Carl Friedberg wrote:
lib/vmsish.t:
1..25


not ok 25 - (stat) UTC: 1064956297  VMS: 1064941897
LIB/VMSISH....FAILED at test 25
Failed 1 test script out of 1, 0.00% okay.

  "SYS$TIMEZONE_DAYLIGHT_SAVING" = "0"
  "SYS$TIMEZONE_DIFFERENTIAL" = "-18000"
  "SYS$TIMEZONE_NAME" = "EST"
  "SYS$TIMEZONE_RULE" = "EST5EDT4,M4.1.0/02,M10.4.0/02"

Yeah, me too, though under 7.3-1. If you go back and test Perl 5.8.1 again you'll find the same thing, which means this worked during daylight savings time (when 5.8.1 was released) but doesn't work now. I think we have one or more genuine bugs in our timezone handling. My suspicion is that Perl compensates for things the C RTL didn't always do, but at least in some cases now does.

For example, from VMS712_ACRTL-v0200:

o Incorrect timestamp returned from stat and fstat functions

         The stat function no longer puts the wrong values into st_?time
         members of the stat structure.  More precisely, during daylight
         savings time the values of st_?time  are  no  longer  one  hour
         ahead  from  file  creation/modification  time  reported by the
         $DIRECTORY command.

Which means that when Perl adds or subtracts an hour based on the DST
setting, we'd need to know we whether we had a CRTL version that needed
extra adjustment. Ick. I think Richard Brodie was looking into this
stuff, but I can't remember if he concluded anything.



Reply via email to