RE: ext/POSIX/t/posix.t on VMS

2001-11-07 Thread Craig A. Berry
At 11:56 AM 11/7/2001 -1000, Tim Jenness wrote: >> not ok 28 # POSIX::errno(): 20, $!: 0 >> >> which I think means autoloading sets errno but isn't supposed to. I suspect >> it's doing a -d somewhere and leaving an errno of ENOTDIR laying about. Hmm. >> > >Well, $! is meant to be exactly er

Re: Dusting out vms/test.com

2001-11-07 Thread PPrymmer
'user1' looks like a volume logical name not a physical device name. configure.com ought to have f$parse()ed out the physical device name. Peter Prymmer "Craig A

Re: Dusting out vms/test.com

2001-11-07 Thread PPrymmer
Do not bother trying to redefine the perlshr logical name. The only one that you need to redefine is PERL_ROOT since PERLSHR is defined in terms of PERL_ROOT. Try this: $ @perl_setup $ show default $ show logical PERL_ROOT $ define/trans=conc PERL_ROOT :[.] Where you will need to deter

Re: Dusting out vms/test.com

2001-11-07 Thread PPrymmer
Type LOGOFF at the command prompt. You can lowercase commands and abbreviate most as well hence: $ lo will generally get you off. N.B. that "exit" will simply stop whatever procedure or program that you ^Y'ed out of. It won't get you off of the machine. "lo" will under most circumstance

Re: Dusting out vms/test.com

2001-11-07 Thread PPrymmer
>From the command line: $ show logical PERLSHR as well as: $ show logical PERL_ROOT will be the (rough) equivalents of: % echo $PATH on Unix - at least as far as PERL is concerned. Craig's suggestion of re-defining your PERL_ROOT to point to your build directory is well worth following in

RE: ext/POSIX/t/posix.t on VMS

2001-11-07 Thread Tim Jenness
> - my $foo = $!; > + my $foo = 0 + $!; # force numeric > my $errno = POSIX::errno(); > print "not " unless $errno == $foo; > - print "ok ", 28 + $test, " # POSIX::errno(): $errno, \$!: $!\n"; > + print "ok ", 28 + $test, " # POSIX::errno(): $errno, \$!:

RE: ext/POSIX/t/posix.t on VMS

2001-11-07 Thread Craig A. Berry
At 09:53 PM 11/7/2001 +0100, Tels wrote: >So, if I read Tim correctly, make that my $foo = $! + 0; and it might >pass the test. Needs probably a bit more munging to print the $! in >numeric context as well as $foo after a failure. Thanks Tels and Tim. With the following additional change: ---

[PATCH pod/perlhack.pod] When to use what test libraries

2001-11-07 Thread Michael G Schwern
This patch adds a basic explanation of what testing bits can be used in what tests. Taking it's own advice, it modernizes the patch example to use t/test.pl and is(). --- pod/perlhack.pod2001/11/07 20:32:08 1.1 +++ pod/perlhack.pod2001/11/07 20:50:18 @@ -1531,47 +1531,42 @@ Unicode

Re: ext/POSIX/t/posix.t on VMS

2001-11-07 Thread Tim Jenness
On Wed, 7 Nov 2001, Craig A. Berry wrote: > I've made some progress getting this test to run (see working patch below) > but I still get one failure: > > not ok 28 # POSIX::errno(): 20, $!: not a directory > > 20 is the correct value for ENOTDIR, but how can POSIX::errno() and $! be > expecte

ext/POSIX/t/posix.t on VMS

2001-11-07 Thread Craig A. Berry
I've made some progress getting this test to run (see working patch below) but I still get one failure: not ok 28 # POSIX::errno(): 20, $!: not a directory 20 is the correct value for ENOTDIR, but how can POSIX::errno() and $! be expected to be numerically equal when $! returns a string? Or i