Re: [boost] [Boost-commit] svn:boost r68201
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Sun, 16 Jan 2011 23:52:02 -0500 (EST) steven@providere-consulting.com wrote:
Author: steven_watanabe Date: 2011-01-16 23:51:54 EST (Sun, 16 Jan 2011) New Revision: 68201 URL: http://svn.boost.org/trac/boost/changeset/68201
Log: Revert addition of BOOST_NIX, etc and everything tied to it. ([66696], [66783], [66804], [66833], and [66834]) Removed:
fenv.h is not a part of the C++ standard library; it is specified in ISO/IEC 9899, section 7.6 (the C99 standard). It's implementation is tied not only to the underlying C library in use, but also to a number of other factors, including but not limited to: - The operating system - The kernel - The compiler's method of implementing header search Attempting to detect the presence of the fenv.h header and the toolchains support for using said header cannot be achieved naively by defining a macro if the GNU, vacpp, sgi or stlport standard library implementation is found. The changes you have introduced will break Boost support for (at least) the following compilers and toolchains: - Sun, HP, Pathscale - Clang on linux or darwin using libc++ - Any compiler using the Apache/Roguewave standard library I am sincerely sorry to have stepped on toes here at Boost, especially yours, as you seem to be hellbent in purging Boost of my contributions, instead of helping to improve upon them. You told me last night in an email that you went back and reviewed every commit that I have ever made to Boost. I would appreciate it if you would let me know how much of my work you have in your crosshairs. Anyone here who knows me is aware that I am young and inexperienced at working on a large-scale open source project. I have been a little over-enthusastic and unintentionally inconsiderate, but I have done my best to fit in here and make contributions to Boost. I do not think that reverting my commits without consulting me first encourages the SVN policies that you and others assert that I have blatantly violated. - -- Bryce Lelbach aka wash boost-spirit.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEARECAAYFAk0z11gACgkQ9cB/V3/s9EyXJQCfesVXrNkhAeNoYCmgWODv0Z7m 6fMAni0Sb7gdtQWHyIj7EVFhrwqvBjCv =Sr5U -----END PGP SIGNATURE-----
On 1/17/2011 12:44 PM, Bryce Lelbach wrote:
On Sun, 16 Jan 2011 23:52:02 -0500 (EST) steven@providere-consulting.com wrote:
Author: steven_watanabe Date: 2011-01-16 23:51:54 EST (Sun, 16 Jan 2011) New Revision: 68201 URL: http://svn.boost.org/trac/boost/changeset/68201
Log: Revert addition of BOOST_NIX, etc and everything tied to it. ([66696], [66783], [66804], [66833], and [66834]) Removed:
<snip>
I am sincerely sorry to have stepped on toes here at Boost, especially yours, as you seem to be hellbent in purging Boost of my contributions, instead of helping to improve upon them.
Careful ...
You told me last night in an email that you went back and reviewed every commit that I have ever made to Boost. I would appreciate it if you would let me know how much of my work you have in your crosshairs.
Careful ...
Anyone here who knows me is aware that I am young and inexperienced at working on a large-scale open source project. I have been a little over-enthusastic and unintentionally inconsiderate, but I have done my best to fit in here and make contributions to Boost. I do not think that reverting my commits without consulting me first encourages the SVN policies that you and others assert that I have blatantly violated.
