From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff King Subject: Re: git-cvsexportcommit fails for huge commits Date: Wed, 12 Dec 2007 23:17:37 -0500 Message-ID: <20071213041737.GA23624@coredump.intra.peff.net> References: <20071211200418.GA13815@mkl-desktop> <20071212083154.GB7676@coredump.intra.peff.net> <46a038f90712121158n674a9044t75ef99473314457c@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Markus Klinik , git@vger.kernel.org To: Martin Langhoff X-From: git-owner@vger.kernel.org Thu Dec 13 05:18:04 2007 Return-path: Envelope-to: gcvg-git-2@gmane.org Received: from vger.kernel.org ([209.132.176.167]) by lo.gmane.org with esmtp (Exim 4.50) id 1J2fWN-0005E4-Oe for gcvg-git-2@gmane.org; Thu, 13 Dec 2007 05:18:04 +0100 Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752173AbXLMERl (ORCPT ); Wed, 12 Dec 2007 23:17:41 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752442AbXLMERl (ORCPT ); Wed, 12 Dec 2007 23:17:41 -0500 Received: from 66-23-211-5.clients.speedfactory.net ([66.23.211.5]:2704 "EHLO peff.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751692AbXLMERk (ORCPT ); Wed, 12 Dec 2007 23:17:40 -0500 Received: (qmail 7099 invoked by uid 111); 13 Dec 2007 04:17:38 -0000 Received: from coredump.intra.peff.net (HELO coredump.intra.peff.net) (10.0.0.2) by peff.net (qpsmtpd/0.32) with SMTP; Wed, 12 Dec 2007 23:17:38 -0500 Received: by coredump.intra.peff.net (sSMTP sendmail emulation); Wed, 12 Dec 2007 23:17:37 -0500 Content-Disposition: inline In-Reply-To: <46a038f90712121158n674a9044t75ef99473314457c@mail.gmail.com> Sender: git-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: git@vger.kernel.org Archived-At: On Thu, Dec 13, 2007 at 08:58:33AM +1300, Martin Langhoff wrote: > IOWs, the protocol *is* atomic, and this patch does make things > slightly more brittle. Perhaps require an option to be set before we > do this? I started writing a patch to let the user specify the limit, but it was just too ugly. What user knows the right limit? We are probably better off just setting the limit at something high and reasonable (like 64K -- it would be nice to get feedback from Markus on what platform he is using and what is a reasonable value), or just using a tempfile with xargs, which should figure out the correct value. -Peff