>I suspect the proper solution is to treat the -e output as shell-like
>From the bitbake/lib/bb/data.py:
"""Emits all items in the data store in a format such that it can be
sourced by a shell."""
so if you say it's shell-like, not shell, shouldn't we replace that to
"""Emits all items in the data
Well, as you've discovered, bitbake variables can have names that are
invalid in shell. I suspect the proper solution is to treat the -e
output as shell-like, and parse it with a proper parser.
Ross
On Thu, 18 Oct 2018 at 09:56, Tomasz Dziendzielski
wrote:
>
> So what is the proper solution?
>
So what is the proper solution?
Best regards,
Tomasz Dziendzielski
wt., 16 paź 2018 o 16:00 Tomasz Dziendzielski <
tomasz.dziendziel...@gmail.com> napisał(a):
> Yes, I want to source the data files. I'm trying to adapt devtool the way
> it's creating the workspace, then I'm sourcing a script
Yes, I want to source the data files. I'm trying to adapt devtool the way
it's creating the workspace, then I'm sourcing a script with all the
necessary variables and then interactively run commands by hand.
I was thinking about the devshell but in devshell variables like ${S}, ${B}
are not set
Do you actually want to source the data files, and if so why? As with
the bitbake -e output, it's shell-like but not actually shell.
Ross
On Tue, 16 Oct 2018 at 13:44, Tomasz Dziendzielski
wrote:
>
> Hi,
> when using archiver.bbclass the do_dumpdata task is using bb.data.emit_env
> function,
Hi,
when using archiver.bbclass the do_dumpdata task is using bb.data.emit_env
function, which should emit variables "in a format such that it can be
sourced by a shell.".
The problem is that the *-showdata.dump file can't be properly sourced,
because variables are not in a correct format.
There
