May 1993 - tz - lists.iana.org

Revised draft strftime.c
by ado 30 May '93

30 May '93

An updated draft is available on elsie.nci.nih.gov as ~ftp/pub/draft-strftime.tar.Z, which includes strftime.c and the header files it references. The draft version incorporates the shift to use of sprintf to generate output. This makes the code simpler; it may irk folks who get the standard input/output library dragged in when they call strftime. The handling of "%D" and "%x" have been separated so that they can be changed independently. Right now they're still treated identically in the code; comments have been added giving other possibilities for "%x" handling. The worry is that some piece of code out there is counting on the current "%x" behavior and will break if it is changed. (Yes, counting on "%x" behavior is worng.) As always, any comments are welcome. --ado

1 0

more on ISO date format
by eggert＠spot.twinsun.com 30 May '93

30 May '93

I did some more research on ISO date format, and discovered that the ISO 2014 standard has been superseded by ISO 8601:1988. The date format hasn't changed, however. You can find a summary of ISO 8601:1988 in ftp.uni-erlangen.de:pub/doc/ISO/english/ISO8601.ps.Z.

1 0

Re: proposed changes to strftime.c
by ado 29 May '93

29 May '93

> Use ISO standard date format for %x; this is more appropriate than the > American date format, which confuses non-Americans. Especially for > dates like 01/02/03; the ISO date 2003-01-02 is much clearer for > everybody, and it sorts correctly too. Does "ISO standard date format" hear mean that YYYY-MM-DD is what the ISO version of the C language standard requires, or does it mean something else? --ado

