Article 13355 of comp.lang.perl: Newsgroups: comp.lang.perl Path: netlabs!news.cerf.net!usc!cs.utexas.edu!uunet!fmrco!fmrco!asherman From: asherman@fmrco.com (Aaron Sherman) Subject: Re: perl 5a2 cannot "die" (plus current list o' bugs) In-Reply-To: wjm@feenix.metronet.com's message of Fri, 20 Aug 1993 21:32:10 GMT Message-ID: Sender: news@fmrco.uucp X-Quote: "...accepting is much harder than giving." -Mike Smith Reply-To: asherman@fmrco.COM Organization: I-Kinetics, 19 Bishop-Allen Dr., Cambridge, MA References: Date: Mon, 23 Aug 1993 14:42:50 GMT Lines: 47 >>>>> wjm@feenix.metronet.com (Bill Middleton) said: wjm> asherman@fmrco.COM writes: >An interesting pair of bugs can be seen in the following output: wjm> I dont think so. Could be in the compilation or something. Did it wjm> pass all tests? Each of the following work fine here on this HP. I tried compiling with Sun's native CC and GCC. Both worked fine, but caused this problem. I'll try it with Larry's original version when I get a chance. wjm> perl5 -e 'die "hello $. \n"; wjm> hello Ah. But, note that the $. STILL isn't working. So only ONE of those bugs did not show. This is my current list of bugs (not complete, but what I've had time to note). Hope it helps: "perl -e die" will cause a seg-fault $. is not updated Memory leak for anonymous arrays: while(1) { @a = (1, 2, 3, [4, 5], 6); @a = (); } Will keep allocating and not freeing memory. "perl -e 'sub foo {print 1} foo'" should either complain or call foo, but does neither. Or, did Larry not impliment the &-less function calling that he was talking about? "perl -le 'sub foo {1} $a = \&foo; print &{$a}; print &{$a} + 1'" should not fail to parse. -AJS -- Aaron Sherman I-Kinetics, Inc. Systems Engineer "Open Systems Stepstones" Voice: (617)661-8181 19 Bishop Allen Dr. Fax: (617)661-8625 Cambridge, MA 02139 Pager: (508)545-0584 asherman@i-kinetics.com