Re: [Chicken-users] choosing right timezone for future date
On Jan 28, 2008 5:27 PM, John Cowan <[EMAIL PROTECTED]> wrote: > Graham Fawcett scripsit: > Here's a demo program that does work: [snip] > However, this doesn't really solve your problem, because it sets > the offset and tzname to the current values, not to the values in > effect at the time. I think I've got it. It's localtime(3) that sets the tz, so converting to time_t and back invokes the TZ lookup: #include #include #include void get_tz(int y, int m, int d, int hh, int mm, char *buf) { struct tm tm; time_t t; if (hh == -1) { // if no hour, assume noon. hh = 12; mm = 0; } memset(&tm, 0, sizeof(tm)); // just to be safe tm.tm_year = y - 1900; tm.tm_mon = m; tm.tm_mday = d; tm.tm_hour = hh; tm.tm_min = mm; t = mktime(&tm); localtime_r(&t, &tm); // localtime sets the tz. strftime(buf, sizeof(buf), "%Z", &tm); } int main() { char buf[200]; get_tz(2008, 1, 1, 0, 0, buf); // January 1 printf("%s\n", buf); get_tz(2008, 8, 1, 0, 0, buf); // August 1 printf("%s\n", buf); } $ ./a.out EST EDT $ TZ=Europe/London ./a.out GMT BST Thanks John & Kon for your help, you set me on the right path. Graham ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] choosing right timezone for future date
On Jan 28, 2008, at 2:27 PM, John Cowan wrote: Graham Fawcett scripsit: My TZ is set properly, and using strace I see that the correct zoneinfo file is being accessed. "date +%Z" works properly. Any ideas? However, this doesn't really solve your problem, because it sets the offset and tzname to the current values, not to the values in effect at the time. The TZ database contains historical DST rules. The usual Unix suspects provide a compiled form (Windows doesn't use the TZ database), SLIB has code to read the binary form (tzfile.scm) and use the info (timecore.scm & timezone.scm). (I have been meaning to steal the algorithms for the srfi-19 egg.) I admit I don't know how to make progress here. Let me think on it. -- John Cowan [EMAIL PROTECTED] http://www.ccil.org/~cowan O beautiful for patriot's dream that sees beyond the years Thine alabaster cities gleam undimmed by human tears! America! America! God mend thine every flaw, Confirm thy soul in self-control, thy liberty in law! -- one of the verses not usually taught in U.S. schools ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users Best Wishes, Kon ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] choosing right timezone for future date
Graham Fawcett scripsit: > My TZ is set properly, and using strace I see that the correct > zoneinfo file is being accessed. "date +%Z" works properly. Any ideas? The problem with your test program is that tm is uninitialized, which strptime does not alter (since there is no zone information in the input) and which causes strftime to bomb. Using memset(&tm, 0, sizeof(tm)) causes %z to be + (wrong) and %Z to be "EST" (correct). So that's not a feasible approach. Here's a demo program that does work: #include #include #include int main() { struct tm * tm; char buf[1000]; time_t t; buf[0] = 0; t = time(NULL); tm = localtime(&t); strptime("2008-11-12 18:31:01", "%Y-%m-%d %H:%M:%S", tm); strftime(buf, sizeof(buf), "%d %b %Y %H:%M %z %Z", tm); puts(buf); return 0; } However, this doesn't really solve your problem, because it sets the offset and tzname to the current values, not to the values in effect at the time. I admit I don't know how to make progress here. Let me think on it. -- John Cowan [EMAIL PROTECTED] http://www.ccil.org/~cowan O beautiful for patriot's dream that sees beyond the years Thine alabaster cities gleam undimmed by human tears! America! America! God mend thine every flaw, Confirm thy soul in self-control, thy liberty in law! -- one of the verses not usually taught in U.S. schools ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] choosing right timezone for future date
On Jan 28, 2008 3:45 PM, Graham Fawcett <[EMAIL PROTECTED]> wrote: > > strftime(3), which should be in unit posix but isn't, is your friend; > > it's the library routine underlying date(1). John, I'm seeing an odd behaviour with strftime on my machine; perhaps you've seen it before and might be able to spot the problem. The manpage for strptime gives the following example program, to which I added "%Z %z" in the strftime line: #include #include int main() { struct tm tm; char buf[255]; strptime("2008-11-12 18:31:01", "%Y-%m-%d %H:%M:%S", &tm); strftime(buf, sizeof(buf), "%d %b %Y %H:%M %Z %z", &tm); puts(buf); return 0; } but the output I get is: $ ./a.out 12 Nov 2008 18:31 \360\204 My TZ is set properly, and using strace I see that the correct zoneinfo file is being accessed. "date +%Z" works properly. Any ideas? Thanks, Graham ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] choosing right timezone for future date
On Jan 28, 2008 2:12 PM, John Cowan <[EMAIL PROTECTED]> wrote: > Graham Fawcett scripsit: > > > This may not be Chicken-specific, but is there an existing library > > that, given a date (year,month,day) and given the current locale, will > > tell me the timezone in which the date falls? I > > The locale as such is not relevant to time zones, which are stored in > the TZ environment variable. ("en_US" is just one locale, but it > spans eight time zones.) D'oh, right. > > I can't find anything in SRFI-19 to help me here. Using date(1), I can > > do 'date -d "mmdd" +%z' but that seems silly... > > strftime(3), which should be in unit posix but isn't, is your friend; > it's the library routine underlying date(1). Thanks, John. Graham ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] choosing right timezone for future date
Graham Fawcett scripsit: > This may not be Chicken-specific, but is there an existing library > that, given a date (year,month,day) and given the current locale, will > tell me the timezone in which the date falls? I'm in the EST zone, but > during warmer months, dates fall within the EDT zone. I'd like to be > able to create a date and have the library provide the correct > timezone. (Unix-specific answers are OK.) The locale as such is not relevant to time zones, which are stored in the TZ environment variable. ("en_US" is just one locale, but it spans eight time zones.) > I can't find anything in SRFI-19 to help me here. Using date(1), I can > do 'date -d "mmdd" +%z' but that seems silly... strftime(3), which should be in unit posix but isn't, is your friend; it's the library routine underlying date(1). -- Winter: MIT, John Cowan Keio, INRIA,[EMAIL PROTECTED] Issue lots of Drafts. http://www.ccil.org/~cowan So much more to understand! Might simplicity return?(A "tanka", or extended haiku) ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users