Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT cygwin DOT com Mail-Followup-To: cygwin AT cygwin DOT com Delivered-To: mailing list cygwin AT cygwin DOT com Subject: Re: Annoucement: GCC-3.4.0 binary release candidat To: mingw-users AT lists DOT sourceforge DOT net, cygwin AT cygwin DOT com Cc: gp AT familiehaase DOT de, tprince AT computer DOT org, hannes AT 2horns DOT com Message-ID: From: "Arthur I Schwarz" Date: Tue, 8 Jun 2004 08:16:29 -0700 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii X-SPAM: 0.00 Hello Tim, At 2004-06-06 16:44 you wrote: > At 04:38 AM 6/6/2004, Gerrit P. Haase wrote: >>Hello Hans, >> >> > I'd like to give this one a test drive! Is it possible to use it under the >> > cygwin gcc frontend (i.e. gcc -mno-cygwin) ?? >> > Or do I need to wait for the cygwin folks to catch up (which may take an >> > eternity and a half :-)? >> >>Please try to build gcc-3.3.3 or gcc-3.4.0 (including ALL languages) >>and if it succeeds, send me your patches! >> >> >>Gerrit >>-- >>=^..^= > I've never seen instructions on where to get the special cygwin patches, > let alone mingw. The public releases up through the latest 3.3.4 build and > pass testsuite fine on cygwin. I suppose the few minor changes since the > supported versions aren't sufficient to motivate anyone to produce and > verify the cygwin patches. 3.4.0 is OK, except that there is no gnu way to > bootstrap ada, and the pch feature has not been ported to cygwin. Current > 3.4.1 exposes an internal error in g++ when I attempt to build libgcj. There was a branch for Cygwin/MinGW at version 3.3.1/3.3.2, (called cygming33x), it includes some MinGW exclusive fixes, some backported patches from HEAD and some MinGW/Cygwin specials to enable the Cygwin gcc version to use the Windows API and produce Cygwin independant binaries (-mno-cygwin and -mwin32 switches). This branch is also the version which the latest release of gcc for Cygwin (3.3.1-3) is based on. 3.4.0: the first build finished, but I got nearly 900 FAILS for the libjava testsuite, every following build of libjava even failed to finish with linking errors when building jv-convert.exe with a longer list of undefined references. As I tried to build from vanilla 3.4.0 sources because I thought, some of the MinGW/Cygwin exclusive extensions may have broken s.th. I saw the 900 FAILS which is about 10 times more FAIL than you reported. 3.3.3/3.3.4: I tried to build 3.3.3 instead of 3.4.0, now I got this error which was reported to be fixed somewhere in 2002: To all; Well I'v tried to compile GCC 3.4.0. My results are given for two machines with Win2000 and the latest version of Cygwin with gcc-3.3.1. My home computer, AMD 2100+. No problems. Ada didn't compile correctly but that's probably my fault. My work computer, Intel 800Mhz. Problems. See below. Configure was vanilla flavored. If I remember, something like "./configure". I'd like to help and (most particularly) I'd like to find out what to do and where to find what to do. art > I've tried to compile and install gcc-3.4.0 under Cygwin on an AMD XP and > Pentium IV computer, both under Win2000. > > On the AMD the bootstrap succeeds with no error and I am ready to install. > > On the Pentium the bootstrap fails (during link?) in Stage 1 when it is > unable to locate 'user32' (-luser32). user32 is located in: C: > \winnt\sysstem32\user32.dll, and this is in my Cygwin PATH variable and in > the Windows System PATH variable. Don't know what to do to discover what's > wrong or work around the problem. Don't understand why I succeed in my home > AMD computer and fail at my work pentium computer. > -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/