Re: [boost] [utility/swap] Internal compiler error Siliconmanborland-6.10.0 (CodeGear?)
David (siliconman) wrote:
It had already been logged when I got to looking: <http://qc.codegear.com/wc/qcmain.aspx?d=68959>
Thanks, David. I've been searching through qc.codegear.com before, but I hadn't found it. I guess I just chose the wrong keyword to search for, e.g., "template", and "Fatal F1004". BTW, I decided to submit the report myself, because I don't have access to borland-6.10.0. As far as I know, it isn't available for free. (I do have the free Turbo C++ 2006 Explorer installed, though.) If I had access to borland-6.10.0, I would have tried to make a smaller test program to reproduce the internal compiler error. Because I wonder if it is specific to the current implementation of the boost::swap utility... Please let me know if you have any suggestion for boost::swap (or its specialized_in_std.cpp test) to work around the compiler error. Kind regards, Niels PS Your replies don't appear on the mail archives: http://lists.boost.org/Archives/boost/2008/11/index.php http://thread.gmane.org/gmane.comp.lib.boost.devel/182214 Is that intentional?
In article <491EA6EA.F9F7C3C6@xs4all.nl>, Niels Dekker - mail address until 2010-10-10 <niels_address_until_2010-10-10@xs4all.nl> wrote:
PS Your replies don't appear on the mail archives:
OK, I've adjusted my headers. X-no-archive was set. This one should appear in the archive. -- -David
David (siliconman) wrote:
OK, I've adjusted my headers. X-no-archive was set. This one should appear in the archive.
Sorry David, your postings still don't appear in the archives. Maybe you should also remove "X-Archive: encrypt"??? Thanks again for providing the link to the CodeGear bug report, "Boost ICE", by Christopher Greene, http://qc.codegear.com/wc/qcmain.aspx?d=68959 Still I decided to dedicate another report to the issue, because I found two simple ways to reproduce the internal compiler error without using any Boost: Report #69196, "Fatal F1004 on std::swap template specialization caused by rvalue reference bug" http://qc.codegear.com/wc/qcmain.aspx?d=69196 I hope I've made clear that the compiler bug is not Boost-specific! Kind regards, Niels
In article <D3A4C3C0F5444AD786AC9C1B21BDF738@4lkebivusnd>, "Niels Dekker - mail address until 2010-10-10" <niels_address_until_2010-10-10@xs4all.nl> wrote:
Sorry David, your postings still don't appear in the archives. Maybe you should also remove "X-Archive: encrypt"???
Trying again...
Thanks again for providing the link to the CodeGear bug report, "Boost ICE", by Christopher Greene, http://qc.codegear.com/wc/qcmain.aspx?d=68959 Still I decided to dedicate another report to the issue, because I found two simple ways to reproduce the internal compiler error without using any Boost: Report #69196, "Fatal F1004 on std::swap template specialization caused by rvalue reference bug" http://qc.codegear.com/wc/qcmain.aspx?d=69196 I hope I've made clear that the compiler bug is not Boost-specific!
It looks like they both point to the same bug in CodeGear's internal system. -- -David
CodeGear 6.10.0 has a bug, which causes an internal compiler error when you try to specialize std::swap. As was reported by Christopher Greene <http://qc.codegear.com/wc/qcmain.aspx?d=68959> and me <http://qc.codegear.com/wc/qcmain.aspx?d=69196>. To my surprise, the compiler error appears to have gone away automatically, and three boost/utility/swap tests now have an "unexpected success": www.boost.org/development/tests/trunk/developer/utility-swap_.html www.boost.org/development/tests/trunk/output/siliconman-boost-bin-v2-libs-utility-swap-test-specialized_in_std-test-borland-6-10-0-debug.html Is there any explanation to this unexpected success? The compiler version hasn't changed, apparently... Kind regards, -- Niels Dekker http://www.xs4all.nl/~nd/dekkerware Scientific programmer at LKEB, Leiden University Medical Center
CodeGear 6.10.0 has a bug, which causes an internal compiler error when you try to specialize std::swap.
To my surprise, the compiler error appears to have gone away automatically, and three boost/utility/swap tests now have an "unexpected success": www.boost.org/development/tests/trunk/developer/utility-swap_.html
Is there any explanation to this unexpected success? The compiler version hasn't changed, apparently...
David Dean informed me that Siliconman temporarily ran a new version of CodeGear, instead of the original CodeGear 6.10.0. Now the original version is restored, so the expected failures are back again: www.boost.org/development/tests/trunk/output/siliconman-boost-bin-v2-libs-utility-swap-test-specialized_in_std-test-borland-6-10-0-debug.html Thanks, David. -- Niels
participants (2)
-
Niels Dekker - mail address until 2010-10-10 -
siliconman