Thanks to all that replied. PGFLQ and WSEXTENT/WSMAX are the critical combination. I'm not sure what the mins are, but I set PGFLQ=100000 and WSEXTENT=40960 and it compiled.
During the test phase, I had one failure -
[.lib]vmsfspec..........FAILED on test 54
Is this a "Bad Thing (tm)"?
- Paul
>-----Original Message-----
>From: Peter Prymmer [mailto:[EMAIL PROTECTED]]
>Sent: Tuesday, January 30, 2001 4:52 PM
>To: [EMAIL PROTECTED]
>Cc: [EMAIL PROTECTED]
>Subject: Re: Error while building Perl 5.6.0
>
>
>
>On Tue, 30 Jan 2001 [EMAIL PROTECTED] wrote:
>
>> I decided today to try building 5.6.0 on my VAX (VMS 6.2,
>DECC 5.7). I
>> pulled latest.tar.gz and applied the patches from
>www.crinoid.com. I then
>> ran CONFIGURE.COM. So far, so good. I then started the
>build (using MMK)
>> and it began compiling. It only got, however, as far as
>OP.C before it
>> stopped with a "%CLI-F-TEXT, Compiler abort - virtual memory limits
>> exceeded." I tried it on another VAX (also VMS 6.2 and DECC 5.7) and
>> encountered the same error at VMS.C. Anyone know the
>process quota minimums
>> to compile Perl?
>
>The big one that used to catch a lot of people off guard was the PGFLQ.
>
>The version of README.vms in the development branch reads in part:
>
> Be sure that the process that you use to build perl has a
>PGFLQ greater
> than 100000. Be sure to have a correct local time zone to UTC offset
> defined (in seconds) in the logical name
>SYS$TIMEZONE_DIFFERENTIAL before
> running the regression test suite.
>
>but I am not sure that the 5.6.0 kit has that advice (it used to be in
>perl 5.002 or so a long time ago). As it turns out you can
>get away with
>less than 100000 (the AUTHORIZE default for any account is
>50000 IIRC) and
>I posted some test results with VMS I7.3 that are in the
>vmsperl mailing
>archives from this past fall.
>
>Peter Prymmer
>
>
