Your message dated Sun, 16 Dec 2018 01:21:35 +0000
with message-id
<RAT8kpcHnxxXi22HB5CtkjaVUMolX_4khVefXryQFG7iImccc3I-2RVHZvDuMKkqpwAHKkoxUEaaZdzPT9lSMjnLq2daAx8O2N6Yvm1zogI=@protonmail.ch>
and subject line Re: Bug#916562: libc6: fputs(), fputc(), and fwrite() do not
fail on write error
has caused the Debian Bug report #916562,
regarding libc6: fputs(), fputc(), and fwrite() do not fail on write error
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact ow...@bugs.debian.org
immediately.)
--
916562: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=916562
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems
--- Begin Message ---
Package: libc6
Version: 2.27-8
Severity: normal
Tags: upstream
Dear Maintainer,
If I use fputs(3), fputc(3), or fwrite(3) to write to a file that can
be opened for writing but cannot be written to (e.g /dev/full), the
functions return 1 rather than the expected EOF (or 0 in the case of
fread()). On the other hand, write(2) returns -1 and sets errno to
ENOSPC an expected.
Expected results:
The functions fputs() and fputc() should return EOF and fwrite()
should return 0. All the functions should set errno to ENOSPC (for
/dev/full).
Actual results:
The functions return 1 (an indication of success) and leave errno
unchanged.
Steps to reproduce:
Compile the attached program (the compiler flag `-fno-builtin' does
not change anything, nor does -O0):
$ gcc -o write write.c
Run the program with /dev/full as the output file:
$ ./write hello /dev/full
fputs() returned 1 and errno is 0 (Success).
fputc() returned 104 (h) and errno is 0 (Success).
fwrite() returned 5 (the length of "hello" is 5) and errno is 0 (Success).
write() returned -1 (the length of "hello" is 5) and errno is 28 (No space left
on device).
The expected output is:
fputs() returned EOF (-1) and errno is 28 (No space left on device).
fputc() returned EOF (-1) and errno is 28 (No space left on device).
fwrite() returned 0 (the length of "hello" is 5) and errno is 28 (No space left
on device).
write() returned -1 (the length of "hello" is 5) and errno is 28 (No space left
on device).
-- System Information:
Debian Release: buster/sid
APT prefers testing
APT policy: (500, 'testing')
Architecture: amd64 (x86_64)
Kernel: Linux 4.18.0-2-amd64 (SMP w/2 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8),
LANGUAGE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
Versions of packages libc6 depends on:
ii libgcc1 1:8.2.0-9
libc6 recommends no packages.
Versions of packages libc6 suggests:
ii debconf [debconf-2.0] 1.5.69
pn glibc-doc <none>
ii libc-l10n 2.27-8
ii locales 2.27-8
-- debconf information:
* glibc/restart-services: cups cron
glibc/kernel-too-old:
glibc/kernel-not-supported:
glibc/restart-failed:
glibc/upgrade: true
glibc/disable-screensaver:
* libraries/restart-without-asking: false
#include <stdio.h>
#include <unistd.h>
#include <string.h>
#include <errno.h>
int main(int argc, char **argv) {
int ret;
size_t string_len;
FILE *output_file;
if (argc != 3) {
fprintf(stderr, "Usage: %s STRING FILE\nWrite STRING to FILE.\n", argv[0]);
return 1;
}
/* Open the file */
if (!(output_file = fopen(argv[2], "w"))) {
fprintf(stderr, "%s: could not open %s: %m\n", argv[0], argv[2]);
return 1;
}
/* Now try to write to the file using fputs() */
errno = 0;
ret = fputs(argv[1], output_file);
/* `ret' should be EOF if an error occured (see fputs(3)) */
printf ("fputs() returned ");
if (ret == EOF)
printf("EOF (%d)", ret);
else
printf("%d", ret);
printf(" and errno is %d (%m).\n", errno);
/* This happens with fputc() too */
errno = 0;
ret = fputc(argv[1][0], output_file);
printf("fputc() returned ");
if (ret == EOF)
printf("EOF (%d)", ret);
else
printf("%d (%c)", ret, (char)ret);
printf(" and errno is %d (%m).\n", errno);
string_len = strlen(argv[1]);
/* This happens with fwrite() too */
errno = 0;
printf("fwrite() returned %lu (the length of \"%s\" is %lu) ",
fwrite(argv[1], 1, string_len, output_file), argv[1], string_len);
printf("and errno is %d (%m).\n", errno);
/* write() gives us expected results */
errno = 0;
printf("write() returned %ld (the length of \"%s\" is %lu) ",
write(fileno(output_file), argv[1], string_len), argv[1], string_len);
printf("and errno is %d (%m).\n", errno);
return 0;
}
--- End Message ---
--- Begin Message ---
On Saturday, December 15, 2018 7:16 PM, Asher Gordon <asd...@protonmail.ch>
wrote:
> On Saturday, December 15, 2018 7:02 PM, Samuel Thibault sthiba...@debian.org
> wrote:
>
> > Hello,
> > Asher Gordon, le sam. 15 déc. 2018 18:51:19 -0500, a ecrit:
> >
> > > If I use fputs(3), fputc(3), or fwrite(3) to write to a file that can
> > > be opened for writing but cannot be written to (e.g /dev/full), the
> > > functions return 1 rather than the expected EOF (or 0 in the case of
> > > fread()).
> >
> > Well, that is not surprising since these functions are buffered. You
> > need to call fflush() to make sure that no error happened on the actual
> > underlying write.
> > Samuel
>
> You're right. I tried it with fflush() and it gave expected results. I guess
> we can close this bug now.
I'm closing the bug now.
--- End Message ---
