Bugs in the generated sample code

[tsdh]

In the generated Java code, the Order member TransactTm ends in a lower-case "m", but in the constructor you access it as TransactTM ending in a capital "M".

In C# and C++, a class Foo must not have a member that's also called Foo. Thus this C# line and this C++ line (and probably more) will give compile errors.

The C++ code won't compile because of multiple reasons. And defining everything including constructors in a header is a no-no anyway.

So is the goal generating code that actually compiles and works, or is being syntactically similar to Java/C#/C++ enough?

[kevinlano]

The transformation should generate correct code. The C++ code

for test1 does compile correctly under Visual Studio, apart from the

Tm/TM problem.

Kevin

---

From: Tassilo Horn notifications@github.com
Sent: 26 March 2014 10:01
To: TransformationToolContest/ttc2014-fixml
Cc: Lano, Kevin
Subject: [ttc2014-fixml] Bugs in the generated sample code (#6)

In the generated Java code, the Order member TransactTmhttps://github.com/TransformationToolContest/ttc2014-fixml/blob/master/test_cases/test1java.txt#L4 ends in a lower-case "m", but in the constructorhttps://github.com/TransformationToolContest/ttc2014-fixml/blob/master/test_cases/test1java.txt#L17 you access it as TransactTM ending in a capital "M".

In C#, a class Foo must not have a member that's also called Foo. Thus this linehttps://github.com/TransformationToolContest/ttc2014-fixml/blob/master/test_cases/test2csharp.txt#L163 (and probably more) will give compile errors.

The C++ code won't compile because of multiple reasons. And defining everything including constructors in a header is a no-no anyway.

So is the goal generating code that actually compiles and works, or is being syntactically similar to Java/C#/C++ enough?

Reply to this email directly or view it on GitHubhttps://github.com//issues/6.

[tsdh]

@kevinlano Yeah, test2 (not test1) doesn't compile with C# and C++ because of members named like their enclosing type.