2 1

strftime.c tuneups
by eggert＠twinsun.com 25 May '93

25 May '93

Here are a couple of performance tuneups for strftime.c. 1. Instead of using arrays of pointers to arrays of chars, use two-dimensional arrays of chars, since the sizes are all known. This tends to use less space (since you don't need to store the pointers) and time (since you don't need to indirect through them. 2. Instead of passing around a pointer and pointer to size, incrementing the former and decrementing the latter indirectly, just pass around a pointer and its limit. That way, you can avoid the extra indirection and decrementation. These two changes lop 10% off the code size (sparc, GCC 2.4.0 with -O2). Here's a patch, which assumes my previous patch was applied. *** draft-strftime1.c Tue May 25 11:28:38 1993 --- draft-strftime2.c Tue May 25 21:07:08 1993 *************** *** 32,56 **** #include "private.h" #include "tzfile.h" ! static const char * const afmt[] = { "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" }; ! static const char * const Afmt[] = { "Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday" }; ! static const char * const bfmt[] = { "Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" }; ! static const char * const Bfmt[] = { "January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December" }; ! static char *_add P((const char *, size_t *, char *)); ! static char *_conv P((int, const char *, size_t *, char *)); ! static char *_fmt P((const char *, const struct tm *, size_t *, char *)); size_t strftime P((char *, size_t, const char *, const struct tm *)); --- 32,56 ---- #include "private.h" #include "tzfile.h" ! static const char afmt[][4] = { "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" }; ! static const char Afmt[][10] = { "Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday" }; ! static const char bfmt[][4] = { "Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" }; ! static const char Bfmt[][10] = { "January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December" }; ! static char *_add P((const char *, char *, const char *)); ! static char *_conv P((int, const char *, char *, const char *)); ! static char *_fmt P((const char *, const struct tm *, char *, const char *)); size_t strftime P((char *, size_t, const char *, const struct tm *)); *************** *** 61,82 **** const char *format; const struct tm *t; { ! size_t gsize; ! gsize = maxsize; ! s = _fmt(format, t, &gsize, s); ! if (gsize <= 0) return 0; ! *s = '\0'; ! return maxsize - gsize; } static char * ! _fmt(format, t, gsizep, pt) const char *format; const struct tm *t; - size_t *gsizep; char *pt; { for (; *format; ++format) { if (*format == '%') { --- 61,81 ---- const char *format; const struct tm *t; { ! char *p; ! p = _fmt(format, t, s, s + maxsize); ! if (p == s + maxsize) return 0; ! *p = '\0'; ! return p - s; } static char * ! _fmt(format, t, pt, ptlim) const char *format; const struct tm *t; char *pt; + const char *ptlim; { for (; *format; ++format) { if (*format == '%') { *************** *** 87,109 **** break; case 'A': pt = _add((t->tm_wday < 0 || t->tm_wday > 6) ? ! "?" : Afmt[t->tm_wday], gsizep, pt); continue; case 'a': pt = _add((t->tm_wday < 0 || t->tm_wday > 6) ? ! "?" : afmt[t->tm_wday], gsizep, pt); continue; case 'B': pt = _add((t->tm_mon < 0 || t->tm_mon > 11) ? ! "?" : Bfmt[t->tm_mon], gsizep, pt); continue; case 'b': case 'h': pt = _add((t->tm_mon < 0 || t->tm_mon > 11) ? ! "?" : bfmt[t->tm_mon], gsizep, pt); continue; case 'c': ! pt = _fmt("%D %X", t, gsizep, pt); continue; case 'C': /* --- 86,108 ---- break; case 'A': pt = _add((t->tm_wday < 0 || t->tm_wday > 6) ? ! "?" : Afmt[t->tm_wday], pt, ptlim); continue; case 'a': pt = _add((t->tm_wday < 0 || t->tm_wday > 6) ? ! "?" : afmt[t->tm_wday], pt, ptlim); continue; case 'B': pt = _add((t->tm_mon < 0 || t->tm_mon > 11) ? ! "?" : Bfmt[t->tm_mon], pt, ptlim); continue; case 'b': case 'h': pt = _add((t->tm_mon < 0 || t->tm_mon > 11) ? ! "?" : bfmt[t->tm_mon], pt, ptlim); continue; case 'c': ! pt = _fmt("%D %X", t, pt, ptlim); continue; case 'C': /* *************** *** 114,120 **** ** (ado, 5/24/93) */ pt = _conv((t->tm_year + TM_YEAR_BASE) / 100, ! "%02d", gsizep, pt); continue; case 'D': case 'x': --- 113,119 ---- ** (ado, 5/24/93) */ pt = _conv((t->tm_year + TM_YEAR_BASE) / 100, ! "%02d", pt, ptlim); continue; case 'D': case 'x': *************** *** 129,138 **** ** numbers (as here) makes Quakers happier. ** This is the ISO standard date format. */ ! pt = _fmt("%Y-%m-%d", t, gsizep, pt); continue; case 'd': ! pt = _conv(t->tm_mday, "%02d", gsizep, pt); continue; case 'E': case 'O': --- 128,137 ---- ** numbers (as here) makes Quakers happier. ** This is the ISO standard date format. */ ! pt = _fmt("%Y-%m-%d", t, pt, ptlim); continue; case 'd': ! pt = _conv(t->tm_mday, "%02d", pt, ptlim); continue; case 'E': case 'O': *************** *** 149,166 **** */ goto label; case 'e': ! pt = _conv(t->tm_mday, "%2d", gsizep, pt); continue; case 'H': ! pt = _conv(t->tm_hour, "%02d", gsizep, pt); continue; case 'I': pt = _conv((t->tm_hour % 12) ? (t->tm_hour % 12) : 12, ! "%02d", gsizep, pt); continue; case 'j': ! pt = _conv(t->tm_yday + 1, "%03d", gsizep, pt); continue; case 'k': /* --- 148,165 ---- */ goto label; case 'e': ! pt = _conv(t->tm_mday, "%2d", pt, ptlim); continue; case 'H': ! pt = _conv(t->tm_hour, "%02d", pt, ptlim); continue; case 'I': pt = _conv((t->tm_hour % 12) ? (t->tm_hour % 12) : 12, ! "%02d", pt, ptlim); continue; case 'j': ! pt = _conv(t->tm_yday + 1, "%03d", pt, ptlim); continue; case 'k': /* *************** *** 173,186 **** ** "%l" have been swapped. ** (ado, 5/24/93) */ ! pt = _conv(t->tm_hour, "%2d", gsizep, pt); continue; #ifdef KITCHEN_SINK case 'K': /* ** After all this time, still unclaimed! */ ! pt = _add("kitchen sink", gsizep, pt); continue; #endif /* defined KITCHEN_SINK */ case 'l': --- 172,185 ---- ** "%l" have been swapped. ** (ado, 5/24/93) */ ! pt = _conv(t->tm_hour, "%2d", pt, ptlim); continue; #ifdef KITCHEN_SINK case 'K': /* ** After all this time, still unclaimed! */ ! pt = _add("kitchen sink", pt, ptlim); continue; #endif /* defined KITCHEN_SINK */ case 'l': *************** *** 195,234 **** */ pt = _conv((t->tm_hour % 12) ? (t->tm_hour % 12) : 12, ! "%2d", gsizep, pt); continue; case 'M': ! pt = _conv(t->tm_min, "%02d", gsizep, pt); continue; case 'm': ! pt = _conv(t->tm_mon + 1, "%02d", gsizep, pt); continue; case 'n': ! pt = _add("\n", gsizep, pt); continue; case 'p': pt = _add(t->tm_hour >= 12 ? "PM" : "AM", ! gsizep, pt); continue; case 'R': ! pt = _fmt("%H:%M", t, gsizep, pt); continue; case 'r': ! pt = _fmt("%I:%M:%S %p", t, gsizep, pt); continue; case 'S': ! pt = _conv(t->tm_sec, "%02d", gsizep, pt); continue; case 'T': case 'X': ! pt = _fmt("%H:%M:%S", t, gsizep, pt); continue; case 't': ! pt = _add("\t", gsizep, pt); continue; case 'U': pt = _conv((t->tm_yday + 7 - t->tm_wday) / 7, ! "%02d", gsizep, pt); continue; case 'u': /* --- 194,233 ---- */ pt = _conv((t->tm_hour % 12) ? (t->tm_hour % 12) : 12, ! "%2d", pt, ptlim); continue; case 'M': ! pt = _conv(t->tm_min, "%02d", pt, ptlim); continue; case 'm': ! pt = _conv(t->tm_mon + 1, "%02d", pt, ptlim); continue; case 'n': ! pt = _add("\n", pt, ptlim); continue; case 'p': pt = _add(t->tm_hour >= 12 ? "PM" : "AM", ! pt, ptlim); continue; case 'R': ! pt = _fmt("%H:%M", t, pt, ptlim); continue; case 'r': ! pt = _fmt("%I:%M:%S %p", t, pt, ptlim); continue; case 'S': ! pt = _conv(t->tm_sec, "%02d", pt, ptlim); continue; case 'T': case 'X': ! pt = _fmt("%H:%M:%S", t, pt, ptlim); continue; case 't': ! pt = _add("\t", pt, ptlim); continue; case 'U': pt = _conv((t->tm_yday + 7 - t->tm_wday) / 7, ! "%02d", pt, ptlim); continue; case 'u': /* *************** *** 238,244 **** ** (ado, 5/24/93) */ pt = _conv((t->tm_wday == 0) ? 7 : t->tm_wday, ! "%d", gsizep, pt); continue; case 'V': /* --- 237,243 ---- ** (ado, 5/24/93) */ pt = _conv((t->tm_wday == 0) ? 7 : t->tm_wday, ! "%d", pt, ptlim); continue; case 'V': /* *************** *** 270,276 **** i = (t->tm_yday + 10 - (t->tm_wday ? (t->tm_wday - 1) : 6)) / 7; pt = _conv((i == 0) ? 53 : i, ! "%02d", gsizep, pt); } continue; case 'v': --- 269,275 ---- i = (t->tm_yday + 10 - (t->tm_wday ? (t->tm_wday - 1) : 6)) / 7; pt = _conv((i == 0) ? 53 : i, ! "%02d", pt, ptlim); } continue; case 'v': *************** *** 279,315 **** ** "date as dd-bbb-YYYY" ** (ado, 5/24/93) */ ! pt = _fmt("%e-%b-%Y", t, gsizep, pt); continue; case 'W': pt = _conv((t->tm_yday + 7 - (t->tm_wday ? (t->tm_wday - 1) : 6)) / 7, ! "%02d", gsizep, pt); continue; case 'w': ! pt = _conv(t->tm_wday, "%d", gsizep, pt); continue; case 'y': pt = _conv((t->tm_year + TM_YEAR_BASE) % 100, ! "%02d", gsizep, pt); continue; case 'Y': pt = _conv(t->tm_year + TM_YEAR_BASE, "%04d", ! gsizep, pt); continue; case 'Z': #ifdef TM_ZONE if (t->TM_ZONE) ! pt = _add(t->TM_ZONE, gsizep, pt); else #endif /* defined TM_ZONE */ if (t->tm_isdst == 0 || t->tm_isdst == 1) { extern char * tzname[2]; pt = _add(tzname[t->tm_isdst], ! gsizep, pt); ! } else pt = _add("?", gsizep, pt); continue; case '%': /* --- 278,314 ---- ** "date as dd-bbb-YYYY" ** (ado, 5/24/93) */ ! pt = _fmt("%e-%b-%Y", t, pt, ptlim); continue; case 'W': pt = _conv((t->tm_yday + 7 - (t->tm_wday ? (t->tm_wday - 1) : 6)) / 7, ! "%02d", pt, ptlim); continue; case 'w': ! pt = _conv(t->tm_wday, "%d", pt, ptlim); continue; case 'y': pt = _conv((t->tm_year + TM_YEAR_BASE) % 100, ! "%02d", pt, ptlim); continue; case 'Y': pt = _conv(t->tm_year + TM_YEAR_BASE, "%04d", ! pt, ptlim); continue; case 'Z': #ifdef TM_ZONE if (t->TM_ZONE) ! pt = _add(t->TM_ZONE, pt, ptlim); else #endif /* defined TM_ZONE */ if (t->tm_isdst == 0 || t->tm_isdst == 1) { extern char * tzname[2]; pt = _add(tzname[t->tm_isdst], ! pt, ptlim); ! } else pt = _add("?", pt, ptlim); continue; case '%': /* *************** *** 321,356 **** break; } } ! if (*gsizep <= 0) return pt; *pt++ = *format; - --*gsizep; } return pt; } static char * ! _conv(n, format, gsizep, pt) int n; const char *format; - size_t *gsizep; char *pt; { char buf[21]; /* Room for - 2**63 ("-9223372036854775808") + null. */ (void) sprintf(buf, format, n); ! return _add(buf, gsizep, pt); } static char * ! _add(str, gsizep, pt) const char *str; - size_t *gsizep; char *pt; { ! while (*gsizep > 0 && (*pt = *str++) != '\0') { ++pt; - --*gsizep; } return pt; } --- 320,353 ---- break; } } ! if (pt == ptlim) return pt; *pt++ = *format; } return pt; } static char * ! _conv(n, format, pt, ptlim) int n; const char *format; char *pt; + const char *ptlim; { char buf[21]; /* Room for - 2**63 ("-9223372036854775808") + null. */ (void) sprintf(buf, format, n); ! return _add(buf, pt, ptlim); } static char * ! _add(str, pt, ptlim) const char *str; char *pt; + const char *ptlim; { ! while (pt < ptlim && (*pt = *str++) != '\0') { ++pt; } return pt; }

