From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Delivered-To: caml-list@yquem.inria.fr Received: from concorde.inria.fr (concorde.inria.fr [192.93.2.39]) by yquem.inria.fr (Postfix) with ESMTP id 4776FBCAE for ; Fri, 24 Jun 2005 18:54:01 +0200 (CEST) Received: from pauillac.inria.fr (pauillac.inria.fr [128.93.11.35]) by concorde.inria.fr (8.13.0/8.13.0) with ESMTP id j5OGs0el019255 for ; Fri, 24 Jun 2005 18:54:00 +0200 Received: from concorde.inria.fr (concorde.inria.fr [192.93.2.39]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id SAA23368 for ; Fri, 24 Jun 2005 18:54:00 +0200 (MET DST) Received: from comtv.ru (mail.comtv.ru [217.10.32.19]) by concorde.inria.fr (8.13.0/8.13.0) with ESMTP id j5OGrxNL019252 for ; Fri, 24 Jun 2005 18:54:00 +0200 X-UCL: actv Received: from av1474.oops ([10.0.66.9] verified) by comtv.ru (CommuniGate Pro SMTP 4.1.8) with ESMTP id 65147667; Fri, 24 Jun 2005 20:53:49 +0400 Date: Fri, 24 Jun 2005 20:54:25 +0400 (MSD) From: malc X-X-Sender: malc@home.oyster.ru To: yminsky@cs.cornell.edu Cc: Caml Mailing List Subject: Re: [Caml-list] Why are backtraces (sometimes) useless In-Reply-To: <891bd339050624093978a3c9a8@mail.gmail.com> Message-ID: References: <891bd339050624093978a3c9a8@mail.gmail.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed X-Miltered: at concorde with ID 42BC3AA8.000 by Joe's j-chkmail (http://j-chkmail.ensmp.fr)! X-Miltered: at concorde with ID 42BC3AA7.001 by Joe's j-chkmail (http://j-chkmail.ensmp.fr)! X-Spam: no; 0.00; malc:01 malc:01 pulsesoft:01 caml-list:01 yaron:01 minsky:01 ocaml:01 threads:01 stack:01 surprising:01 native-code:01 backtrace:01 backtrace:01 pulsesoft:01 2005,:98 X-Spam-Checker-Version: SpamAssassin 3.0.2 (2004-11-16) on yquem.inria.fr X-Spam-Status: No, score=0.1 required=5.0 tests=FORGED_RCVD_HELO autolearn=disabled version=3.0.2 X-Spam-Level: On Fri, 24 Jun 2005, Yaron Minsky wrote: > I've had a lot of experience lately with OCaml backtraces sucking, and I'm > not sure what's going on. We're working on a decent-sizes project (about > 10-kloc) that uses threads and a decent amount of functorization. One thing > we've noticed is that backtraces are almost completely useless --- generally > speaking, they have lots of "Called from unknown location" messages, and > almost no useful data. They often skip over functions on the call stack of > the error in surprising ways. > > I guess the question is this: what kind of situations cause backtraces to > drop information in the ways described above? > > Also, is there anyone out there actively using the native-code backtrace > patches that people have come up with? I'm curious to hear other people's > experiences. Native code backtracing i did is constrained by the fact that only functions that contain gc call site are included into backtrace. I even included an example which showcases this limitation. -- mailto:malc@pulsesoft.com