Any way to speed my use of boost_pp macros?
The file, http://groups.yahoo.com/group/boost/files/managed_ptr/ctor_fwder.zip , shows a simple program using boost/preprocessor macros taking 17secs to execute. If anyone has some suggestions for speeding this up, I'd certainly appreciate it. TIA.
On 03/25/2004 01:58 PM, Larry Evans wrote:
The file, http://groups.yahoo.com/group/boost/files/managed_ptr/ctor_fwder.zip , Just a more descriptive subject.
I've found an odd bug with vc 7.1 and lambda: namespace { struct test_struct { int b; }; void f() { namespace ll = boost::lambda; ll::bind(&test_struct::b, ll::_1); } } If I try to compile this code vc 7.1 will rapidly chew up all available memory until it runs out. I'm not sure if it will ever finish or not. If you remove test_struct from the anonymous namespace it compiles just fine. I've never seen vc 7.1 do this before. Any ideas? Thanks, Brock
Brock Peabody <brock.peabody@npcinternational.com> wrote:
namespace it compiles just fine. I've never seen vc 7.1 do this before. Any ideas?
following test file:
begin #include <boost/lambda/lambda.hpp> #include <boost/lambda/bind.hpp>
namespace { struct test_struct { int b; }; void f() { namespace ll = boost::lambda; ll::bind(&test_struct::b, ll::_1); } } int main() { f(); }
end
compiles successfully on my MSVC71, compiler options are: /O2 /GL /GR /GX /MD. Do you want me to try different options or different test file? Regards B.
On Behalf Of Bronek Kozicki
compiles successfully on my MSVC71, compiler options are: /O2 /GL /GR /GX /MD. Do you want me to try different options or different test file?
I was compiling from the IDE, but sure enough when I used the command line with your options it compiled successfully. I tracked it down to the /Zi option. If you use it you should be able to get the same behavior. I've tried this on two computers.
Brock Peabody <brock.peabody@npcinternational.com> wrote:
line with your options it compiled successfully. I tracked it down to the /Zi option. If you use it you should be able to get the same
OK, I got it. It also happens with /ZI and /Z7, but not with /Zd. Reported to Microsoft, I will try to reproduce it without including boost headers. B.
Bronek Kozicki <brok@rubikon.pl> wrote:
OK, I got it. It also happens with /ZI and /Z7, but not with /Zd.
from discussion on MS newsgroup I see that other person is unable to reproduce this problem. Could you post some more info about your setup (like OS, RAM, processors, boost version)? B.
Bronek Kozicki <brok@rubikon.pl> wrote:
from discussion on MS newsgroup I see that other person is unable to reproduce this problem. Could you post some more info about your setup (like OS, RAM, processors, boost version)?
Hmm that's odd. You didn't have any trouble recreating it once you use /Zi, right? I'm using boost version 1.31. I've tried it on two computers: P4 2.4 with 768MB of RAM, Windows XP-SP1 And AMD 2600+ with 1GB of RAM, Windows XP-SP1.
"Brock Peabody" <brock.peabody@npcinternational.com> writes:
I've found an odd bug with vc 7.1 and lambda:
From the Posting Nanny:
When starting a new topic, please do it with a fresh message instead of changing the subject of a reply to some other thread. Many mailers (like mine) are able to detect the thread you started with and will show the new message as part of the original thread. The message I'm replying to shows up as part of the "faster ctor forwarder macros?" thread, for example. I ask this for your own sake as well as for others': Often people scanning for relevant messages will decide they're done with a topic and hide or kill the entire thread: your message will be missed, and you won't get the help you're looking for. Thanks, -- Dave Abrahams Boost Consulting www.boost-consulting.com
On Behalf Of David Abrahams
When starting a new topic, please do it with a fresh message instead of changing the subject of a reply to some other thread. Many mailers (like mine) are able to detect the thread you started with and will show the new message as part of the original thread.
I apologize, I had no idea that this was possible.
"Brock Peabody" <brock.peabody@npcinternational.com> writes:
On Behalf Of David Abrahams
When starting a new topic, please do it with a fresh message instead of changing the subject of a reply to some other thread. Many mailers (like mine) are able to detect the thread you started with and will show the new message as part of the original thread.
I apologize, I had no idea that this was possible.
I just updated the posting guidelines to reflect that fact. -- Dave Abrahams Boost Consulting www.boost-consulting.com
-----Original Message----- [mailto:boost-bounces@lists.boost.org] On Behalf Of Larry Evans
Subject: [boost] Any way to speed my use of boost_pp macros?
The file, http://groups.yahoo.com/group/boost/files/managed_ptr/ctor_fwd er.zip , shows a simple program using boost/preprocessor macros taking 17secs to execute. If anyone has some suggestions for speeding this up, I'd certainly appreciate it.
Hi Larry, (I'd guess that you are using an EDG-based compiler.) Can you send me the file directly? Regards, Paul Mensonides
On 03/25/2004 06:26 PM, Paul Mensonides wrote:
-----Original Message----- [mailto:boost-bounces@lists.boost.org] On Behalf Of Larry Evans
(I'd guess that you are using an EDG-based compiler.) Can you send me the file directly?
I did and Paul replied with a much improved version. If there's any interest, I'll post it. This post is essentially a duplicate of one I sent in response to the "faster ctor forwarder macros?" thread, which, as I just learned from David, may have been missed.
Larry Evans <cppljevans@cox-internet.com> writes:
On 03/25/2004 06:26 PM, Paul Mensonides wrote:
-----Original Message----- [mailto:boost-bounces@lists.boost.org] On Behalf Of Larry Evans (I'd guess that you are using an EDG-based compiler.) Can you send me the file directly?
I did and Paul replied with a much improved version. If there's any interest, I'll post it.
This post is essentially a duplicate of one I sent in response to the "faster ctor forwarder macros?" thread, which, as I just learned from David, may have been missed.
Your post was fine. I was referring to the post from "Brock Peabody" <brock.peabody@npcinternational.com> beginning:
I've found an odd bug with vc 7.1 and lambda:
-- Dave Abrahams Boost Consulting www.boost-consulting.com
participants (5)
-
Brock Peabody -
Bronek Kozicki -
David Abrahams -
Larry Evans -
Paul Mensonides