Hi Kim ! On Wednesday 18 February 2009, Kim Barrett wrote:

At 1:01 PM +0800 2/18/09, Dean Michael Berris wrote:

...

I have recently upgraded to Boost 1.38 from Boost 1.35 and suddenly I get the following warnings from GCC 4.3.2:

../boost_1_38_0/boost/test/floating_point_comparison.hpp:224: error: 'boost::test_tools::<unnamed>::check_is_close' defined but not used ../boost_1_38_0/boost/test/floating_point_comparison.hpp:246: error: 'boost::test_tools::<unnamed>::check_is_small' defined but not used

Sigh, yes those were introduced in 1.37 (maybe 1.36). A search in the Trac reveals that this is #2390 https://svn.boost.org/trac/boost/ticket/2390

...

I have warnings treated as errors in my build and I haven't encountered this issue before.

Can this be considered a regression and if yes what can I do to work around it?

Well, those issues are fixed in trunk, but were not merged to release in time. I hope Gennadyi will merge them for 1.39.

I ran into this same problem with the upgrade. For the moment my solution is to comment out the offending definitions for check_is_close and check_is_small, and change test_tools.hpp:

- replace all (8) occurrences of check_is_close with check_is_close_t() - replace all (4) occurrences of check_is_small with check_is_small_t()

This is definitely ugly ;-)) The real patch is avaiable at http://svn.boost.org/trac/boost/changeset/49865 And you might want to apply the second part which is https://svn.boost.org/trac/boost/changeset/50388 This fixes all warnings for me. Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! Ingenieurgesellschaft für * voice: ++49 511 262926 57 ! Verkehrs- und Eisenbahnwesen mbH * fax : ++49 511 262926 99 ! Lister Straße 15 * juergen.hunold@ivembh.de ! www.ivembh.de * * Geschäftsführer: ! Sitz des Unternehmens: Hannover * Prof. Dr.-Ing. Thomas Siefer ! Amtsgericht Hannover, HRB 56965 * PD Dr.-Ing. Alfons Radtke !

Hi Dean ! On Thursday 19 February 2009 02:34:05 Dean Michael Berris wrote:

Thanks for these links! Did they apply cleanly to Boost 1.38.0 for you? I tried both (in the given order above) and I got some rejections when I applied them to a cleanly unpacked distribution.

Well, actually not ;-(( I think you can safely ignore the patch to utils/runtime/cla/parameter.hpp as this just fixes a regression which happened in between the changes. I'll dig further into this and ask the maintainer for permission to merge at least the warning fixes to the release branch. Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! Ingenieurgesellschaft für * voice: ++49 511 262926 57 ! Verkehrs- und Eisenbahnwesen mbH * fax : ++49 511 262926 99 ! Lister Straße 15 * juergen.hunold@ivembh.de ! www.ivembh.de * * Geschäftsführer: ! Sitz des Unternehmens: Hannover * Prof. Dr.-Ing. Thomas Siefer ! Amtsgericht Hannover, HRB 56965 * PD Dr.-Ing. Alfons Radtke !

Hi Gennadiy ! On Thursday 19 February 2009, Gennadiy Rozental wrote:

Juergen Hunold <juergen.hunold <at> ivembh.de> writes:

...

Well, those issues are fixed in trunk, but were not merged to release in time. I hope Gennadiy will merge them for 1.39.

Due to lack of time need to fix some regressions with older compilers, I've decided to skip 1.38. I'll try to make it into 1.39

Well, the warning patches http://svn.boost.org/trac/boost/changeset/49865 https://svn.boost.org/trac/boost/changeset/50388 should not be the cause of regressions. It's a pity that 50388 partly fixes a warning in https://svn.boost.org/trac/boost/changeset/50354 which is a massive refactoring. I would still like to merge 49865 and the other half on 50388 to the release branch using svnmerge.py. It seems these warnings are really annyoing some people ... Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! Ingenieurgesellschaft für * voice: ++49 511 262926 57 ! Verkehrs- und Eisenbahnwesen mbH * fax : ++49 511 262926 99 ! Lister Straße 15 * juergen.hunold@ivembh.de ! www.ivembh.de * * Geschäftsführer: ! Sitz des Unternehmens: Hannover * Prof. Dr.-Ing. Thomas Siefer ! Amtsgericht Hannover, HRB 56965 * PD Dr.-Ing. Alfons Radtke !