- tz - lists.iana.org

Palestine time
by amos＠nsof.co.il Feb. 15, 1998

Feb. 15, 1998

In the asia file you include: > # Zone NAME GMTOFF RULES FORMAT [UNTIL] > Zone Asia/Gaza 2:17:52 - LMT 1900 Oct > 2:00 - EET 1957 May 10 > 2:00 EgyptAsia EE%sT 1967 Jun 30 > 2:00 Zion I%sT From 1917 until 1948/5/15, all of Palestine, including the parts now known as the Gaza Strip and the West Bank, was under British rule. Therefore the rules given for Israel for that period, apply there too. (Though I find the name "Zion" rather inappropriate -- why not just call it "Israel" or "Palestine"?) The Gaza Strip was under Egyptian rule between 1948/5/15 until 1967/6/5 (except a short occupation by Israel from 1956/11 till 1957/3, but no time zone was affected then). It was never formally annexed to Egypt, though. The rest of Palestine was under Jordanian rule at that time, formally annexed in 1950 as the West Bank (and the word "Trans" was dropped from the country's previous name of "the Hashemite Kingdom of the Trans-Jordan"). So the rules for Jordan for that time apply. Major towns in that area are Nablus (Shchem), El-Halil (Hebron), Ramallah, and East Jerusalem. Both areas were occupied by Israel in June 1967, but not annexed (except for East Jerusalem). They were on Israel time since then; there might have been a Military Governor's order about time zones, but I'm not aware of any (such orders may have been issued semi-annually whenever summer time was in effect, but maybe the legal aspect of time was just neglected). The Palestinian Authority was established in 1993, and got hold of most towns in the West Bank and Gaza by 1995. I know that in order to demonstrate some manner of independence, they have been switching to summer time and back on a different schedule than Israel's, but I don't know when this was started, or what algorithm is used (most likely the Jordanian one). To summarize, the table should probably look something like that: Area \ when | 1918-1947 | 1948-1967 | 1967-1995 | 1996- ------------+-----------+-----------+-----------+----------- Israel | Zion | Zion | Zion | Zion West bank | Zion | Jordan | Zion | Jordan Gaza | Zion | Egypt | Zion | Jordan I guess more info may be available from the PA's web page (if/when they have one). -- Amos Shapir Paper: nSOF Parallel Software, Ltd. Givat-Hashlosha 48800, Israel Tel: +972 3 9388551 Fax: +972 3 9388552 GEO: 34 55 15 E / 32 05 52 N

1 0

Brazil can't decide on daylight savings
by Daniel C. Sobral Feb. 15, 1998

Feb. 15, 1998

Ok, my previous patch had an error. Here is a newer and, hopefully, correct patch. -- Daniel C. Sobral (8-DCS) dcs(a)gns.com.br When it comes to helping you, some people stop at nothing.

1 0

FW: Brazil can't decide on daylight savings
by Olson, Arthur David Feb. 12, 1998

Feb. 12, 1998

---------- From: Daniel C. Sobral[SMTP:dcs@gns.com.br] Sent: Thursday, February 12, 1998 10:50 AM To: tz(a)elsie.nci.nih.gov Subject: Brazil can't decide on daylight savings This is a patch I submitted to FreeBSD regarding recent and very recent changes on Daylight Savings beginning and ending dates on Brazil. In 1997, the DS began on October 6. The stated reason was that because international television networks ignored Brazil's policy on DS, they bought the wrong times on satellite for coverage of Pope's visit. This year, the ending date of DS was postponed to March first to help dealing with the shortages of electric power. I was also told to send you any citations demonstrating the announcements dates, but I can't. The ending date was changed this week, I'm almost sure the announcement was made tuesday (it was being expected, and I was first notified of it through a backweb newspaper service). As for the change in 1997, I really don't recall any dates (besides the ones that have direct impact to me, being a sysadm, of course). The patch is attached. -- Daniel C. Sobral (8-DCS) dcs(a)gns.com.br Every program has (at least) two purposes: the one for which it was written and another for which it wasn't. <</usr/src/share/zoneinfo/patch>> begin 600 patch M*BHJ('-O=71H86UE<FEC82YO<FEG"5=E9"!&96(@,3$@,C$Z,#8Z,#D@,3DY M.`T*+2TM('-O=71H86UE<FEC80E7960@1F5B(#$Q(#(Q.C`V.C4X(#$Y.3@- M"BHJ*BHJ*BHJ*BHJ*BHJ*@T**BHJ(#(W,BPR-S<@*BHJ*@T*+2TM(#(W,BPR M-SD@+2TM+0T*("!2=6QE"4)R87II;`DQ.3DT"3$Y.34)+0E&96()4W5N/CTQ M-0D@,#HP,`DP"5,-"B`@4G5L90E"<F%Z:6P),3DY-@EM87@)+0E&96()4W5N M/CTQ,0D@,#HP,`DP"5,-"B`@4G5L90E"<F%Z:6P),3DY-@EM87@)+0E/8W0) M4W5N/CTQ"2`P.C`P"3$Z,#`)1`T**R!2=6QE"4)R87II;`DQ.3DW"6]N;'D) M+0E/8W0)(#8)(#`Z,#`),3HP,`E$#0HK(%)U;&4)0G)A>FEL"3$Y.3@);VYL M>0DM"4UA<@D@,0D@,#HP,`DP"5,-"B`@#0H@(",(a)6F]N90E.04U%"0E'351/ >1D8)4E5,15,)1D]234%4"5M53E1)3%T-"B`@(PT* ` end

