Closed p5pRT closed 20 years ago
The build itself was uneventful\, though it now requires the patch to MMK (the most commonly used make utility) located here:
\<http://www.xray.mpe.mpg.de/mailing-lists/vmsperl/2000-12/msg00146.html>
There were two test failures\, given here in verbose form:
[.OP]UTF8DECODE.........1..78
ok 1
ok 2
\
[.LIB]NET-HOSTENT.......1..7 ok 1 ok 2 ok 3 ok 4 Bad arg length for Socket::inet_ntoa\, length is 58\, should be 4 at [.lib]net-hostent.t line 34. FAILED on test 5
Environment summary:
DEC C V5.2-003 on OpenVMS Alpha V7.1 Process Software MultiNet V4.2 Rev A-X\, AlphaServer 2100 5/250\, OpenVMS AXP V7.1
The build itself was uneventful\, though it now requires the patch to MMK (the most commonly used make utility) located here:
\<http://www.xray.mpe.mpg.de/mailing-lists/vmsperl/2000-12/msg00146.html>
Building of Perl really does exercise the systems...
There were two test failures\, given here in verbose form:
[.OP]UTF8DECODE.........1..78 ok 1 ok 2 \
ok 51 # 4.2.1: Malformed UTF-8 character (2 bytes\, need 1) in unpack at [.op]utf8decode.t line 142. ok 52 # 4.2.2: Malformed UTF-8 character (3 byte(null)\, need 2) in unpack at [.op]utf8decode.t line 142. 4.2.2: unpack('U*') false positive not ok 53
Curious: why does this one particular test fail in VMS\, but not in many UNIXes... the "(null)" is another oddity\, it looks like the standard case of stdio telling that it got a NULL to print as a string)\, but off-hand I can't see how that could happen (in utf8.c:utf8_to_uv()).
FAILED on test 53
[.LIB]NET-HOSTENT.......1..7 ok 1 ok 2 ok 3 ok 4 Bad arg length for Socket::inet_ntoa\, length is 58\, should be 4 at [.lib]net-hostent.t line 34. FAILED on test 5
Huh? Hasn't this been working just fine in VMS?
At 6:45 PM -0600 1/16/01\, Jarkko Hietaniemi wrote:
[.OP]UTF8DECODE.........1..78 . . . not ok 53
Curious: why does this one particular test fail in VMS\, but not in many UNIXes...
It may well be a compiler problem\, not an OS problem per se. This particular compiler (DEC C 5.2) is pretty old.
[.LIB]NET-HOSTENT.......1..7 ok 1 ok 2 ok 3 ok 4 Bad arg length for Socket::inet_ntoa\, length is 58\, should be 4 at [.lib]net-hostent.t line 34. FAILED on test 5
Huh? Hasn't this been working just fine in VMS?
Define "VMS"; there are several TCP/IP stacks\, and this test has not worked yet on Multinet as far as I know. I've been meaning to track it down\, but . . . .
I'm about to try a build with a later compiler and the Compaq-supplied TCIP/IP stack. I expect it to be less interesting.
Migrated from rt.perl.org#5212 (status was 'resolved')
Searchable as RT5212$