Peter -
I cannot say why I had problems with perlshr defined, but as soon as I
deassign'd it, most of the errors went away. The errors (which I helpfully did
not save) referred to problems activating the shared image PERLSHR.EXE.
I just tested the glob-basic.t patch on Win32 and it skipped the test (as it
should, apparently). I suspect that the test was originally for just one case
and someone hacked it quickly and blew the boolean logic. On any platform
except Win32, the first half of the test is always true and the VMS half is
never tested because of the ||. I tend to prefer eq over ne because then the
sense of the || works better for me.
I don't know grep well enough to understand what Test 8 is doing; I suspect it
is looking for a file called TES* but why I only see two values @a = ('a', 'b').
Is there supposed to be a file called TEST. in the directory someplace? That
did not seem to help (I created on in the t directory).
My SYS$TIMEZONE_DIFFERENTIAL was set to -14400 for some unknown reason; that
appears to be the Atlantic Timezone, not Eastern US. Can't find what script
that is in just yet, but I'll track it down eventually. That fixed all of the
problems in VMSISH.T. Thanks
John Peacock
____________________Reply Separator____________________
Subject: Re: [ID 20000330.044] Not OK: perl v5.6.0 on VMS_AXP V7.1
Author: <[EMAIL PROTECTED]>
Date: 3/30/00 5:30 PM
In <[EMAIL PROTECTED]> John Peacock wrote Thu, 30 Mar 2000:
> Overall a clean build under VMS (WOW!!!). If perlshr is already
> defined (i.e. an existing installation), there are failures in
> [.t.lib]vmsish.t and [.t.pragma]warnings.t (Configure.com should
> check for this and suggest that perlshr be deassigned).
Some folks have reported clean builds and tests with a pre-extant
PERL_ROOT and PERLSHR logical. In general the test suite is quite
careful about avoinding dependencies on PERL_ROOT and PERLSHR to
avoid such upgrade problems.
Is PERL_ROOT=DRA0:[PERL5.] the new or old installation location?
> Once that is done:
>
> Failed two tests in [.t.lib]glob-basic.t:
>
> Test 3 - Patch below (sorry for the lame VMS diff)
>
> File PERL_SRC:[PERL-5_6_0.T.LIB]GLOB-BASIC.T;1
> 42 if ($^O ne 'MSWin32' || $^O ne 'VMS') {
> 43 eval {
> ******
> File PERL_SRC:[PERL-5_6_0.T.LIB]GLOB-BASIC.T;2
> 42 if ($^O eq 'MSWin32' || $^O eq 'VMS') {
> 43 print "ok 3 # skipped\n";
> 44 }
> 45 else {
> 46 eval {
> ************
> ************
> File PERL_SRC:[PERL-5_6_0.T.LIB]GLOB-BASIC.T;1
> 52 }
> 53 print "ok 3\n";
> 54
> ******
> File PERL_SRC:[PERL-5_6_0.T.LIB]GLOB-BASIC.T;2
> 55 print "ok 3\n";
> 56 }
> 57
> ************
Thanks (did you test on Win32?).
> Test 8 - I do not understand what it is doing but @a is
> being set to ('a','a') so it fails all tests.
For me on one Alpha test 7 fails when run under `MMS test`
but they all pass (including 7) when I type:
@perl_setup
perl t/lib/vmsish.t
> Failed Tests 14-16 in [.t.lib]vmsish.t (no patch supplied)
>
> not ok 14 # (time) UTC: 954450111 VMS: 954432111
> not ok 15 # (localtime)
> # UTC: 51 1 11 30 2 100 4 89 0
> # VMS: 51 1 16 30 2 100 4 89 0
> not ok 16 # (gmtime)
> # UTC: 51 1 16 30 2 100 4 89 0
> # VMS: 51 1 21 30 2 100 4 89 0
What does
$ show logical sys$timezone_differential
tell you about your system's UTC TDF? I am in the PST so I have:
$ sho log sys$timezone_differential
"SYS$TIMEZONE_DIFFERENTIAL" = "-28800" (LNM$SYSTEM_TABLE)
Peter Prymmer
