From shiva@qued.com Wed Jun 30 04:38:11 1999 Received: from locus.qued.com ([216.101.181.7]) by spog.gaertner.de (8.8.8/8.8.8/Nase) with ESMTP id EAA01593; Wed, 30 Jun 1999 04:38:09 +0200 Received: from mochas (vpn-mail [216.101.181.251]) by locus.qued.com (8.8.7/8.8.7) with SMTP id TAA20287; Tue, 29 Jun 1999 19:37:55 -0700 Message-Id: <002e01bec2a2$059b31e0$74273b0a@mochas.qued.com> Date: Tue, 29 Jun 1999 19:40:27 -0700 From: "miles zarathustra" <shiva@qued.com> To: "miles zarathustra" <shiva@qued.com>, "Martin Neitzel" <neitzel@gaertner.de>, <jkt@qued.com> Cc: <bugs@spog.gaertner.de>, "Eric Iverson" <eiverson@interlog.com> Subject: memr >> end of j session >Number: 31 >Category: j-engine >Synopsis: memr >> end of j session >Confidential: no >Severity: serious >Priority: low >Responsible: neitzel >State: closed >Class: mistaken >Submitter-Id: net >Arrival-Date: Wed Jun 30 04:52:01 MET DST 1999 >Last-Modified: Fri May 30 19:35:30 MET DST 2003 >Originator: miles zarathustra >Organization: >Release: >Environment: >Description: I hope I'm sending this to the right address for bug reports. I had a mapped file, but when I unmapped it and tried an memr from it it ended my J session. I don't know if you're planning to catch these segv's, but I thought I'd send it anyway. The script file is attached that defines c_mmap, &c. it's 'in progress.' [elided, neitzel] >How-To-Repeat: 0!:0<'mmap.ijs' c_open 'test.txt';RDWR +-+--------+-+ |3|test.txt|2| +-+--------+-+ c_mmap (<0);100;3;2;3;0 +----------+---+---+-+-+-+-+ |1073790976|+-+|100|3|2|3|0| | ||0|| | | | | | | |+-+| | | | | | +----------+---+---+-+-+-+-+ addr=.1073790976 memr addr,0,100 this is a test NB. worked fine until I unmapped it c_munmap (<addr);100 +-+------------+---+ |0|+----------+|100| | ||1073790976|| | | |+----------+| | +-+------------+---+ $memr addr,0,100 [locus-~/j/mmap] # end of j session >Fix: >Audit-Trail: State-Changed-From-To: open-feedback State-Changed-By: neitzel State-Changed-When: Tue Mar 14 02:47:48 MET 2000 State-Changed-Why: This is tyiding up an old bug report probably already answered before. Accessing memory after munmap(2)ing it is simply not valid. J cannot detect this violation. Handling segfaults in general is not something we would do unless we'd be 100% that J itself will never ever segfault -- which we aren't. So, no, this is not on the list of things planned for implementation. Responsible-Changed-From-To: gnats-admin->neitzel Responsible-Changed-By: neitzel Responsible-Changed-When: Tue Mar 14 02:47:48 MET 2000 Responsible-Changed-Why: [removed the PR from the "pending" queue.] State-Changed-From-To: feedback-closed State-Changed-By: neitzel State-Changed-When: Fri May 30 19:31:23 MET DST 2003 State-Changed-Why: PR database cleanup [no feedback after 5 years] >Unformatted: