August 2004 - Boost - lists.boost.org

[info] More Boost.org copyrights removed
by John Maddock 15 Aug '04

15 Aug '04

Here's the final batch of file to have their boost.org copyrights removed, as before: If your name has been added when it should not have been, or if your name should have been added to the copyright declarations, but has not been, then please speak up now. Here are the files affected: boost/cast.hpp: Kevlin Henney and Dave Abrahams boost/checked_delete.hpp: Peter Dimov Daniel Frey Howard Hinnant boost/integer_fwd.hpp: Dave Abrahams. boost/math_fwd.hpp: Dave Abrahams. boost/smart_ptr.hpp: Peter Dimov libs/conversion/cast_test.cpp: Beman Dawes, Dave Abrahams. libs/utility/noncopyable_test.cpp: Beman Dawes. Thanks, John Maddock

2 2

[Info] Updated Boost.org copyrights
by John Maddock 15 Aug '04

15 Aug '04

We agreed a while ago that the "Boost.org" copyrights should be removed from source files (on the grounds that there is no legal entity called Boost.org) So, the following files have had their copyright declarations updated as below, copyright assignments are based on cvs history. If I have either: missed your name off the list, or added your name when you don't think that it should be there, then please speak up now. Here's the list of assignments (so far): boost/io_fwd.hpp: Beman Dawes and Daryle Walker boost/iterator.hpp: Beman Dawes boost/limits.hpp: John Maddock, Dave Abrahams. boost/mpl/if.hpp: Aleksey Gurtovoy boost/next_prior.hpp: Aleksey Gurtovoy boost/noncopyable.hpp: Beman Dawes boost/numeric/conversion/cast.hpp: Fernando Luis Cacciola Carballal boost/utility.hpp Aleksey Gurtovoy boost/utility_fwd.hpp: Beman Dawes boost/version.hpp: John Maddock Thanks, John Maddock.

2 2

Fw: Boost,License Permission
by John Maddock 15 Aug '04

15 Aug '04

Forwarded here as a permanent record: ----- Original Message ----- From: "Eric Ford" <un5o6n902(a)sneakemail.com> To: <john(a)johnmaddock.co.uk> Sent: Friday, August 13, 2004 2:01 AM Subject: Boost,License Permission > I, Eric Ford, give permission for all my contributions to Boost to be > moved to the Boost Software License Version 1 >

1 0

[ATT:Jim Apple] Memoization Library status?
by Mattias Flodin 15 Aug '04

15 Aug '04

In December last year, Jim Apple asked if there was interest in a memoization library he was working on. [1] It seems there were several people interested in this, but I have seen no more posts about it. Was this project discontinued? Are there other boosters working on a library like this? Dave Abrahams pointed out (in [2]) that a library such as this may need the ability to invalidate parts of the cache, but was met with confusion and I don't think the motivation for the need was widely understood. I have had increasing need for a memoization library such as this in our current project, so I thought I'd try to summarize it as a kind of use case. The project is a medical diagnosis application using in the order of a hundred image processing step for diagnosis of digitized images. Some of these steps can take a very significant amount of time, and many of them may need to use intermediate results that were already computed in other steps. I am currently obtaining the result image recursively through a very rudimentary non-generic memoization cache class, where recursive calls are explicitly routed through the cache object (by a string->funptr mapping). Thanks to the caching I am now probably cutting away a good 90% of processing time. Although all processing-step functions now (of necessity) have the same signature, many of them use various globally set parameters related to the algorithm, such as threshold settings for neural-network outputs. Some parameters are not used until late in the process, whereas other parameters affect early steps. To optimize the parameter values we must perform a lot of searches through the parameter space using various parameter optimization techniques. Currently, with every parameter change we must re-run the entire process, even if the parameter actually only affects one of the very last steps. With a memoization library that knows about parameter dependencies, we would be able to try many different values of this parameter and only the dependent parts of the cache would be invalidated, which could save hours if not days of processing time. /Mattias [1] http://lists.boost.org/MailArchives/boost/msg56780.php [2] http://lists.boost.org/MailArchives/boost/msg56913.php

