> Am 20.06.2025 um 12:01 schrieb Peter Dyballa <peter_dyba...@web.de>:
>
> Comparing the function definition in both versions 3.12 and 3.10 I see that
> time is handled differently – cause found? In the end both functions simply
> have:
>
> set_color_context (RESET_CONTEXT);
> putc ('\n', outfile);
>
> The newer version has
>
> 24 int nsec = ts.tv_nsec;
>
> On Tiger /usr/include/time.h has:
>
> 87 #ifndef _TIMESPEC 88 #define _TIMESPEC
> 89 struct timespec {
> 90 time_t tv_sec; /* seconds */
> 91 long tv_nsec; /* and nanoseconds */
> 92 };
> 93 #endif
> 94
> 95 struct tm {
> 96 int tm_sec; /* seconds after the minute [0-60] */
> 97 int tm_min; /* minutes after the hour [0-59] */
> 98 int tm_hour; /* hours since midnight [0-23] */
> 99 int tm_mday; /* day of the month [1-31] */
> 100 int tm_mon; /* months since January [0-11] */
> 101 int tm_year; /* years since 1900 */
> 102 int tm_wday; /* days since Sunday [0-6] */
> 103 int tm_yday; /* days since January 1 [0-365] */
> 104 int tm_isdst; /* Daylight Savings Time flag */
> 105 long tm_gmtoff; /* offset from CUT in seconds */
> 106 char *tm_zone; /* timezone abbreviation */
> 107 };
In print_context_label()
61 ts = get_stat_mtime (&inf->stat);
is executed. get_stat_mtime() is defined in lib/stat-time.h as:
144 /* Return *ST's data modification time. */
145 _GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE
146 get_stat_mtime (struct stat const *st)
147 {
148 #ifdef STAT_TIMESPEC
149 return STAT_TIMESPEC (st, st_mtim);
150 #else
151 return (struct timespec) { .tv_sec = st->st_mtime,
152 .tv_nsec = get_stat_mtime_ns (st) };
153 #endif
154 }
GDB goes to line #149, so STAT_TIMESPEC must be TRUE. ts reports that I have
already spent 1¾·10⁹ seconds in front of my tv – plus zero nsec. Then in
print_context_label()
73 struct tm const *tm = localtime (&ts.tv_sec);
74 int nsec = ts.tv_nsec;
is executed. If tm_mon = 0 stands for January, then tm holds the correct date
(since 1900-01-01T00:00:00), nsec is zero. On nect line
75 if (tm && nstrftime (buf, sizeof buf, time_format, tm, localtz,
nsec))
nstrftime() is called, which is #define'd in
lib/nstrftime.c:18:#define my_strftime nstrftime
as:
1057 /* Write information from TP into S according to the format
1058 string FORMAT, writing no more that MAXSIZE characters
==> shouldn't it be "than"?
1059 (including the terminating '\0') and returning number of
1060 characters written. If S is NULL, nothing will be written
1061 anywhere, so to determine how many characters would be
1062 written, use NULL for S and (size_t) -1 for MAXSIZE. */
1063 size_t
1064 my_strftime (STREAM_OR_CHAR_T *s, STRFTIME_ARG (size_t maxsize)
1065 const CHAR_T *format,
1066 const struct tm *tp extra_args_spec LOCALE_PARAM)
1067 {
1068 bool tzset_called = false;
1069 return __strftime_internal (s, STRFTIME_ARG (maxsize) format, tp,
false,
1070 ZERO_PAD, -1,
1071 &tzset_called extra_args LOCALE_ARG);
1072 }
nstrftime is called with (s=0xbfffcb80 "\277\377\313\320", maxsize=52,
format=0x5be2c "%Y-%m-%d %H:%M:%S.%N %z", tp=0xa000fd84, tz=0x0, ns=0). The
value of localtz seems to be wrong, because it is never declared (tm contains
tm_gmtoff = 7200 (2 hrs) and tm_zone = "CEST", both correct).
••• And in the definition of my_strftime() it seems at least two COMMAs are
missing: •••
between "STRFTIME_ARG (size_t maxsize)" and "const CHAR_T *format",
between "const struct tm *tp" and "extra_args_spec LOCALE_PARAM",
and also a variable for nsec, separated by another COMMA? (lib/strftime.c has
not changed from 3.11 to 3.12; 3.10 does not have this file, all is in
lib/nstrftime.c while this file is in 3.11 and 3.12 almost empty.) diffutils
3.10 has the same function definition with missing COMMAs…
The call on line #1069 is:
__strftime_internal (s=0xbfffcb80 "\277\377\313\320", maxsize=52,
format=0x5be2c "%Y-%m-%d %H:%M:%S.%N %z", tp=0xa000fd84, upcase=false,
yr_spec=ZERO_PAD, width=-1, tzset_called=0xbfffcb08, tz=0x0, ns=0) at
strftime.c:1093
Near the end a breakpoint reveals that the form(at) has been filled in:
__strftime_internal (s=0xbfffcb80 "2025-06-17 08:43:05.000000000
+0200", maxsize=52, format=0x5be2c "%Y-%m-%d %H:%M:%S.%N %z", tp=0xa000fd84,
upcase=false, yr_spec=ZERO_PAD, width=-1, tzset_called=0xbfffcb08, tz=0x0,
ns=0) at strftime.c:2240
and so __strftime_internal() finishes with:
return i=35
nstrftime (s=0xbfffcb80 "2025-06-17 08:43:05.000000000 +0200",
maxsize=52, format=0x5be2c "%Y-%m-%d %H:%M:%S.%N %z", tp=0xa000fd84, tz=0x0,
ns=0) at strftime.c:1072
So both
99 print_context_label ("---", &inf[0], names[0], file_label[0]);
100 print_context_label ("+++", &inf[1], names[1], file_label[1]);
succeed, with obviously no failure when accessing the files' time stamps.
--
Greetings
Pete <]
o __o |__ o HPV, the real
___o /I -\<, |o \ -\),-% high speed!
___/\ /\___./ \___...O/ O____.....`-O-'-()--o_________________
