Hello,
Flavio Cruz, le mar. 05 mars 2024 02:13:57 -0500, a ecrit:
> The perl test suite has a test where it reads all the block or char devices
> under /dev without following the translators. Then it compares it against a
> list of devices that read the translated nodes stat info.
>
> The patch changes how the the device files are created initially so that the
> stat information
> is identical and makes the Hurd environment appear more similar to other
> operating
> systems.
Oh, good, thanks!
> ---
> sutils/MAKEDEV.sh | 67 +++++++++++++++++++++++++++++++----------------
> 1 file changed, 44 insertions(+), 23 deletions(-)
>
> diff --git a/sutils/MAKEDEV.sh b/sutils/MAKEDEV.sh
> index 0702663..79757fc 100644
> --- a/sutils/MAKEDEV.sh
> +++ b/sutils/MAKEDEV.sh
> @@ -81,6 +81,27 @@ st() {
> if [ "$KEEP" ] && showtrans "$NODE" > /dev/null 2>&1 ; then
> return;
> fi
> + local NODE_TYPE="$1"
> + case "$NODE_TYPE" in
> + b|c|d|f)
> + shift 1
> + ;;
> + *)
> + NODE_TYPE="f"
> + ;;
I'd rather not have differing behavior: better always eat the node
type, and shout out if it is unknown.
> + esac
> + if [ ! -e "$NODE" ]; then
> + case "$NODE_TYPE" in
> + b|c)
> + cmd mknod "$NODE" "$NODE_TYPE" 0 0
> + ;;
> + d)
> + cmd mkdir "$NODE"
> + ;;
> + *)
> + ;;
Here as well, better not have a catch-all case.
> @@ -235,7 +256,7 @@ mkdev() {
> # Linux, we tell tmpfs to set the size to half the physical RAM
> # in the machine.
> shm)
> - st $I root 644 /hurd/tmpfs --mode=1777 50%
> + st $I root 644 d /hurd/tmpfs --mode=1777 50%
> ;;
>
> pseudo-root)
pseudo-root is a block device too.
> @@ -246,7 +267,7 @@ mkdev() {
> loop*)
> # In Linux an inactive "/dev/loopN" device acts like /dev/null.
> # The `losetup' script changes the translator to "activate" the device.
> - st $I root 640 /hurd/null
> + st $I root 640 c /hurd/null
> ;;
>
> *)
Samuel
