On 18/04/2026 05:54, Collin Funk wrote:
Bruno Haible via GNU coreutils General Discussion <[email protected]>
writes:
On NetBSD 10, there is one test failure (already reported for the
coreutils 9.10 prerelease):
FAIL: tests/tail/pipe-f2
This one is interesting. I haven't looked at a fix yet, but here is the
issue.
When standard output is not redirected 'tail' behaves as we expect:
$ mkfifo fifo
$ echo 1 > fifo &
[1] 23114
$ ./src/timeout -v 10 ./src/tail -s.1 \
--max-unchanged-stats=1 -f fifo
1
timeout: sending signal TERM to command './src/tail'
[1]+ Done echo 1 > fifo
However, when we redirect standard output nothing is printed:
$ echo 1 > fifo &
[1] 10274
$ ./src/timeout -v 10 ./src/tail -s.1 \
--max-unchanged-stats=1 -f fifo > out
timeout: sending signal TERM to command './src/tail'
[1]+ Done echo 1 > fifo
$ ./src/od out
0000000
Collin
Interesting. Perhaps different flushing behaviour between
tail_forever() and tail_inotify_forever()?
I'd put some debug to stderr at each fflush() point.
cheers,
Padraig
