Richard Webb wrote:

Beman Dawes <bdawes <at> acm.org> writes:

...

The release branch is frozen. Please make no further changes without explicit permission from me.

RC1 is being generated.

--Beman

I notice that the release branch test results are showing some Wave failures that only appeared recently. Do these need sorting before the release?

Good question! Wave folks? --Beman

Beman Dawes wrote:

Tobias Schwinger wrote:

...

Beman Dawes wrote:

...

The release branch is frozen. Please make no further changes without explicit permission from me. OK, I'd love to stage FunctionTypes to the release branch within the next couple of days.

All changes are about greening the test matrix and fixing the tickets regarding the build process.

I have (hopefully) just added support for Sun (thanks to Noel Belcourt) and I am ready to go once I see the test results from the trunk.

Let's make sure we are on the same page. FunctionTypes is already in the release branch, and you just want to apply some fixes, which are now being tested already on trunk. Is that right?

Yes.

If so, go ahead as long as the trunk tests are clean and you can do it within the next three days. Say by the end of Monday.

OK. Thanks, Tobias

Ion Gaztañaga wrote:

Beman Dawes wrote:

...

The release branch is frozen. Please make no further changes without explicit permission from me.

I've received a trivial to fix bug report (#1683):

#1683: Boost Intrusive - Superflous template parameter --------------------------------+------------------------------------------- Reporter: boost@strigeus.com | Owner: igaztanaga Type: Bugs | Status: new Milestone: Boost 1.35.0 | Component: intrusive Version: Boost 1.34.1 | Severity: Problem Keywords: | --------------------------------+------------------------------------------- In boost/intrusive/unordered_set.hpp

template<class KeyType, class KeyHasher, class KeyValueEqual, class Disposer> size_type count(const KeyType& key, KeyHasher hash_func, KeyValueEqual equal_func) const { return table_.count(key, hash_func, equal_func); }

The 'class Disposer' template argument is superflous. With that there, the compiler is not able to automatically infer the template types so I can't write foo.count<>(key,hash,equal).

/////////////////////

The fix is trivial and does not affect regression testing (because if this function would be instantiated by any test case, it wouldn't compile). Could I fix this both in release and trunk?

Fix it in trunk, then hold a day or so to let tests cycle, then fix in release if it's OK in trunk testing. Even though the fix is trivial, there is always the chance of some unexpected snafu. We want to get RC2 ready by the middle of the week, and want minimal chance of delay. Thanks, --Beman

Ion Gaztañaga wrote:

Beman Dawes wrote:

...

Fix it in trunk, then hold a day or so to let tests cycle, then fix in release if it's OK in trunk testing.

Even though the fix is trivial, there is always the chance of some unexpected snafu. We want to get RC2 ready by the middle of the week, and want minimal chance of delay.

Thanks,

--Beman

Ok. Fixed in trunk.

Thanks,

Trunk tests are ok so I've fixed it in release: #1683: Boost Intrusive - Superflous template parameter Revision 46397. Regards, Ion

Eric Niebler wrote:

Beman Dawes wrote:

...

The release branch is frozen. Please make no further changes without explicit permission from me.

Can I merge some documentation changes? http://svn.boost.org/trac/boost/changeset/43663/

Yes. --Beman

Doug Gregor wrote:

On Mar 14, 2008, at 8:33 AM, Beman Dawes wrote:

...

The release branch is frozen. Please make no further changes without explicit permission from me.

I have some graph and MPI documentation fixes that I'd like to push into the release branch. Is that still okay?

Yes. --Beman