On Fri, 1 Sep 2000, Craig A. Berry wrote:
> At 01:48 PM 9/1/2000 -0700, Peter Prymmer wrote:
>
> >Perhaps we should add this:
> >
> > $cc_cmd =~ s/\/prefix=all//i;
> >
> >before trying to figure out the gcc version? I'll see if I can test that.
>
> I was just giving a for instance; I don't know the exact qualifiers used and
> they should all be removed when just getting the version. I don't even know
> if gcc has a /version qualifier, BTW.
I didn't think that it did either, but I wondered if it had a dual
unix (-flag) vs. vms (/qualifier) nature like vmstar. It should certainly
all for:
gcc --version
to determine the version. Unfortunately I have no working gcc
installation on VMS here.
> $cc_version_cmd = $cc_cmd;
> $cc_version_cmd =~ s/\/?.*//i;
> $cc_version_cmd .= '/VERSION';
>
> [off the cuff and untested]
OK I'll see if I can test something like that (we are getting really close
to the deadline, I'd really like to test a non default prefix mms install
too :-)
> >> >The regression tests still skip lines:
>
> >I was using DECC sockets on a UCX (er, TCP/IP Services), not Multinet
> >(although I could).
>
> What OS version. I'm currently testing on 7.2-1 and I think the last time I
> saw this was on a 7.1 machine with a rather out-of-date C RTL.
Yep. VMS V7.1 with anal/image sys$library:decc$crtl.exe turns up:
image format major id: 3, minor id: 0
Image Identification Information
image name: "DECC$CRTLMAP"
image file identification: "V4.5"
image file build identification: ""
link date/time: 13-MAY-1998 03:43:37.61
linker identification: "A11-39"
The problem with the regression tests did not appear on this
machine with any of the bleadperl kits until a couple of days ago.
Basically after the fwrite() and vmspipe patches were put in.
Like I said previously I can rule out the fwrite patch as the cause,
I am not sure about the vmspipe patch though.
Peter Prymmer
