meulenbr@cstw01.UUCP (Frans Meulenbroeks) (01/02/89)
Happy New Year with much minix!
As probably all people dedicated to minix have noticed there is a bug in
the date program from minix (at least in the ST version, but I think
that it appears in the PC version as well).
The problem is that date does not work properly on 12/31 on leap years.
(like 12/31/88). This is due to a bug in the cv_time module.
The (untested) shar at the end cures this problem.
I got also a bug report that cc would not work properly on this day.
I have not been able to verify this. The person who told me about this,
said that it happened on an ST. I'm unsure if this happens on a PC as
well.
Something unrelated:
Is someone aware of a PD SCCS/RCS clone/alternative?
Any pointers would be appreciated greatly.
Frans (.sig at end)
# This is a shell archive.
# Remove everything above and including the cut line.
# Then run the rest of the file through sh.
-----cut here-----cut here-----cut here-----cut here-----
#!/bin/sh
# shar: Shell Archiver
# Run the following text with /bin/sh to create:
# date.cdiff
# This archive created: Mon Jan 2 09:51:45 1989
sed 's/^X//' << \SHAR_EOF > date.cdiff
X*** date.c.orig Mon Jan 2 09:46:53 1989
X--- date.c Mon Jan 2 09:47:48 1989
X*************** long t
X*** 65,72 ****
X while (t >= s_p_year) {
X if (((tm.year + 2) % 4) == 0)
X t -= s_p_day;
X! tm.year += 1;
X! t -= s_p_year;
X }
X if (((tm.year + 2) % 4) == 0)
X days_per_month[1]++;
X--- 65,74 ----
X while (t >= s_p_year) {
X if (((tm.year + 2) % 4) == 0)
X t -= s_p_day;
X! if (t >= s_p_year) {
X! tm.year += 1;
X! t -= s_p_year;
X! }
X }
X if (((tm.year + 2) % 4) == 0)
X days_per_month[1]++;
SHAR_EOF
# End of shell archive
exit 0
--
Frans Meulenbroeks (meulenbr@cst.prl.philips.nl)
Centre for Software Technology
( or try: ...!mcvax!philmds!prle!cst!meulenbr)meulenbr@cst.UUCP (Frans Meulenbroeks) (01/03/89)
Oops I goofed. As I said I sent an untested shar. Well I learned to do this never again. I've canceled the original article. Here is it again but now with a proper and tested fix. Sorry, Frans. >Happy New Year with much minix! > >As probably all people dedicated to minix have noticed there is a bug in meant was: addicted >the date program from minix (at least in the ST version, but I think >that it appears in the PC version as well). > >The problem is that date does not work properly on 12/31 on leap years. >(like 12/31/88). This is due to a bug in the cv_time module. >The (untested) shar at the end cures this problem. now it is tested, and it works > >I got also a bug report that cc would not work properly on this day. >I have not been able to verify this. The person who told me about this, >said that it happened on an ST. I'm unsure if this happens on a PC as >well. I could repeat this. /usr/lib/cem crashes with signal 11. This only happens on 12/31 on leap years. No explanation. No fix, since I do not have source from cem. Perhaps the same problem as above. Perhaps this problem only exist in Minix/ST. > >Something unrelated: >Is someone aware of a PD SCCS/RCS clone/alternative? >Any pointers would be appreciated greatly. > >Frans (.sig at end) > # This is a shell archive. # Remove everything above and including the cut line. # Then run the rest of the file through sh. -----cut here-----cut here-----cut here-----cut here----- #!/bin/sh # shar: Shell Archiver # Run the following text with /bin/sh to create: # date.cdiff # This archive created: Mon Jan 2 19:28:48 1989 sed 's/^X//' << \SHAR_EOF > date.cdiff X*** date.c.orig Tue Sep 6 21:30:13 1988 X--- date.c Sat Dec 31 12:43:35 1988 X*************** X*** 63,70 **** X tm.min = 0; X tm.sec = 0; X while (t >= s_p_year) { X! if (((tm.year + 2) % 4) == 0) X! t -= s_p_day; X tm.year += 1; X t -= s_p_year; X } X--- 63,75 ---- X tm.min = 0; X tm.sec = 0; X while (t >= s_p_year) { X! if (((tm.year + 2) % 4) == 0) X! { X! if (t >= s_p_year + s_p_day) { X! t -= s_p_day; X! } X! else break; /* 12/3 in a leap year */ X! } X tm.year += 1; X t -= s_p_year; X } SHAR_EOF # End of shell archive exit 0 -- Frans Meulenbroeks (meulenbr@cst.prl.philips.nl) Centre for Software Technology ( or try: ...!mcvax!philmds!prle!cst!meulenbr)