1 0

updated strftime.c
by ado 25 May '93

25 May '93

An updated draft strftime.c, with changes from Bradley White (bww+(a)transarc.com) to avoid problems on multiprocessor systems, is available for anonymous ftp from elsie.nci.nih.gov (128.231.16.1) in the "pub" directory, file "draft-strftime.c". As Bradley noted in correspondence, it also: - removes _conv()'s static [non-const] buffer - corrects a problem with _conv() most easily seen with "%k" when tm_hour == 0 - can use the first char of _conv()'s new, non-static buffer - eliminates _conv()'s ASCII dependency - adds a missing _fmt()-parameter in the "%v" case (I've reformatted "strftime.c" just slightly from what bww sent me.) Again, any and all feedback is welcome; correspondence might best be sent to tz(a)elsie.nci.nih.gov. --ado

2 1

impending tzcode93b.tar.Z
by ado 24 May '93

24 May '93

Having received useful changes from several sources, I'm hoping to put up tzcode93b.tar.Z for anonymous ftp soon. One thing I'd like to do is bring "strftime" up to date with new standards (akin to Arnold Robbins' version). A proposed "strftime" is attached; grepping for "5/24/93" will show up where changes have happened. Any comments would be appreciated; I'd particularly like word on the matter marked by "XXX" below. --ado #ifndef lint #ifndef NOID static char elsieid[] = "@(#)strftime.c 7.2"; /* ** Based on the UCB version whose ID appears below. ** This is ANSIish only when time is treated identically in all locales and ** when "multibyte character == plain character". */ #endif /* !defined NOID */ #endif /* !defined lint */ /* * Copyright (c) 1989 The Regents of the University of California. * All rights reserved. * * Redistribution and use in source and binary forms are permitted * provided that the above copyright notice and this paragraph are * duplicated in all such forms and that any documentation, * advertising materials, and other materials related to such * distribution and use acknowledge that the software was developed * by the University of California, Berkeley. The name of the * University may not be used to endorse or promote products derived * from this software without specific prior written permission. * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. */ #if defined(LIBC_SCCS) && !defined(lint) static char sccsid[] = "@(#)strftime.c 5.4 (Berkeley) 3/14/89"; #endif /* LIBC_SCCS and not lint */ #include "sys/types.h" #include "time.h" #include "tzfile.h" static char *afmt[] = { "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat", }; static char *Afmt[] = { "Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday", }; static char *bfmt[] = { "Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec", }; static char *Bfmt[] = { "January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December", }; static size_t gsize; static char *pt; static void _add(); static void _conv(); static void _fmt(); size_t strftime(s, maxsize, format, t) char *s; char *format; size_t maxsize; struct tm *t; { pt = s; gsize = maxsize; _fmt(format, t); if (gsize <= 0) return(0); *pt = '\0'; return(maxsize - gsize); } static void _fmt(format, t) register char *format; struct tm *t; { for (; *format; ++format) { if (*format == '%') { label: switch(*++format) { case '\0': --format; break; case 'A': if (t->tm_wday < 0 || t->tm_wday > 6) _add("?"); else _add(Afmt[t->tm_wday]); continue; case 'a': if (t->tm_wday < 0 || t->tm_wday > 6) _add("?"); else _add(afmt[t->tm_wday]); continue; case 'B': if (t->tm_mon < 0 || t->tm_mon > 11) _add("?"); else _add(Bfmt[t->tm_mon]); continue; case 'b': case 'h': if (t->tm_mon < 0 || t->tm_mon > 11) _add("?"); else _add(bfmt[t->tm_mon]); continue; case 'c': _fmt("%D %X", t); continue; case 'C': /* ** %C used to do a... ** _fmt("%a %b %e %X %Y", t); ** ...whereas now POSIX 1003.2 calls for ** something completely different. ** (ado, 5/24/93) */ _conv((t->tm_year + TM_YEAR_BASE) / 100, 2, '0'); continue; case 'D': case 'x': /* ** Version 3.0 of strftime from Arnold Robbins ** (arnold(a)skeeve.atl.ga.us) does the ** equivalent of... ** _fmt("%a %b %e %Y"); ** ...for %x; since the X3J11 C language ** standard calls for "date, using locale's ** date format," anything goes. Using just ** numbers (as here) makes Quakers happier. ** (ado, 5/24/93) */ _fmt("%m/%d/%y", t); continue; case 'd': _conv(t->tm_mday, 2, '0'); continue; case 'E': case 'O': /* ** POSIX locale extensions, a la ** Arnold Robbins' strftime version 3.0. ** The sequences ** %Ec %EC %Ex %Ey %EY ** %Od %oe %OH %OI %Om %OM ** %OS %Ou %OU %OV %Ow %OW %Oy ** are supposed to provide alternate ** representations. ** (ado, 5/24/93) */ goto label; case 'e': _conv(t->tm_mday, 2, ' '); continue; case 'H': _conv(t->tm_hour, 2, '0'); continue; case 'I': _conv(t->tm_hour % 12 ? t->tm_hour % 12 : 12, 2, '0'); continue; case 'j': _conv(t->tm_yday + 1, 3, '0'); continue; case 'k': /* ** This used to be... ** _conv(t->tm_hour % 12 ? ** t->tm_hour % 12 : 12, 2, ' '); ** ...and has been changed to the below to ** match SunOS 4.1.1 and Arnold Robbins' ** strftime version 3.0. That is, "%k" and ** "%l" have been swapped. ** (ado, 5/24/93) */ _conv(t->tm_hour, 2, ' '); continue; #ifdef KITCHEN_SINK case 'K': /* ** After all this time, still unclaimed! */ _add("kitchen sink"); continue; #endif /* defined KITCHEN_SINK */ case 'l': /* ** This used to be... ** _conv(t->tm_hour, 2, ' '); ** ...and has been changed to the below to ** match SunOS 4.1.1 and Arnold Robbin's ** strftime version 3.0. That is, "%k" and ** "%l" have been swapped. ** (ado, 5/24/93) */ _conv(t->tm_hour % 12 ? t->tm_hour % 12 : 12, 2, ' '); continue; case 'M': _conv(t->tm_min, 2, '0'); continue; case 'm': _conv(t->tm_mon + 1, 2, '0'); continue; case 'n': _add("\n"); continue; case 'p': _add(t->tm_hour >= 12 ? "PM" : "AM"); continue; case 'R': _fmt("%H:%M", t); continue; case 'r': _fmt("%I:%M:%S %p", t); continue; case 'S': _conv(t->tm_sec, 2, '0'); continue; case 'T': case 'X': _fmt("%H:%M:%S", t); continue; case 't': _add("\t"); continue; case 'U': _conv((t->tm_yday + 7 - t->tm_wday) / 7, 2, '0'); continue; case 'u': /* ** From Arnold Robbins' strftime version 3.0: ** "ISO 8601: Weekday as a decimal number ** [1 (Monday) - 7]" ** (ado, 5/24/93) */ _conv(((t->tm_wday == 0) ? 7 : t->tm_wday), 1, '0'); continue; case 'V': /* ** From Arnold Robbins' strftime version 3.0: ** "the week number of the year (the first ** Monday as the first day of week 1) as a ** decimal number (01-53). The method for ** determining the week number is as specified ** by ISO 8601 (to wit: if the week containing ** January 1 has four or more days in the new ** year, then it is week 1, otherwise it is ** week 53 of the previous year and the next ** week is week 1)." ** (ado, 5/24/93) */ /* ** XXX--If January 1 falls on a Friday, ** January 1-3 are part of week 53 of the ** previous year. By analogy, if January ** 1 falls on a Thursday, are December 29-31 ** of the PREVIOUS year part of week 1??? ** (ado 5/24/93) /* ** You are understood not to expect this. */ { register int i; i = (t->tm_yday + 10 - (t->tm_wday ? (t->tm_wday - 1) : 6)) / 7; _conv((i == 0) ? 53 : i, 2, '0'); } continue; case 'v': /* ** From Arnold Robbins' strftime version 3.0: ** "date as dd-bbb-YYYY" ** (ado, 5/24/93) */ _fmt("%e-%b-%Y"); continue; case 'W': _conv((t->tm_yday + 7 - (t->tm_wday ? (t->tm_wday - 1) : 6)) / 7, 2, '0'); continue; case 'w': _conv(t->tm_wday, 1, '0'); continue; case 'y': _conv((t->tm_year + TM_YEAR_BASE) % 100, 2, '0'); continue; case 'Y': _conv(t->tm_year + TM_YEAR_BASE, 4, '0'); continue; case 'Z': #ifdef TM_ZONE if (t->TM_ZONE) _add(t->TM_ZONE); else #endif /* defined TM_ZONE */ if (t->tm_isdst == 0 || t->tm_isdst == 1) { extern char * tzname[2]; _add(tzname[t->tm_isdst]); } else _add("?"); continue; case '%': /* * X311J/88-090 (4.12.3.5): if conversion char is * undefined, behavior is undefined. Print out the * character itself as printf(3) also does. */ default: break; } } if (gsize <= 0) return; *pt++ = *format; --gsize; } } static void _conv(n, digits, fill) int n, digits, fill; { static char buf[10]; register char *p; for (p = buf + sizeof(buf) - 2; n > 0 && p > buf; n /= 10, --digits) *p-- = n % 10 + '0'; while (p > buf && digits-- > 0) *p-- = fill; _add(++p); } static void _add(str) register char *str; { for (;; ++pt, --gsize) { if (gsize <= 0) return; if (!(*pt = *str++)) return; } }

