57 timel-demo java[23887]: at java.util.stream.ReferencePipeline$3$1.accept(ReferencePipeline.java:193) ~[na:1.8.0_151]
Aug 28 20:54:57 timel-demo java[23887]: at java.util.LinkedList$LLSpliterator.forEachRemaining(LinkedList.java:1235) ~[na:1.8.0_151]

Expected behavior
ParseException should be thrown instead, including the suggested signatures.

Runtime

• OS: Any
• JVM version: Any
• TimEL version: 0.9.3-SNAPSHOT (14e3ae9574c5d66e73f6e51554349219b7761c34)

Additional context
/

Describe the bug
TimEL's grammar supports unicode escape sequences as \uHHHH and \UHHHHHHHH to be used in identifiers, but these are not handled correctly.

Actual behavior
Running hello\U0001f602 = 1 defines a variable called hello\U0001f602

Expected behavior
Running hello\U0001f602 = 1 defines a variable called hello😂

Runtime
Any

Additional context
/

Is your feature request related to a problem? Please describe.
TimEL should support long constants via the L suffix

Describe the solution you'd like
Add IntegerSuffix to the grammar, so that maybe one day one can use it also for other data types (bigdecimals?):

fragment
IntegerConstant
    :   DecimalConstant IntegerSuffix?
    |   OctalConstant IntegerSuffix?
    |   HexadecimalConstant IntegerSuffix?
    ;

fragment
IntegerSuffix
    :   LongSuffix
    ;

fragment
LongSuffix
    :   [lL]
    ;

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
Add any other context or screenshots about the feature request here.

Currently the last statement should not be semicolon terminated, which is extremely annoying for the users.

Change the behavior so that the last expression can be compiled with a closing semicolon without affecting the semantic

Evaluate the "Pay-per-use scenario" on the web console with a stoptime as 03/05/2020 12:00 AM, leads to 500:

Value 2 for hourOfDay is not supported: Illegal instant due to time zone offset transition (daylight savings time 'gap'): 2019-03-31T02:00:00.000 (Europe/Rome)

To be investigated!

Describe the bug
Many Functions throw RuntimeException when an evaluation exception happens. These should be migrated to a specific checked exception EvaluationException.

Actual behavior
Evaluation of every(1, "x", "x") throws java.lang.RuntimeException: Invalid field specified: x

Expected behavior
Evaluation of every(1, "x", "x") throws EvaluationException.

Runtime
Any

Additional context
/

Is your feature request related to a problem? Please describe.
For the sake of compability with the rest of the world, TimEL should be able to understand octal constants.

Describe the solution you'd like
Add octal support:

fragment
IntegerConstant
    :   OctalConstant

fragment
OctalConstant
    :   '0' OctalDigit*
    ;

Describe alternatives you've considered
/

Additional context
/

Describe the bug
Compile (Integer<1>) 1

Actual behavior
Throws a ParseException wrapped in a RuntimeException.

Expected behavior
Throws a ParseException.

Runtime

• TimEL 0.9.3-SNAPSHOT (14e3ae9574c5d66e73f6e51554349219b7761c34)

Additional context
/