From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 21433 invoked by alias); 18 Dec 2007 15:54:42 -0000 Received: (qmail 21425 invoked by uid 22791); 18 Dec 2007 15:54:41 -0000 X-Spam-Check-By: sourceware.org Received: from mail.codesourcery.com (HELO mail.codesourcery.com) (65.74.133.4) by sourceware.org (qpsmtpd/0.31) with ESMTP; Tue, 18 Dec 2007 15:54:31 +0000 Received: (qmail 26685 invoked from network); 18 Dec 2007 15:54:29 -0000 Received: from unknown (HELO digraph.polyomino.org.uk) (joseph@127.0.0.2) by mail.codesourcery.com with ESMTPA; 18 Dec 2007 15:54:29 -0000 Received: from jsm28 (helo=localhost) by digraph.polyomino.org.uk with local-esmtp (Exim 4.68) (envelope-from ) id 1J4em4-0000HV-N8; Tue, 18 Dec 2007 15:54:28 +0000 Date: Tue, 18 Dec 2007 16:30:00 -0000 From: "Joseph S. Myers" To: Kai Tietz cc: GCC Patches , NightStrike Subject: Re: Ping - old patch from April - mingw support for I32/I64 MS printf formatters to c-format.c In-Reply-To: Message-ID: References: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2007-12/txt/msg00885.txt.bz2 On Tue, 18 Dec 2007, Kai Tietz wrote: > I added updated the patch against current head version of gcc and add a > test-case for the common case that the system (crt) version of formatter > is used. For mingw targets there is no %lld formatter existing. Instead MS > functions are using %I64d and %I64d. This patch additionally allows the > user to define a alternative formatter beside the system one for i386 > targets. The default formatters are named "gnu_*" and the MS specific > "ms_*". > > This patch is tested for *-pc-mingw32 (using MS %I32 and %I64) and for > i686-pc-cygwin (for the gnu variant). The new test needs to go in gcc.dg/format, and where possible tests should use format.h. In addition to testing the new Windows format, you need to test that all the standard and GNU features not supported by Windows receive proper diagnostics. Did you really get clean test results for all the existing testcases in gcc.dg/format on MinGW? Since they use system functions such as printf, I'd have expected many failures as they test features not supported by Windows. This would require appropriate arrangements for the functions to be declared with the gnu_* attributes on Windows, or to be mapped with #define to functions so declared (as the minimal change to keep the tests working) for most tests, with separate Windows-only tests that appropriate features are accepted and rejected for the system and ms_ formats. Does the set of format features supported on Windows depend at all on the version of Windows? If so, for what version of Windows did you prepare the tables of supported features? -- Joseph S. Myers joseph@codesourcery.com