1 0

A brief history of (and agenda for) time zone libraries
by eggert＠twinsun.com 15 May '93

15 May '93

Summary: The GNU C library is derived from Arthur David Olson's 1989 time zone code, with extensive changes. The good news is that these changes improve the modularity of Olson's code; the bad news is that it is now hard to merge some of Olson's more recent improvements into the GNU C library. Despite the bad news, Olson's fixes and improvements should be merged into the GNU C library somehow. While we're on the subject, the Linux time zone code also comes from Olson's 1989 code, but has only few minor changes; it should be updated to the current edition of Olson's code. __________ Olson's 1989 code. The original timezone code was written by Arthur David Olson. This public domain code has a good reputation. Even the Posix standard comes as close as a standard can to saying ``This is good stuff -- check it out!''; see Posix 1003.1-1990 section B.8.1.1, lines 4195-4201. The most famous publication of Olson's code was in comp.sources.unix 18, 111-117 (19 April 1989) -- see ftp.uu.net:/usenet/comp.sources.unix/volume18/localtime3/*. GNU C Library. The GNU C Library (see prep.ai.mit.edu:pub/gnu/glibc-1.05.tar.z) is derived from Olson's 1989 code. Here are its changes: 1. Rewrite the Makefile to use GNU make and GNU libc standards. 2. Split localtime.c into several smaller, more modular files. These files are: __tzset.c gmtime.c mktime.c offtime.c tzfile.c tzset.c. Completely rewrite these files in the GNU style. 3. Add some time functions not in Olson's package, e.g. adjtime. 4. Update some of the New Zealand time zone rules. Linux. Linux's code for timezones (see tsx-11.mit.edu:pub/linux/sources/sbin/timesrc.tar.Z) is also derived from Olson's 1989 code, but it has only the following minor changes: 1. Rewrite the Makefile to meet Linux standards. 2. Add a few #includes, or change them from "" form to <> form. Olson's current edition. Arthur David Olson is still maintaining both the code and the corresponding timezone database (see elsie.nci.nih.gov:pub/tz*). The differences between Olson's 1989 code and his current edition are as follows: 1. Fix some minor Makefile portability problems. 2. Fix some bugs in the C code, and avoid some possible core dumps. 3. Bring the leap second database up-to-date through June 1993. 4. Update the time zone database for Israel, Australia (which is particularly wild and woolly -- they hold voter referendums to change the timezone rules!), New Zealand, Britain, Portugal, Starke County Indiana, Saskatchewan. Remove the nonexistent US Pacific Presidential Election Time, which caused some problems late last year on some West Coast Suns. The New Zealand updates are a superset of the GNU C library's updates. 5. Add a -y option to zic to specify a command that tests for arbitrary year types; this is used in some of the new time zone rules. Proposal. Olson's changes should be propagated into both Linux and the GNU C library. Linux should update to Olson's current edition; this should be a simple merge. By the way, the Linux code should also give credit to Olson instead of silently removing his README file; this will help prevent the current situation from recurring. The GNU C library changes fall into two categories: (A) changes to Olson's localtime.c, and (B) everything else. Merging (B) is easy, almost as easy as merging the Linux timezone code, and I've already submitted patches to the GNU C library developers to do this. Merging (A) is harder, since the GNU C library has split up localtime.c into lots of little pieces. The GNU C library changes generally improve Olson's code -- they make the module interfaces cleaner, and the code is easier to follow. Unfortunately, it is now hard to tell how to propagate Olson's localtime.c changes and improvements into the GNU C library. The best person to do this propagation is the person who split up localtime.c in the first place. Another possibility is for the GNU C library to switch to a version of localtime.c that is as close as possible to Olson's latest version; that might simplify future maintenance.