ast@cs.vu.nl (Andy Tanenbaum) (01/04/89)
In article <321@cstw01.UUCP> meulenbr@cst.UUCP () writes: >Happy New Year with much minix! > >As probably all people dedicated to minix have noticed there is a bug in >the date program from minix (at least in the ST version, but I think >that it appears in the PC version as well). The bug was discovered and fixed after MINIX-ST went to press, but it has been fixed in 1.3. In the interest of keeping date-ST == date-PC, here is date-PC etc. Andy Tanenbaum : This is a shar archive. Extract with sh, not csh. : This archive ends with exit, so do not worry about trailing junk. : --------------------------- cut here -------------------------- PATH=/bin:/usr/bin:/usr/ucb echo Extracting '/local/ast/minix/tape3/commands/date.c' sed 's/^X//' > '/local/ast/minix/tape3/commands/date.c' << '+ END-OF-FILE ''/local/ast/minix/tape3/commands/date.c' X/* date - print or set time and date Author: Jan Looyen */ X X#include <stdio.h> X#include <time.h> X X#define MIN 60L /* # seconds in a minute */ X#define HOUR (60 * MIN) /* # seconds in an hour */ X#define DAY (24 * HOUR) /* # seconds in a day */ X#define YEAR (365 * DAY) /* # seconds in a year */ X Xchar *ctime(); X Xmain(argc, argv) Xint argc; Xchar **argv; X{ X int qflag; X long t, time(); X char time_buf[15]; X X if (argc > 2) usage(); X if (argc == 2) { X if (*argv[1] == '-' && (argv[1][1] | 0x60) == 'q') { X freopen(stdin, "/dev/tty0", "r"); X printf("\nPlease enter date: MMDDYYhhmmss. Then hit RETURN.\n"); X gets(time_buf); X set_time(time_buf); X } X else X set_time(argv[1]); X } X time(&t); X printf("%s", ctime(&t)); X exit(0); X} X X Xset_time(t) Xchar *t; X{ X char *tp; X long ct, time(); X int len; X static int days_per_month[] = { X 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 X }; X struct tm *p, *localtime(); X X time(&ct); X p = localtime(&ct); X p->tm_year -= 1970; X p->tm_mon++; X len = strlen(t); X if (len != 12 && len != 10 && len != 6 && len != 4) usage(); X tp = t; X while (*tp) X if (!isdigit(*tp++)) X bad(); X if (len == 6 || len == 12) X p->tm_sec = conv(&tp, 59); X p->tm_min = conv(&tp, 59); X p->tm_hour = conv(&tp, 23); X if (len == 12 || len == 10) { X p->tm_year = conv(&tp, 99); X p->tm_mday = conv(&tp, 31); X p->tm_mon = conv(&tp, 12); X p->tm_year -= 70; X if (p->tm_year < 0) X p->tm_year += 100; X } X ct = p->tm_year * YEAR; X ct += ((p->tm_year + 1) / 4) * DAY; X days_per_month[1] = 28; X if (((p->tm_year + 2) % 4) == 0) X days_per_month[1]++; X len = 0; X p->tm_mon--; X while (len < p->tm_mon) X ct += days_per_month[len++] * DAY; X ct += --p->tm_mday * DAY; X ct += p->tm_hour * HOUR; X ct += p->tm_min * MIN; X ct += p->tm_sec; X if (stime(&ct)) X fprintf(stderr, "Set date not allowed\n"); X} X Xconv(ptr, max) Xchar **ptr; Xint max; X{ X int buf; X X *ptr -=2; X buf = atoi(*ptr); X **ptr = 0; X if (buf < 0 || buf > max) X bad(); X return(buf); X} X Xbad() X{ X fprintf(stderr, "Date: bad conversion\n"); X exit(1); X} X Xusage() X{ X fprintf(stderr, "Usage: date [-q] [[MMDDYY]hhmm[ss]]\n"); X exit(1); X} X Xisdigit(c) Xchar c; X{ X if (c >= '0' && c <= '9') X return(1); X else X return(0); X} + END-OF-FILE /local/ast/minix/tape3/commands/date.c chmod 'u=r,g=r,o=r' '/local/ast/minix/tape3/commands/date.c' set `wc -c '/local/ast/minix/tape3/commands/date.c'` count=$1 case $count in 2223) :;; *) echo 'Bad character count in ''/local/ast/minix/tape3/commands/date.c' >&2 echo 'Count should be 2223' >&2 esac echo Extracting '/local/ast/minix/tape3/lib/ctime.c' sed 's/^X//' > '/local/ast/minix/tape3/lib/ctime.c' << '+ END-OF-FILE ''/local/ast/minix/tape3/lib/ctime.c' X#include <time.h> X Xstatic int days_per_month[] = { X 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 X}; Xstatic char *months[] = { X "Jan", "Feb", "Mar", "Apr", "May", "Jun", X "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" X}; Xstatic char *days[] = { X "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" X}; X X#define MIN 60L /* # seconds in a minute */ X#define HOUR (60 * MIN) /* # seconds in an hour */ X#define DAY (24 * HOUR) /* # seconds in a day */ X#define YEAR (365 * DAY) /* # seconds in a year */ X Xstatic struct tm tm; Xstatic char buf[26]; X Xchar *ctime(pt) Xlong *pt; X{ X register long t = *pt; X X long year; X X tm.tm_year = 0; X tm.tm_mon = 0; X tm.tm_mday = 1; X tm.tm_hour = 0; X tm.tm_min = 0; X tm.tm_sec = 0; X X /* t is elapsed time in seconds since Jan 1, 1970. */ X tm.tm_wday = (int) (t/DAY + 4L) % 7; /* Jan 1, 1970 is 4th wday */ X while (t >= (year=((tm.tm_year%4)==2) ? YEAR+DAY : YEAR)) { X tm.tm_year += 1; X t -= year; X } X tm.tm_year += 1970; X X /* t is now the offset into the current year, in seconds. */ X tm.tm_yday = (t/DAY); /* day # of the year, Jan 1 = 0 */ X X days_per_month[1] = 28; X if ((tm.tm_year % 4) == 0) /* check for leap year */ X days_per_month[1]++; X X /* Compute month. */ X while (t >= (days_per_month[tm.tm_mon] * DAY)) X t -= days_per_month[tm.tm_mon++] * DAY; X X /* Month established, now compute day of the month */ X while (t >= DAY) { X t -= DAY; X tm.tm_mday++; X } X X /* Day established, now do hour. */ X while (t >= HOUR) { X t -= HOUR; X tm.tm_hour++; X } X X /* Hour established, now do minute. */ X while (t >= MIN) { X t -= MIN; X tm.tm_min++; X } X X /* Residual time is # seconds. */ X tm.tm_sec = (int) t; X X /* Generate output in ASCII in buf. */ X sprintf(buf, "%s %s %2d %02d:%02d:%02d %d\n", X days[tm.tm_wday], months[tm.tm_mon], X tm.tm_mday, tm.tm_hour, tm.tm_min, tm.tm_sec, tm.tm_year); X return buf; X} X Xstruct tm *localtime(pt) Xlong *pt; X{ X ctime(pt); X return &tm; X} X Xstruct tm *gmtime(pt) Xlong *pt; X{ X ctime(pt); X return &tm; X} + END-OF-FILE /local/ast/minix/tape3/lib/ctime.c chmod 'u=r,g=r,o=r' '/local/ast/minix/tape3/lib/ctime.c' set `wc -c '/local/ast/minix/tape3/lib/ctime.c'` count=$1 case $count in 1960) :;; *) echo 'Bad character count in ''/local/ast/minix/tape3/lib/ctime.c' >&2 echo 'Count should be 1960' >&2 esac exit 0
ast@cs.vu.nl (Andy Tanenbaum) (01/04/89)
In article <321@cstw01.UUCP> meulenbr@cst.UUCP () writes: >Happy New Year with much minix! Indeed. On my system this was message number 4000. I am not sure if we have gotten all of them, but 4000 messages from all over the world in < 2 years is a milestone of sorts. Unlike many newsgroups, I regard the signal to noise ratio of this group as being quite good. If we can just keep the PDP-11/45 memory management unit from rearing its ugly head in 1989, we'll be in good shape. Andy Tanenbaum (ast@cs.vu.nl)
ncoverby@ndsuvax.UUCP (Glen Overby) (01/04/89)
In article <324@cst.UUCP> meulenbr@cst.UUCP (Frans Meulenbroeks) writes: >Oops I goofed. As I said I sent an untested shar. Well I learned to do >this never again. I hope so. Getting untested bug fixes scares me; there is no way that you can truly say that the purported bugs are fixed. Or that the "fix" didn't break something else. So far I haven't noticed any real problems with this. We've been lucky. Another thing I find lacking in this group which I enjoy from comp.sources.bugs and comp.bugs.{4bsd,sysv,...} is an explanation of what the bug is and how to duplicate it. I like to know WHAT is wrong so I can be on the lookout for any other problems like it. >>I got also a bug report that cc would not work properly on this day. >>I have not been able to verify this. The person who told me about this, >>said that it happened on an ST. I'm unsure if this happens on a PC as >>well. >I could repeat this. /usr/lib/cem crashes with signal 11. >This only happens on 12/31 on leap years. No explanation. No fix, since >I do not have source from cem. Perhaps the same problem as above. >Perhaps this problem only exist in Minix/ST. I burned a LOT of CPU time on the 31st compiling under Minix-PC ("1.3d" with 1.2 compiler binaries) and did not encounter any compiler crashes. Not all of the output code worked, but that was my fault :-) >>Something unrelated: >>Is someone aware of a PD SCCS/RCS clone/alternative? >>Any pointers would be appreciated greatly. Why an RCS clone? You can get the real thing from arthur.cs.purdue.edu via anonymous FTP, and GNU has a diff3. I haven't tried putting the two together under Minix (I keep all my RCS stuff on a VAX). Glen Overby ncoverby@plains.nodak.edu uunet!ndsuvax!ncoverby ncoverby@ndsuvax (Bitnet)