From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 2355 invoked by alias); 4 Jun 2010 20:11:19 -0000 Received: (qmail 2162 invoked by uid 48); 4 Jun 2010 20:10:45 -0000 Date: Fri, 04 Jun 2010 20:11:00 -0000 Message-ID: <20100604201045.2161.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug pch/14933] missing pre-compiled header depends with -MD In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "kristiaan dot lenaerts at gmail dot com" Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org X-SW-Source: 2010-06/txt/msg00528.txt.bz2 ------- Comment #13 from kristiaan dot lenaerts at gmail dot com 2010-06-04 20:10 ------- Created an attachment (id=20847) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20847&action=view) Updated fix for improper PCH dependency tracking I got hit by this bug a few weeks ago when working on a C++ project. This project has a setup for which using -fpch-deps will not work, most pch files include local headers that have to be referenced by their complete path elsewhere. I rebased the proposed fix from Mario on the 4.5 branch (I created it when 4.5 was not yet out). I'm using this patch for about a month now without any obvious problems, but I would like to get this in mainline. What are the steps that I should take? I have not yet run the testsuite. (this is on a corporate laptop running under cygwin - windows XP and full hd-encryption - so compiling gcc was painfull, and I imagine running the testsuite will be worse) Would it be ok if I run the testsuite on cygwin with and without the patch, and if the results are the same submit this patch to gcc-patches (with meaningfull changelog entry)? I could also test on a x64_64-pc-linux-gnu machine at home, but it could take a while before I can test there. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14933