2 # Detect printf(3) failure even when it doesn't set stream error indicator
4 # Copyright (C) 2007-2009 Free Software Foundation, Inc.
6 # This program is free software: you can redistribute it and/or modify
7 # it under the terms of the GNU General Public License as published by
8 # the Free Software Foundation, either version 3 of the License, or
9 # (at your option) any later version.
11 # This program is distributed in the hope that it will be useful,
12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 # GNU General Public License for more details.
16 # You should have received a copy of the GNU General Public License
17 # along with this program. If not, see <http://www.gnu.org/licenses/>.
21 if test "$VERBOSE" = yes; then
31 # Up to coreutils-6.9, "printf %.Nf 0" would encounter an ENOMEM internal
32 # error from glibc's printf(3) function whenever N was large relative to
33 # the size of available memory. As of Oct 2007, that internal stream-
34 # related failure was not reflected (for any libc I know of) in the usual
35 # stream error indicator that is tested by ferror. The result was that
36 # while the printf command obviously failed (generated no output),
37 # it mistakenly exited successfully (exit status of 0).
39 # Testing it is tricky, because there is so much variance
40 # in quality for this corner of printf(3) implementations.
41 # Most implementations do attempt to allocate N bytes of storage.
42 # Using the maximum value for N (2^31-1) causes glibc-2.7 to try to
43 # allocate almost 2^64 bytes, while freeBSD 6.1's implementation
44 # correctly outputs almost 2GB worth of 0's, which takes too long.
45 # We want to test implementations that allocate N bytes, but without
46 # triggering the above extremes.
48 # Some other versions of glibc-2.7 have a snprintf function that segfaults
49 # when an internal (technically unnecessary!) memory allocation fails.
51 # The compromise is to limit virtual memory to something reasonable,
52 # and to make an N-byte-allocating-printf require more than that, thus
53 # triggering the printf(3) misbehavior -- which, btw, is required by ISO C99.
55 mkfifo fifo || framework_failure
57 # Disable MALLOC_PERTURB_, to avoid triggering this bug
58 # http://bugs.debian.org/481543#77
59 export MALLOC_PERTURB_=0
61 head -c 10 fifo > out &
62 ( ulimit -v 10000; env $prog %20000000f 0 2>err-msg > fifo )
65 # Map this longer, and rarer, diagnostic to the common one.
66 # printf: cannot perform formatted output: Cannot allocate memory" \
67 sed 's/cannot perform .*/write error/' err-msg > k && mv k err-msg
68 err_msg=$(cat err-msg|tr '\n' :)
70 # By some bug, on Solaris 11 (5.11 snv_86), err_msg ends up
71 # containing '1> fifo:printf: write error:'. Recognize that, too.
74 "$prog: write error:") diagnostic=y ;;
75 "1> fifo:$prog: write error:") diagnostic=y ;;
77 *) diagnostic=unexpected ;;
81 case $n_out:$diagnostic:$exit in
82 10:n:0) ;; # ok, succeeds w/no diagnostic: FreeBSD 6.1
83 0:y:1) ;; # ok, glibc-2.8 and newer, when printf(3) fails with ENOMEM
85 # With MALLOC_PERTURB_=0, this no longer happens.
86 # *:139) # segfault; known bug at least in debian unstable's libc6 2.7-11
87 # echo 1>&2 "$0: bug in snprintf causes low-mem use of printf to segfault"
90 # 10:y) ;; # Fail: doesn't happen: nobody succeeds with a diagnostic
91 # 0:n) ;; # Fail pre-patch: no output, no diag