From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 2066 invoked by alias); 8 Aug 2008 14:41:01 -0000 Received: (qmail 5946 invoked by uid 22791); 8 Aug 2008 12:27:13 -0000 X-Spam-Check-By: sourceware.org Message-ID: <24446FFC01E14C3891D7A63850C90C46@desktop2> From: "Sisyphus" To: "Brian Gough" Cc: References: <87k5f69jeb.wl%bjg@network-theory.co.uk> In-Reply-To: Subject: Re: Some long double problems on Win32 Date: Fri, 08 Aug 2008 14:41:00 -0000 MIME-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=original Content-Transfer-Encoding: 7bit X-Mailer: Microsoft Windows Mail 6.0.6001.18000 Mailing-List: contact gsl-discuss-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gsl-discuss-owner@sourceware.org X-SW-Source: 2008-q3/txt/msg00014.txt.bz2 ----- Original Message ----- From: "Brian Gough" To: "Sisyphus" Cc: Sent: Friday, August 08, 2008 8:32 PM Subject: Re: Some long double problems on Win32 > At Sun, 3 Aug 2008 19:45:33 +1000, > Sisyphus wrote: >> But the library seems to be mssing the following symbols (which are to be >> found in gsl_block_complex_long_double.h and gsl_block_long_double.h): >> >> `gsl_block_complex_long_double_fscanf' >> `gsl_block_complex_long_double_fprintf' >> `gsl_block_complex_long_double_raw_fscanf' >> `gsl_block_complex_long_double_raw_fprintf' >> `gsl_block_long_double_fscanf' >> `gsl_block_long_double_fprintf' >> `gsl_block_long_double_raw_fscanf' >> >> Anyone know why that might be ? > > The configure test for fprintf/fscanf with long-double failed so those > functions are not compiled. Yes - it finally dawned on me. Just last night I got around to rewriting the test so that it would pass and, sure enough, the library that got built included those symbols. The tests failed, of course :-) What is now puzzling me, however, is whether it's impossible to incorporate those symbols into a Win32 gsl library. MinGW certainly knows about the 12 byte 'long double', but afaik the Microsoft C runtime (which MinGW-built apps use) has no formatter to printf() it. That doesn't quite make sense to me. Why would MinGW provide a 12 byte long double if there's no way of printf()'ing it out ? (I suppose this aspect would be better raised on the MinGW list.) Thanks for the reply, Brian. Cheers, Rob