I was just about to write the same, but Erik was faster... :-) The point is that the mode is process-global state, and we can't really change that. I think the main point is not about 80 or 64 bit, but being consistent in what we calculate. This issue arose last week, where it became obvious that we e.g. can't use fptan sometimes and call the libc runtime function at other times. Using -ffloat-store might help with the problem on MinGW32, but perhaps there are other consistency problems.
Cheers, S. On Mon, Mar 5, 2012 at 08:02, Erik Corry <[email protected]> wrote: > This seems a rather drastic move. According to this: > http://www.christian-seiler.de/projekte/fpmath/ MingW32 is normally > run in extended precision mode. This means that any application that > both uses long double and V8 is now broken. > > V8 on Linux 32 bit mode is also run in extended precision mode (glibc > requires it, and switching is expensive). If we can get away with > that, can we not also get away with using extended precision on > MingW32? > > On Sun, Mar 4, 2012 at 11:50 AM, <[email protected]> wrote: > > Revision: 10913 > > Author: [email protected] > > Date: Sun Mar 4 02:49:25 2012 > > Log: Set x86 FPU precision to 64-bit for MinGW. Original patch by > > Jonathan Liu <[email protected]>. > > > > BUG=v8:1062 > > TEST=test-strtod.cc > > > > Review URL: https://chromiumcodereview.appspot.com/9599006 > > Patch from Jonathan Liu <[email protected]>. > > http://code.google.com/p/v8/source/detail?r=10913 > > > > Modified: > > /branches/bleeding_edge/src/platform-win32.cc > > /branches/bleeding_edge/src/win32-headers.h > > /branches/bleeding_edge/test/cctest/test-strtod.cc > > > > ======================================= > > --- /branches/bleeding_edge/src/platform-win32.cc Sat Mar 3 > 03:07:29 > > 2012 > > +++ /branches/bleeding_edge/src/platform-win32.cc Sun Mar 4 > 02:49:25 > > 2012 > > @@ -539,6 +539,10 @@ > > > > > > void OS::SetUp() { > > +#ifdef __MINGW32__ > > + // Set x87 FPU precision to 64-bit to be consistent with MSVC. > > + _controlfp(_PC_53, _MCW_PC); > > +#endif > > // Seed the random number generator. > > // Convert the current time to a 64-bit integer first, before > converting > > it > > // to an unsigned. Going directly can cause an overflow and the seed > to be > > ======================================= > > --- /branches/bleeding_edge/src/win32-headers.h Wed Feb 22 04:26:36 2012 > > +++ /branches/bleeding_edge/src/win32-headers.h Sun Mar 4 02:49:25 2012 > > @@ -63,6 +63,7 @@ > > // header files to expose getaddrinfo. > > #undef _WIN32_WINNT > > #define _WIN32_WINNT 0x501 > > +#include <float.h> // For _controlfp(). > > #endif // __MINGW32__ > > #if !defined(__MINGW32__) || defined(__MINGW64_VERSION_MAJOR) > > #include <dbghelp.h> // For SymLoadModule64 and al. > > @@ -78,7 +79,7 @@ > > #ifndef __MINGW32__ > > #include <wspiapi.h> > > #endif // __MINGW32__ > > -#include <process.h> // for _beginthreadex() > > +#include <process.h> // For _beginthreadex(). > > #include <stdlib.h> > > #endif // V8_WIN32_HEADERS_FULL > > > > ======================================= > > --- /branches/bleeding_edge/test/cctest/test-strtod.cc Wed Feb 9 > 06:12:31 > > 2011 > > +++ /branches/bleeding_edge/test/cctest/test-strtod.cc Sun Mar 4 > 02:49:25 > > 2012 > > @@ -1,4 +1,29 @@ > > -// Copyright 2006-2008 the V8 project authors. All rights reserved. > > +// Copyright 2012 the V8 project authors. All rights reserved. > > +// Redistribution and use in source and binary forms, with or without > > +// modification, are permitted provided that the following conditions > are > > +// met: > > +// > > +// * Redistributions of source code must retain the above copyright > > +// notice, this list of conditions and the following disclaimer. > > +// * Redistributions in binary form must reproduce the above > > +// copyright notice, this list of conditions and the following > > +// disclaimer in the documentation and/or other materials provided > > +// with the distribution. > > +// * Neither the name of Google Inc. nor the names of its > > +// contributors may be used to endorse or promote products derived > > +// from this software without specific prior written permission. > > +// > > +// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS > > +// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT > > +// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR > > +// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT > > +// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, > > +// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT > > +// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, > > +// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY > > +// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT > > +// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE > > +// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. > > > > #include <stdlib.h> > > > > @@ -23,6 +48,8 @@ > > > > > > TEST(Strtod) { > > + v8::V8::Initialize(); > > + > > Vector<const char> vector; > > > > vector = StringToVector("0"); > > > > -- > > v8-dev mailing list > > [email protected] > > http://groups.google.com/group/v8-dev > > > > -- > Erik Corry, Software Engineer > Google Denmark ApS - Frederiksborggade 20B, 1 sal, > 1360 København K - Denmark - CVR nr. 28 86 69 84 > > -- > v8-dev mailing list > [email protected] > http://groups.google.com/group/v8-dev > -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
