RFR: 8330846: Add stacks of mounted virtual threads to the HotSpot thread dump [v7]

Inigo Mediavilla Saiz duke at openjdk.org
Tue Jun 4 13:27:46 UTC 2024


On Tue, 4 Jun 2024 08:50:25 GMT, Thomas Stuefe <stuefe at openjdk.org> wrote:

>> Actually I'm not sure what this indentation actually does at this location and its affect on other user's of this API. I would have expected the caller to set up the necessary indentation in the stream that gets passed in. I see you inc the indentation but then use the current indentation to insert multiple tabs - which should not be necessary if the stream indentation works correctly. ???
>
> Note that We are in the process of adding better and saner auto-indentation to outputStream. See https://github.com/openjdk/jdk/pull/19461 . I don't think that PR is going to take long.
> 
> If you don't want to wait, please:
> - As David wrote, use spaces, not tabs
> - Today's pattern for using outputStream indentation is:
>   - set up indentation, preferably with streamIndentor, not manually with inc/dec
>   - then, before printing each line, call stream->indent()
>  
> This pattern would also help us to later identify and remove this manual indentation pattern if auto-indent becomes a thing.
> 
> But really, waiting for https://github.com/openjdk/jdk/pull/19461 would be preferable. Then, all you have to do is place a streamIndentor around stack printing. Sub-function printing is then indented automatically.

Thanks !

Your PR looks very promising @tstuefe, I would indeed prefer to wait for your changes as a way to add additional indentation to the stack of the virtual thread.

What do you think if I leave the current PR with the indentation that is already used for the stack of the carrier thread and I create a separate PR based on yours to add the additional indentation ?

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/19482#discussion_r1626005247


More information about the hotspot-runtime-dev mailing list