Hi Harald,

On 2020-07-01 16:38, Harald Seiler wrote:
> Use a ProgressHandler to report how far postprocessing has progressed.
> Because we can't know the exact timing of all postprocessing tasks, the
> percentages are split evenly for all steps.

I think I wanted to do this myself at some point. So great that I don't
have to ;)

I think I would probably have used the MultiStageProgressReporter here
as well, like with it is done in `do_rootfs_install` and just entered
some measured values for the tasks I got locally as weights. The idea is
that the weights are just a rough indicator. Since `do_rootfs_install`
has some tasks that take much longer than others, it sort of makes
sense. If the postprocessing tasks are mostly equal, then the
ProgressHandler is ok.

Maybe you can just try measuring it with the MultiStageProgressReporter
and then decide.

IIRC, I though at that time that the post-process tasks are finished
pretty fast and so decided not to invest the time to add a progress bar.
But its long ago, so I might be mistaken.

regards,
Claudius

> 
> Signed-off-by: Harald Seiler <hws@denx.de>
> ---
>  meta/classes/rootfs.bbclass | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/meta/classes/rootfs.bbclass b/meta/classes/rootfs.bbclass
> index 806e824c825a..889272e7bee4 100644
> --- a/meta/classes/rootfs.bbclass
> +++ b/meta/classes/rootfs.bbclass
> @@ -248,12 +248,16 @@ python do_rootfs_postprocess() {
>      # removed on a previous execution of this task:
>      bb.build.exec_func('rootfs_do_qemu', d)
>  
> +    progress_reporter = bb.progress.ProgressHandler(d)
> +    progress_reporter.update(0)
> +
>      cmds = d.getVar("ROOTFS_POSTPROCESS_COMMAND")
>      if cmds is None or not cmds.strip():
>          return
>      cmds = cmds.split()
> -    for cmd in cmds:
> +    for i, cmd in enumerate(cmds):
>          bb.build.exec_func(cmd, d)
> +        progress_reporter.update(int(i / len(cmds) * 100))
>  }
>  addtask rootfs_postprocess before do_rootfs
>  
>