Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!samsung!zaphod.mps.ohio-state.edu!rpi!bu.edu!nntp-read!composer From: composer@bucsf.bu.edu (Jeff Kellem) Newsgroups: comp.lang.perl Subject: io.dup test also fails on Encore Multimax, UMAX 4.2 Message-ID: Date: 11 Aug 90 20:46:51 GMT References: Sender: news@bu.edu.bu.edu Reply-To: composer@cs.bu.edu Followup-To: comp.lang.perl Distribution: comp Organization: Boston University Computer Science Department Lines: 44 In-reply-to: nozoe@telcom.oki.CO.JP's message of 11 Aug 90 09:47:39 GMT In article nozoe@telcom.oki.CO.JP (Katsumi Nozoe) writes: | Subject: Re: perl-3.0 pl#27 "make test" NG on VAX Ultirx V3.0 | | Machine : VAX 8810 | OS : Ultirx-32 V3.0 (Rev 64) | Compiler: "gcc-1.37.9" and "cc" | | nozoe@sh8810% make test | chmod +x t/TEST t/base.* t/comp.* t/cmd.* t/io.* t/op.*; \ | cd t && (rm -f perl; ln -s ../perl .) && ./perl TEST [ ...text deleted... ] | io.argv.........ok | io.dup..........FAILED on test 2 | io.fs...........ok [ ...text deleted... ] The `io.dup' test also fails on an Encore Multimax running UMAX 4.2 Release 3.3. I've tried compiling with gcc 1.37.1 and cc; and, tried both bison 1.11 and yacc. None have been able to produce a perl binary that could execute `io.dup' properly. The output from `io.dup' is: 1..6 ok 1 ok 6 It seems that STDOUT and STDERR aren't being reset properly, or some such thing, right before trying to cat the file "Io.dup". The "Io.dup" file has the correct data in it. Any clues, Larry? Note that things do work properly on a Sun 3 running SunOS 4.0.3, using gcc 1.37.1 or cc, and bison 1.11. The `io.dup' test also works properly on the Multimax with perl at version 3.0 patchlevel 18. Good luck! -jeff Jeff Kellem INTERNET: composer@cs.bu.edu (or composer@bu.edu) UUCP: ...!harvard!bu-cs!composer