2 2

Re: A brief history of (and agenda for) time zone libraries
by ado 15 May '93

15 May '93

Improvidently, a couple of things make buying back the GNU changes at this end difficult. One is not so bad. Since the GNU folks are clearly implementers (in the ANSI sense of the word), they're entitled to use names from the implementers' name space in their code; this means that strictly internal functions can be split into separate source files without polluting the user's name space. It's not clear that this is kosher when the timezone package is distributed independently. But some care in packaging and some creative use of the preprocessor would avoid most challenges here. One is not so tractable. Buying back the GNU code at this point would also involve putting things under the GPL, which I'm loathe to do. --ado

1 0

Re: A brief history of (and agenda for) time zone libraries
by ado 15 May '93

15 May '93

> From roland(a)gnu.ai.mit.edu Sat May 15 18:14:48 1993 > Return-Path: <roland(a)gnu.ai.mit.edu> > Received: from churchy.gnu.ai.mit.edu by elsie.nci.nih.gov (4.1/SMI-4.1) > id AA22472; Sat, 15 May 93 18:14:34 EDT > Received: by churchy.gnu.ai.mit.edu (5.65/4.0) > id <AA21556(a)churchy.gnu.ai.mit.edu>; Sat, 15 May 93 18:14:01 -0400 > Date: Sat, 15 May 93 18:14:01 -0400 > From: roland(a)gnu.ai.mit.edu (Roland McGrath) > Message-Id: <9305152214.AA21556(a)churchy.gnu.ai.mit.edu> > To: eggert(a)twinsun.com > Cc: libc-alpha(a)gnu.ai.mit.edu, tz(a)elsie.nci.nih.gov > Subject: Re: A brief history of (and agenda for) time zone libraries > In-Reply-To: Paul Eggert's message of Sat, 15 May 93 00:08:44 PDT <9305150708.AA18312(a)bi.twinsun.com> > X-Shopping-List: (1) Hot mutant holidays > (2) Apologetic interruptions > (3) Frenetic lesions > > > Olson's 1989 code. > > > > The original timezone code was written by Arthur David Olson. This public > > domain code has a good reputation. Even the Posix standard comes as close as > > a standard can to saying ``This is good stuff -- check it out!''; see Posix > > 1003.1-1990 section B.8.1.1, lines 4195-4201. The most famous publication of > > Olson's code was in comp.sources.unix 18, 111-117 (19 April 1989) -- see > > ftp.uu.net:/usenet/comp.sources.unix/volume18/localtime3/*. > > I wish I had known that this code was being maintained, and where to get it > from, before now. > > > Proposal. > > > > Olson's changes should be propagated into both Linux and the GNU C library. > > Indeed. Linux is not my concern, but I plan to update glibc today. > > > The GNU C library changes fall into two categories: (A) changes to > > Olson's localtime.c, and (B) everything else. Merging (B) is easy, > > almost as easy as merging the Linux timezone code, and I've already > > submitted patches to the GNU C library developers to do this. Merging > > (A) is harder, since the GNU C library has split up localtime.c into lots > > of little pieces. The GNU C library changes generally improve Olson's > > code -- they make the module interfaces cleaner, and the code is easier > > to follow. > > These issues are very important to the GNU project, and to me. > > > Unfortunately, it is now hard to tell how to propagate Olson's > > localtime.c changes and improvements into the GNU C library. The best > > person to do this propagation is the person who split up localtime.c in > > the first place. > > I am that person, and I plan to do the work. > > > Another possibility is for the GNU C library to switch to a version of > > localtime.c that is as close as possible to Olson's latest version; that > > might simplify future maintenance. > > I really don't want to do this. It took me quite a while to untangle the > spaghetti in the first place. This would certainly not simplify > maintenance of the glibc code; it would return it to the obfuscated > spaghetti I started with. Having the code be modular and clean simplifies > maintenance for me. Using code close Olsen's code only simplifies future > merging with Olsen's code. >

1 0

proactive DST discovery?
by ado 12 May '93

12 May '93

If you have insights on the advisability of doing network polling to learn of time changes, or have notions on how it might best be done, I'd appreciate your sharing them with tz(a)elsie.nci.nih.gov. -- Arthur David Olson ado(a)elsie.nci.nih.gov ADO and Elsie are Ampex and Borden trademarks "REM AND STINPY"--found in Bill Gates' startup file

1 0