On Fri, Feb 14, 2025 at 05:19:48 -0600, MacBeth wrote:
> On Fri, Feb 14, 2025 at 4:31 AM MacBeth <macbeth.112...@gmail.com> wrote:
> > # fails:
> > $ history -r <(echo 'echo bad')
>
> Unless the reason is due to the file seek issue as described here...?
> https://unix.stackexchange.com/a/164109
hobbit:~$ echo ': testing file y' > y
hobbit:~$ strace bash -c 'history -r y'
[...]
openat(AT_FDCWD, "y", O_RDONLY) = 3
newfstatat(3, "", {st_mode=S_IFREG|0644, st_size=17, ...}, AT_EMPTY_PATH) = 0
read(3, ": testing file y\n", 17) = 17
close(3) = 0
[...]
hobbit:~$ strace bash -c 'history -r <(echo ": testing procsub")'
[...]
openat(AT_FDCWD, "/dev/fd/63", O_RDONLY) = 3
newfstatat(3, "", {st_mode=S_IFIFO|0600, st_size=0, ...}, AT_EMPTY_PATH) = 0
close(3) = 0
close(63) = 0
[...]
Looks like it never even tries to read the bytes from the procsub. It
checks the file's type, sees S_IFIFO, and silently ignores it.
You'd think it would at least give a warning.
