From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dan.rpsys.net (5751f4a1.skybroadband.com [87.81.244.161]) by mail.openembedded.org (Postfix) with ESMTP id 4AC4D605D2 for ; Fri, 17 Apr 2015 07:15:55 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by dan.rpsys.net (8.14.4/8.14.4/Debian-4.1ubuntu1) with ESMTP id t3H7FqrR028776; Fri, 17 Apr 2015 08:15:52 +0100 Received: from dan.rpsys.net ([127.0.0.1]) by localhost (dan.rpsys.net [127.0.0.1]) (amavisd-new, port 10024) with LMTP id o-4mKHwtfZtM; Fri, 17 Apr 2015 08:15:52 +0100 (BST) Received: from [192.168.3.10] ([192.168.3.10]) (authenticated bits=0) by dan.rpsys.net (8.14.4/8.14.4/Debian-4.1ubuntu1) with ESMTP id t3H7FdYL028770 (version=TLSv1/SSLv3 cipher=AES128-GCM-SHA256 bits=128 verify=NOT); Fri, 17 Apr 2015 08:15:50 +0100 Message-ID: <1429254936.6976.191.camel@linuxfoundation.org> From: Richard Purdie To: Khem Raj Date: Fri, 17 Apr 2015 08:15:36 +0100 In-Reply-To: References: <20150415100850.GA2338@jama> X-Mailer: Evolution 3.12.10-0ubuntu1~14.10.1 Mime-Version: 1.0 Cc: OE Core mailing list Subject: Re: autom4te segfault in Dumper.so when configuring intltool X-BeenThere: openembedded-core@lists.openembedded.org X-Mailman-Version: 2.1.12 Precedence: list List-Id: Patches and discussions about the oe-core layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 17 Apr 2015 07:15:59 -0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit On Thu, 2015-04-16 at 22:09 -0400, Khem Raj wrote: > On Wed, Apr 15, 2015 at 5:37 PM, Andre McCurdy wrote: > > > > Hiding OE sysroot perl when configuring intltool and using only host > > perl does also prevent the segfault. > > > > I'm not sure what the correct fix is though. Perhaps have the > > autoconf-native scripts run host perl via a wrapper which removes any > > reference to OE sysroot perl from the PATH? > > Not sure why Richard suggested that but may be his intention was to > build automake-native early enough since perl takes a bit of time to > build and it would serialize the build but we dont know unless we > measure, may be there are enough native tools to build to keep > processors spinning while its building perl-native > in theory correct fix is to add dependency on perl native for autoconf > native and do the necessary surgeries (if required to use perl-native) autoconf-native is near enough one of the primary pieces of the dependency chain. Having it depend on perl-native is a very very bad idea, if its possible at all. It might be "correct" in principle but in practise its not workable. Cheers, Richard