Bryce, I know you're upset at having your contributions rolled back. But I suggest you choose your words with a little more care. You're tone with Steven is unwarranted. We're all trying to make Boost better. I guarantee that Steven doesn't have some personal vendetta. You made some check-ins without going through the usual channels, and a few turned out to have problems. On a big project like Boost, the safest course is to roll them back and try again later. Nothing out of the ordinary. We (the moderators) are trying to enforce policy. If we didn't there'd be chaos. I'm sure you can understand. If you have something to contribute, make your case on the list and put together a patch. I promise you that valuable and correct contributions won't be rejected. -- Eric Niebler BoostPro Computing http://www.boostpro.com
Log: Revert addition of BOOST_NIX, etc and everything tied to it. ([66696], [66783], [66804], [66833], and [66834]) Removed:
fenv.h is not a part of the C++ standard library; it is specified in ISO/IEC 9899, section 7.6 (the C99 standard). It's implementation is tied not only to the underlying C library in use, but also to a number of other factors, including but not limited to:
- The operating system - The kernel - The compiler's method of implementing header search
To be fair to Bryce, I approved the changes to fenv.hpp. Please note also that I merged the current Boost.Config state to release a short while ago, so that needs to be re-synchronised with Trunk once the dust has settled, but in any case before the next release. Probably not paying enough attention yours, John.
On 17 January 2011 10:04, John Maddock <boost.regex@virgin.net> wrote:
To be fair to Bryce, I approved the changes to fenv.hpp. Please note also that I merged the current Boost.Config state to release a short while ago, so that needs to be re-synchronised with Trunk once the dust has settled, but in any case before the next release.
I agree, I don't think this is a good time to revert changes from a couple of months ago, especially as some of these changes seem to mix a few things up (a large part of this problem is due to how we do version control and how we manage boost). If this is something that's icky but not broken maybe we should just remove the documentation for the controversial macros. Daniel
AMDG On 1/17/2011 3:36 AM, Daniel James wrote:
On 17 January 2011 10:04, John Maddock<boost.regex@virgin.net> wrote:
To be fair to Bryce, I approved the changes to fenv.hpp. Please note also that I merged the current Boost.Config state to release a short while ago, so that needs to be re-synchronised with Trunk once the dust has settled, but in any case before the next release.
I agree, I don't think this is a good time to revert changes from a couple of months ago, especially as some of these changes seem to mix a few things up (a large part of this problem is due to how we do version control and how we manage boost). If this is something that's icky but not broken maybe we should just remove the documentation for the controversial macros.
Good point. I was trying to get this done before the release, to avoid releasing new features only to pull them out again. I've already made the changes in the trunk, but I can merge only the documentation now and re-sync everything else after the release. That will also give me a little more time to restore the fenv bug-fixes. In Christ, Steven Watanabe
AMDG On 1/17/2011 2:04 AM, John Maddock wrote:
Log: Revert addition of BOOST_NIX, etc and everything tied to it. ([66696], [66783], [66804], [66833], and [66834]) Removed:
fenv.h is not a part of the C++ standard library; it is specified in ISO/IEC 9899, section 7.6 (the C99 standard). It's implementation is tied not only to the underlying C library in use, but also to a number of other factors, including but not limited to:
- The operating system - The kernel - The compiler's method of implementing header search
To be fair to Bryce, I approved the changes to fenv.hpp.
That's fine. It only got reverted because it was mixed up with the changes that I was trying to revert.
Please note also that I merged the current Boost.Config state to release a short while ago, so that needs to be re-synchronised with Trunk once the dust has settled, but in any case before the next release.
Probably not paying enough attention yours, John.
In Christ, Steven Watanabe
AMDG On 1/16/2011 9:44 PM, Bryce Lelbach wrote:
I would appreciate it if you would let me know how much of my work you have in your crosshairs.
There's nothing except what I've already told you about. It's just that reverting individual features is harder when you commit multiple independent changes simultaneously. This is especially true if the log messages don't mention everything that the changeset does. In Christ, Steven Watanabe
AMDG On 1/16/2011 9:44 PM, Bryce Lelbach wrote:
Attempting to detect the presence of the fenv.h header and the toolchains support for using said header cannot be achieved naively by defining a macro if the GNU, vacpp, sgi or stlport standard library implementation is found. The changes you have introduced will break Boost support for (at least) the following compilers and toolchains:
My mistake. The commit that you changed this in was just mislabeled. I'll put it back. In Christ, Steven Watanabe
participants (5)
-
Bryce Lelbach -
Daniel James -
Eric Niebler -
John Maddock -
Steven Watanabe