1 0

Brazil can't decide on daylight savings
by Daniel C. Sobral Feb. 12, 1998

Feb. 12, 1998

This is a patch I submitted to FreeBSD regarding recent and very recent changes on Daylight Savings beginning and ending dates on Brazil. In 1997, the DS began on October 6. The stated reason was that because international television networks ignored Brazil's policy on DS, they bought the wrong times on satellite for coverage of Pope's visit. This year, the ending date of DS was postponed to March first to help dealing with the shortages of electric power. I was also told to send you any citations demonstrating the announcements dates, but I can't. The ending date was changed this week, I'm almost sure the announcement was made tuesday (it was being expected, and I was first notified of it through a backweb newspaper service). As for the change in 1997, I really don't recall any dates (besides the ones that have direct impact to me, being a sysadm, of course). The patch is attached. -- Daniel C. Sobral (8-DCS) dcs(a)gns.com.br Every program has (at least) two purposes: the one for which it was written and another for which it wasn't.

1 0

symlinks
by Olson, Arthur David Feb. 3, 1998

Feb. 3, 1998

We do indeed need an addition to #define HAVE_SYMLINK in private.h; I've added it. The peculiar #if (HAVE_SYMLINK - 0) compiles when HAVE_SYMLINK is undefined; it's the form used in many places in the code. Since the behavior of zic is being changed, a warning seems mandatory. Since there's been one reported challenge worldwide with the lack of symlinks over the last decade, I don't sense an urgent need to be able to disable the warning. --ado *** 7.46/private.h Tue Feb 3 09:20:01 1998 --- 7.47/private.h Tue Feb 3 09:20:01 1998 *************** *** 21,27 **** #ifndef lint #ifndef NOID ! static char privatehid[] = "@(#)private.h 7.46"; #endif /* !defined NOID */ #endif /* !defined lint */ --- 21,27 ---- #ifndef lint #ifndef NOID ! static char privatehid[] = "@(#)private.h 7.47"; #endif /* !defined NOID */ #endif /* !defined lint */ *************** *** 46,51 **** --- 46,55 ---- #define HAVE_STRERROR 0 #endif /* !defined HAVE_STRERROR */ + #ifndef HAVE_SYMLINK + #define HAVE_SYMLINK 1 + #endif /* !defined HAVE_SYMLINK */ + #ifndef HAVE_UNISTD_H #define HAVE_UNISTD_H 1 #endif /* !defined HAVE_UNISTD_H */

1 0

Re: symlinks
by Ulrich Drepper Jan. 31, 1998

Jan. 31, 1998

[I reposted this because it bounced. Sorry if you get this twice.] ado(a)elsie.nci.nih.gov (Arthur David Olson) writes: > --- 601,619 ---- > if (!itsdir(toname)) > (void) remove(toname); > if (link(fromname, toname) != 0) { > + int result; > + > if (mkdirs(toname) != 0) > (void) exit(EXIT_FAILURE); > ! result = link(fromname, toname); > ! #if (HAVE_SYMLINK - 0) > ! if (result != 0) { > ! result = symlink(fromname, toname); > ! if (result == 0) > ! warning(_("hard link failed, symbolic link used")); > ! } > ! #endif > ! if (result != 0) { > const char *e = strerror(errno); > > (void) fprintf(stderr, If you decide to emit a warning in the case an symlink is used there should also be a possibility to prevent this. It's not nice for people who have a setup which is effected by this change to be "punished" with a warning everytime though they already know this happens. So I suggest to use either an option to force using symlinks or an option to make zic quiet. I very much prefer the first. Thanks, -- Uli ---------------. drepper at gnu.org ,-. Rubensstrasse 5 Ulrich Drepper \ ,-------------------' \ 76149 Karlsruhe/Germany Cygnus Solutions `--' drepper at cygnus.com `------------------------ --UAA02321.886276016/ilk.de--

