On Wed, Dec 10, 2025 at 10:58:45AM +0100, Heinrich Schuchardt wrote: > On 12/10/25 00:07, Tom Rini wrote: > > On Wed, Dec 10, 2025 at 12:03:18AM +0100, Heinrich Schuchardt wrote: > > > Add missing boardenv files for > > > > > > * qemu-riscv64_smode_defconfig > > > * qemu-riscv64_smode_acpi_defconfig > > > > > > to enable network testing. > > > > > > Signed-off-by: Heinrich Schuchardt <[email protected]> > > > --- > > > .../u_boot_boardenv_qemu_riscv64_smode_acpi_na.py | 10 ++++++++++ > > > py/travis-ci/u_boot_boardenv_qemu_riscv64_smode_na.py | 10 ++++++++++ > > > 2 files changed, 20 insertions(+) > > > create mode 100644 > > > py/travis-ci/u_boot_boardenv_qemu_riscv64_smode_acpi_na.py > > > create mode 100644 py/travis-ci/u_boot_boardenv_qemu_riscv64_smode_na.py > > > > These are identical yes? So they can be symlinks. Probably to > > u_boot_boardenv_qemu_riscv64_na.py ? I forget if the file > > "py/travis-ci/u_boot_boardenv_qemu_riscv64.py" would be read by all > > iterations of qemu_riscv64. > > > > Hello Tom, > > My understanding is that the files are per boards. I am always irritated by > the _na suffixes. Maybe we can get rid of these. Or at least we should > document these.
Yes. I recall seeing in the code a month or two back that we have some fall-back logic so that for the conf files, we can omit the _na suffix. I don't recall if the python import portion has similar logic however. > I remember you being very hesitant about symlinks in git repos because they > are not supported on all file-systems to which a user might check out at git > repo (e.g. on Windows). Ah, I had forgotten that myself, but, we have several symlinks in the main repository itself now, so I think it's become a moot point. Especially for trade-offs in reducing duplication. -- Tom
signature.asc
Description: PGP signature
