* testsuite/read-error-stale-errno.sh: Skip when setting O_NONBLOCK on an fd in a child process has no effect on the fd in the parent process. Reported in https://lists.gnu.org/r/sed-devel/2026-04/msg00016.html
86 lines
2.9 KiB
Bash
Executable File
86 lines
2.9 KiB
Bash
Executable File
#!/bin/sh
|
|
# Verify ck_getdelim no longer reports a misleading strerror(errno)
|
|
# when ferror is already set from a prior I/O operation.
|
|
|
|
# Copyright (C) 2026 Free Software Foundation, Inc.
|
|
|
|
# This program is free software: you can redistribute it and/or modify
|
|
# it under the terms of the GNU General Public License as published by
|
|
# the Free Software Foundation, either version 3 of the License, or
|
|
# (at your option) any later version.
|
|
|
|
# This program is distributed in the hope that it will be useful,
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
# GNU General Public License for more details.
|
|
|
|
# You should have received a copy of the GNU General Public License
|
|
# along with this program. If not, see <https://www.gnu.org/licenses/>.
|
|
. "${srcdir=.}/testsuite/init.sh"; path_prepend_ ./sed
|
|
print_ver_ sed
|
|
|
|
# Verify that O_NONBLOCK is preserved across program invocations.
|
|
# In particular, verify that when a 'dd' child process sets O_NONBLOCK
|
|
# on stdin, this flag is then active in the parent process and in all
|
|
# its future child processes.
|
|
# This is not the case on Cygwin, because the major part of the
|
|
# O_NONBLOCK state is stored in an 'fhandler_base' object, not in the
|
|
# Windows HANDLE.
|
|
mkfifo probe || framework_failure_
|
|
{ exec sleep 1; } > probe &
|
|
probe_pid=$!
|
|
(
|
|
exec < probe
|
|
dd iflag=nonblock count=0 2>/dev/null
|
|
dd bs=1 count=1 2>/dev/null
|
|
)
|
|
probe_rc=$?
|
|
kill $probe_pid 2>/dev/null; wait
|
|
rm -f probe
|
|
test $probe_rc = 0 \
|
|
&& skip_ "O_NONBLOCK is not preserved across program invocations"
|
|
|
|
# Use a FIFO with non-blocking I/O. After sed reads one line,
|
|
# the next read(2) fails with EAGAIN (not EOF, since a writer is
|
|
# still attached), setting ferror on the stream.
|
|
#
|
|
# With sed -s, test_eof's getc triggers EAGAIN, setting ferror;
|
|
# ck_getdelim is then called with ferror already set.
|
|
#
|
|
# Before the fix: "sed: read error on stdin: Resource temporarily unavailable"
|
|
# After the fix: "sed: read error on stdin"
|
|
|
|
mkfifo fifo || framework_failure_
|
|
|
|
# A writer that keeps the FIFO's write-end open.
|
|
{ printf 'hello\n'; exec sleep 99; } > fifo &
|
|
writer_pid=$!
|
|
|
|
# "dd iflag=nonblock count=0" sets O_NONBLOCK on the file description
|
|
# underlying stdin (a no-op read that just changes the flag).
|
|
# That flag is then inherited by sed.
|
|
# Skip if dd doesn't support iflag=nonblock.
|
|
(
|
|
exec < fifo
|
|
dd iflag=nonblock count=0 2>/dev/null || exit 77
|
|
sed -s -n '$p'
|
|
) 2>err
|
|
rc=$?
|
|
|
|
kill $writer_pid 2>/dev/null
|
|
wait
|
|
|
|
test $rc -eq 77 && skip_ "dd does not support iflag=nonblock"
|
|
|
|
# sed should have panicked (exit 4) due to the read error.
|
|
test $rc -eq 4 || fail=1
|
|
|
|
# The error message must mention "read error" ...
|
|
grep 'read error' err > /dev/null || fail=1
|
|
|
|
# ... but must NOT include a ": strerror" suffix, which would be misleading
|
|
# (before the fix, this was e.g., ": Resource temporarily unavailable").
|
|
grep 'read error.*: ' err && fail=1
|
|
|
|
Exit $fail
|