Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin AT sources DOT redhat DOT com Message-ID: <002501c022bc$731f48c0$6a7f8ad1@timayum4srqln4> From: "Tim Prince" To: Subject: gcc-2.96 progress Date: Tue, 19 Sep 2000 21:37:12 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 I've been testing gcc-2.96 snapshots, mostly on cygwin, for quite a while, and yesterday's version provides a satisfactory g77 for the first time in many months. A noticeable difference is that the options -Os and -O2 at last live up to their intentions; -O2 is as fast as anything and -Os produces smaller code. I'm not so impressed with the reliability of the gcc or g++ compilers, given the inability to complete a bootstrap and the number of make check failures in g++ and its libraries, but it does look like progress. g77 runs make check and many other tests with the same result as on linux, aside from the continued habit of dejagnu of running the tests twice. I've made a long overdue revision of the procedure description for building gcc on cygwin, at http://members.aol.com/n8tm/cygwingccg77.htm If anyone has ways to make it better, please let me know; there's much room for improvement. Yes, I know the reasons why some of the steps I had to take aren't recommended. But anyway, gcc/g77 performs in the same league with the Compaq, Intel, and Lahey compilers on my T20 laptop, and the cygwin and gcc developers deserve a great deal of credit. -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com