On Thu, 28 Nov 2024 11:33:43 GMT, Jonathan Lampérth <d...@openjdk.org> wrote:
> This PR includes a suggested change in behaviour of `javap -l` without `-v`
> or `-c`.
> Previously it was possible to print `LineNumberTable` and
> `LocalVariableTable` without disassembled code output. This didn't make much
> sense, as there is no context for this output. This PR proposes to output a
> warning for this case, as well as not print the `LineNumberTable`and
> `LocalVariableTable` attrubutes. They are nested attributes of the
> `CodeAttribute` and thus should only be printed if the `CodeAttribute` is
> also printed.
>
> Small changes were also made to existing tests, which seem unproblematic to
> adjust.
>
> <details>
> <summary>Before: `javac -l EmptyLoop`</summary>
>
>
> Compiled from "EmptyLoop.java"
> public class EmptyLoop {
> public EmptyLoop();
> Code:
> LineNumberTable:
> line 1: 0
>
> public void emptyLoop();
> Code:
> LineNumberTable:
> line 3: 0
> line 5: 14
> }
>
>
> </details>
>
> <details>
> <summary>After: `javac -l EmptyLoop`</summary>
>
>
> Warning: bad combination of options: -l without -c - line number and local
> variable tables will not be printed
> Compiled from "EmptyLoop.java"
> public class EmptyLoop {
> public EmptyLoop();
>
> public void emptyLoop();
> }
>
>
> </details>
>
> Stems from discussion in: https://github.com/openjdk/jdk/pull/22359
While I am not sure, looking at
[JDK-8322041](https://bugs.openjdk.org/browse/JDK-8322041)
and[JDK-8333827](https://bugs.openjdk.org/browse/JDK-8333827), it seems that
the translations are updated by certain people after RDP1 and then are
backported.
At least, it seems out of scope for this change.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/22434#issuecomment-2506458780