2 1

Re:Re: [boost] [serialization] export macros
by Robert Ramey 14 Aug '04

14 Aug '04

Christoph Ludwig wrote: >> template<class> struct pack; >> >> template<class T> struct pack<void (T)> { >> typedef T type; >> }; >> >> BOOST_CLASS_EXPORT( pack<void (DerivedStatus< Algo<int, double > >> >)>::type ); >That did the trick! Great! >I am not sure which of Boost's libraries is the most adequate place for >such a tool (preprocessor? mpl?), but it is certainly valuable. And if / >once there is such a tool in Boost, then the documentation of macros that >are likely to take classes genareted from templates as arguments should >refer to it. Hmmm - maybe we should just roll this into export.hpp - Would this create any other problems? Robert Ramey

3 6

remove_pointer type trait metafunction and smart pointers
by Jeff Mirwaisi 14 Aug '04

14 Aug '04

Currently (afaict) there isnt a specialization of remove_pointer that handles smart pointers. Providing a very generic specialization: template <typename T, template<typename> class Ptr> struct remove_pointer<Ptr<T> >; or a set of specializations that cover shared_ptr etc + auto_ptr explicitly, "seems" like a rational extension. Are there any potential problems, and if not can we get this added to type traits?

2 2

Directory for the regression test results doesn't exist.
by Martin Wille 14 Aug '04

14 Aug '04

Hi, http://boost.sourceforge.net/regression-logs doesn't exist. Instead, there is a tar file owned by root which seems to contain the result directory. What's happening there? Regards, m

2 1

[repost] regressions in como-win32-4.3.3-vc7.1
by JOAQUIN LOPEZ MU?Z 14 Aug '04

14 Aug '04

This post caught nobody's attention two days ago, but the error is still showing... -- Original message -- Please see http://tinyurl.com/5f7ba "C:\Comeau\libcomo\cnames\cstring", line 31: error #247: function "std::strlen" has already been defined extern "C++" inline size_t strlen(const char *arg) { return ::strlen(arg); } ^ "C:\Comeau\libcomo\cnames\cstring", line 31: error #247: function "std::strlen" has already been defined extern "C++" inline size_t strlen(const char *arg) { return ::strlen(arg); } ^ 2 errors detected in the compilation of "..\libs\multi_index\test\test_basic_main.cpp". The problem has nothing to do with Boost.MultiIndex, actually, as test_basic_main.cpp does not do anything more than including <boost/test/included/test_exec_monitor.hpp> These tests used to pass one or two days ago. I've inspected Boost.Test for recent changes and found none. Other libraries are affected by this problem: algorithm/minmax algorithm/string program_options ... Thanks Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

3 5

regression - failed tests for date_time and regex in vc7.1 toolkit
by Bronek Kozicki 14 Aug '04

14 Aug '04

I've noticed that regression test I'm running are failing on almost all tests in date_time and many tests in regex. Here's what I've found: bjam is building static library: libboost_date_time-71-mt-gd-1_32.lib Please note that there's no "vc71", but "71" instead. File: libboost_date_time-vc71-mt-gd-1_32.lib is not build, thus all tests linking to this file are failing. When I copy libboost_date_time-71-mt-gd-1_32.lib into libboost_date_time-vc71-mt-gd-1_32.lib , date_time tests are no longer failing. The same applies to dynamic library: what is actually being build is named: boost_date_time-71-mt-gd-1_32.lib boost_date_time-71-mt-gd-1_32.dll but for tests to succeed there should be: boost_date_time-vc71-mt-gd-1_32.lib boost_date_time-vc71-mt-gd-1_32.dll What's even more interesting, library files "71" are actually being used, too (besides "vc71" files). I've found that "cmd" files are referring to "71" files. I do not know what is referring to "vc71" files (maybe autolink feature?) The same problem plagued regression tests for regex library - libraries build by bjam are named : libboost_regex-71-mt-gd-1_32.lib boost_regex-71-mt-gd-1_32.lib boost_regex-71-mt-gd-1_32.dll while tests require also files named : libboost_regex-vc71-mt-gd-1_32.lib boost_regex-vc71-mt-gd-1_32.lib boost_regex-vc71-mt-gd-1_32.dll Any ideas what might be wrong? B.

3 6

"Making a Boost Library" - article for C/C++ user's journal - request for critique.
by Robert Ramey 14 Aug '04

14 Aug '04

Fellow boosters, Editors of the C/C++ users journal accepted my proposal for an article titled "Making a Boost Library" (subject to final review of course). Rather than being "code centric" as other articles for this journal are, this will focus on the Boost library development and review process from the vantage point of a library submitter. I've got the article done. Its 8 pages (3500 words) in Microsoft word format. If anyone is interested in reading this and critiquing it, please advise me by private email and I'll send you a copy for your review. Robert Ramey

1 0