Your message dated Thu, 8 Sep 2022 23:48:55 +0200 with message-id <[email protected]> and subject line Re: libc6: mktime() does not set errno when it fails has caused the Debian Bug report #874798, regarding libc6: mktime() does not set errno when it fails 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 [email protected] immediately.) -- 874798: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=874798 Debian Bug Tracking System Contact [email protected] with problems
--- Begin Message ---Package: libc6 Version: 2.24-11+deb9u1 Severity: normal Tags: upstream When mktime() fails to convert a struct tm to a time_t, it returns -1. It should also set errno to EOVERFLOW in order to distinguish the failure from the legitimate case of converting "1 second before the epoch". The following program, when run on a 32-bit system, shows that mktime does not comply to the standard. On recent amd64 systems, time_t is 64 bit long, therefore failure doesn't occur for the same broken down times. Best regards, g.b. #define _POSIX_C_SOURCE 200112L #include <stdio.h> #include <time.h> #include <assert.h> #include <errno.h> #include <stdlib.h> int main() { struct tm tm0; struct tm *ptm; time_t t; int eno; assert(sizeof t == 4); /* need 32 bit to trigger failure in 1900 */ /* nail timezone to UTC */ setenv("TZ", "UTC", 1); tzset(); t = -1; /* 1 sec before epoch */ ptm = gmtime_r(&t, &tm0); assert(ptm != NULL); errno = 0; t = mktime(&tm0); eno = errno; printf("struct TM: %d-%d-%d %02d:%02d:%02d (dst=%d)\n", 1900 + tm0.tm_year, 1 + tm0.tm_mon, tm0.tm_mday, tm0.tm_hour, tm0.tm_min, tm0.tm_sec, tm0.tm_isdst); printf("mktime(): retval %ld (expected -1), " "errno %d (expected 0)\n", (long) t, eno); tm0.tm_year = 0; /* warp to year 1900 */ /* bad conversion for 32-bit time_t, should return -1 and set errno */ errno = 0; t = mktime(&tm0); eno = errno; printf("struct TM: %d-%d-%d %02d:%02d:%02d (dst=%d)\n", 1900 + tm0.tm_year, 1 + tm0.tm_mon, tm0.tm_mday, tm0.tm_hour, tm0.tm_min, tm0.tm_sec, tm0.tm_isdst); printf("mktime(): retval %ld (expected -1), " "errno %d (expected EOVERFLOW=%d)\n", (long) t, eno, EOVERFLOW); return 0; } -- System Information: Debian Release: 9.1 APT prefers stable APT policy: (500, 'stable') Architecture: i386 (i686) Kernel: Linux 4.9.0-3-686 (SMP w/1 CPU core) 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: sysvinit (via /sbin/init) Versions of packages libc6 depends on: ii libgcc1 1:6.3.0-18 libc6 recommends no packages. Versions of packages libc6 suggests: ii debconf [debconf-2.0] 1.5.61 pn glibc-doc <none> ii libc-l10n 2.24-11+deb9u1 ii locales 2.24-11+deb9u1 -- debconf information excluded
--- End Message ---
--- Begin Message ---Version: 2.29-1 Hi, On 2017-09-09 19:29, g1 wrote: > Package: libc6 > Version: 2.24-11+deb9u1 > Severity: normal > Tags: upstream > > When mktime() fails to convert a struct tm to a time_t, it returns -1. > It should also set errno to EOVERFLOW in order to distinguish the failure > from the legitimate case of converting "1 second before the epoch". > > The following program, when run on a 32-bit system, shows that mktime > does not comply to the standard. On recent amd64 systems, time_t is 64 > bit long, therefore failure doesn't occur for the same broken down times. > This bug has been fixed in glibc 2.29-1. Closing the bug accordingly. Regards, Aurelien -- Aurelien Jarno GPG: 4096R/1DDD8C9B [email protected] http://www.aurel32.net
--- End Message ---