1 0

symlinks
by ado Jan. 31, 1998

Jan. 31, 1998

Below is my guess as to the best way to handle the symlinks situation; absent adverse feedback, it will show up in the next tzcode1998... --ado *** 7.64/Makefile Sat Jan 31 13:53:19 1998 --- 7.65/Makefile Sat Jan 31 13:53:19 1998 *************** *** 1,4 **** ! # @(#)Makefile 7.64 # Change the line below for your time zone (after finding the zone you want in # the time zone files, or adding it to a time zone file). --- 1,4 ---- ! # @(#)Makefile 7.65 # Change the line below for your time zone (after finding the zone you want in # the time zone files, or adding it to a time zone file). *************** *** 91,96 **** --- 91,97 ---- # -DHAVE_SETTIMEOFDAY=2 if settimeofday uses 2nd arg (4.3BSD) # -DHAVE_SETTIMEOFDAY=3 if settimeofday ignores 2nd arg (4.4BSD) # -DHAVE_STRERROR=1 if `strerror' works + # -DHAVE_SYMLINK=0 if your system lacks the symlink function # -DLOCALE_HOME=\"path\" if locales are in "path", not "/usr/lib/locale" # -DHAVE_UNISTD_H=0 if your compiler lacks a "unistd.h" (Microsoft C++ 7?) # -DHAVE_UTMPX_H=1 if your compiler has a "utmpx.h" *** 7.93/zic.c Sat Jan 31 13:53:21 1998 --- 7.94/zic.c Sat Jan 31 13:53:21 1998 *************** *** 1,6 **** #ifndef lint #ifndef NOID ! static char elsieid[] = "@(#)zic.c 7.93"; #endif /* !defined NOID */ #endif /* !defined lint */ --- 1,6 ---- #ifndef lint #ifndef NOID ! static char elsieid[] = "@(#)zic.c 7.94"; #endif /* !defined NOID */ #endif /* !defined lint */ *************** *** 601,609 **** if (!itsdir(toname)) (void) remove(toname); if (link(fromname, toname) != 0) { if (mkdirs(toname) != 0) (void) exit(EXIT_FAILURE); ! if (link(fromname, toname) != 0) { const char *e = strerror(errno); (void) fprintf(stderr, --- 601,619 ---- if (!itsdir(toname)) (void) remove(toname); if (link(fromname, toname) != 0) { + int result; + if (mkdirs(toname) != 0) (void) exit(EXIT_FAILURE); ! result = link(fromname, toname); ! #if (HAVE_SYMLINK - 0) ! if (result != 0) { ! result = symlink(fromname, toname); ! if (result == 0) ! warning(_("hard link failed, symbolic link used")); ! } ! #endif ! if (result != 0) { const char *e = strerror(errno); (void) fprintf(stderr,

1 0

another proposol to fix the zic -l bug
by Ulrich Drepper Jan. 20, 1998

Jan. 20, 1998

Hi, could the appended patch be considered for the next tzcode release? It is an IMO cleaner rewrite since it preserve the old semantic and allows the symlink function (which isn't available on SysV) be skipped by defining a simple symbol. Thanks, -- Uli ---------------. drepper at gnu.org ,-. Rubensstrasse 5 Ulrich Drepper \ ,-------------------' \ 76149 Karlsruhe/Germany Cygnus Solutions `--' drepper at cygnus.com `------------------------ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Index: zic.c =================================================================== RCS file: /egcs/carton/cvsfiles/libc/time/zic.c,v retrieving revision 1.22 diff -c -r1.22 zic.c *** zic.c 1997/12/22 20:49:08 1.22 --- zic.c 1998/01/20 23:09:26 *************** *** 601,610 **** if (!itsdir(toname)) (void) remove(toname); if (link(fromname, toname) != 0) { ! if (mkdirs(toname) != 0) ! (void) exit(EXIT_FAILURE); ! if (link(fromname, toname) != 0) { ! const char *e = strerror(errno); (void) fprintf(stderr, _("%s: Can't link from %s to %s: %s\n"), --- 601,623 ---- if (!itsdir(toname)) (void) remove(toname); if (link(fromname, toname) != 0) { ! int failure = errno; ! if (failure == ENOENT) ! if (mkdirs(toname) != 0) ! failure = errno; ! else if (link(fromname, toname) == 0) ! failure = 0; ! else ! failure = errno; ! #ifndef MISSING_SYMLINK ! if (failure == EXDEV) ! if (symlink(fromname, toname) != 0) ! failure = errno; ! else ! failure = 0; ! #endif ! if (failure) { ! const char *e = strerror(failure); (void) fprintf(stderr, _("%s: Can't link from %s to %s: %s\n"),

