JIRA JASSIST-227: include INCREMENT/DECREMENT on case IINC

[caot]

No description provided.

[chibash]

Why is the patch for ClassPool necessary? Could you explain?

[caot]

int i = 0;
i1 += 5;
i1 -= 5;

the related bytecode:
Code:
0: iconst_0
1: istore_3
2: iinc 2, 5
5: iinc 2, -5

Obviously, the javassist ignjored the increament/decreament (5/-5), and
generated the following:

Code:
0: iconst_0
1: istore_3
2: iinc 2
5: iinc 2

On Mon, Jul 21, 2014 at 10:35 PM, Shigeru Chiba notifications@github.com
wrote:

Why is the patch for ClassPool necessary? Could you explain?

—
Reply to this email directly or view it on GitHub
#13 (comment)
.

[chibash]

That part is fine. I'm wondering about the necessity of the patch "classes of java.lang".

[caot]

Otherwise, any classname under package java.lang will throw Exception in
the call of "get0(classname, true)" for it is not in classpool.

On Tue, Jul 22, 2014 at 9:24 PM, Shigeru Chiba notifications@github.com
wrote:

That part is fine. I'm wondering about the necessity of the patch "classes
of java.lang".

—
Reply to this email directly or view it on GitHub
#13 (comment)
.

[chibash]

Really? I could not reproduce that bug. Maybe because ClassPath is not appropriately given.

[chibash]

I've merged the first patch (d267e47).
Thank you!

[caot]

To reproduce the bug through the return type String or other class under
package java.lang of a method.

On Wed, Jul 23, 2014 at 6:16 AM, Shigeru Chiba notifications@github.com
wrote:

I've merged the first patch (d267e47
d267e47).
Thank you!

—
Reply to this email directly or view it on GitHub
#13 (comment)
.

[chibash]

I did but I couldn't see the problem. Can you show me a program for reproduction?