1 0

Re: libc/413: "zic -l" fails (cross-device link)
by Zack Weinberg Jan. 18, 1998

Jan. 18, 1998

The appended patch should fix the problem you reported. I am not certain that this fix is the right one, and am cc:ing the tz maintainers for comments. zw Sun Jan 18 16:37:02 1998 Zack Weinberg <zack(a)rabi.phys.columbia.edu> * time/zic.c (dolink): Try symlink() if link() fails with EXDEV. --- zic.c.old Sun Jan 18 16:27:39 1998 +++ zic.c Sun Jan 18 16:36:29 1998 @@ -600,10 +600,10 @@ */ if (!itsdir(toname)) (void) remove(toname); + if (mkdirs(toname) != 0) + (void) exit(EXIT_FAILURE); if (link(fromname, toname) != 0) { - if (mkdirs(toname) != 0) - (void) exit(EXIT_FAILURE); - if (link(fromname, toname) != 0) { + if (errno != EXDEV || symlink(fromname, toname) != 0) { const char *e = strerror(errno); (void) fprintf(stderr,

1 0

FW: zic using hardlinks:-(
by Olson, Arthur David Jan. 17, 1998

Jan. 17, 1998

Andreas Jaeger is not on the time zone list; be sure to include Andreas in any reply. --ado ---------- From: Andreas Jaeger[SMTP:aj@arthur.rhein-neckar.de] Sent: Thursday, January 15, 1998 3:34 PM To: tz(a)elsie.nci.nih.gov Subject: zic using hardlinks:-( Hi, I received the following bug report for GNU libc. Jochen says that using a hard link doesn't work for his system since /etc and /usr/share are on different filesystems. zic is from tzcode1997h. I see three possibilities: - changing the link call in zic.c to symlink. - if link fails with error EXDEV (oldpath and newpath are not on the same filesystem) retry with a symlink. - leave it the way it is. Should symbolic links used in general or only for localtime? Andreas -- Andreas Jaeger aj(a)arthur.rhein-neckar.de >Number: 413 >Category: libc >Synopsis: `zic -l' tries to hard link a file between different disk partitions >Confidential: no >Severity: non-critical >Priority: low >Responsible: libc-gnats >State: open >Class: sw-bug >Submitter-Id: unknown >Arrival-Date: Thu Jan 15 07:32:17 EST 1998 >Last-Modified: >Originator: Jochen Voss >Organization: Department of Mathematics, University of Kaiserslautern/Germany >Release: libc-2.0.6 >Environment: Host type: i486-pc-linux-gnu System: Linux tatonka 2.0.33 #2 Fri Jan 9 20:20:45 MET 1998 i486 Architecture: i486 Addons: localedata crypt linuxthreads Build CFLAGS: -O3 -m486 -fomit-frame-pointer Build CC: gcc -B$(common-objpfx) Build shared: yes Build profile: no Build omitfp: no Stdio: libio >Description: The command "zic -l MET" should set the local time zone by linking the file "/usr/share/zoneinfo/MET" to "/etc/localtime". It uses a hard link for this purpose. On my system this fails, because the directories "/etc" and "/usr/share/zoneinfo" are located on different disk partitions on my system. This results in the output zic: Kann nicht von /usr/share/zoneinfo/MET nach /etc/localtime linken: Ungltiger Link ber Gertegrenzen hinweg when I execaute the above command (obviously I use `LANG=de'). >How-To-Repeat: Find a system, which uses disk partitions as described above, and execute the command "zic -l MET". >Fix: Modify the code in "time/zic.c" to use symlinks (if available) or to copy the file if the hard link fails. >Audit-Trail: >Unformatted:

2 1