[unordered] Status, incompatible with Interprocess
The 18.03.2007 20:42 upload of unordered.tar.gz in the vault, which as far as I know is the most recent copy of the proposed Boost unordered containers library, is no longer compatible with Ion Gaztañaga's Boost Interprocess. See <http://lists.boost.org/boost-users/2007/06/28971.php> for the details of this change. Other than this recent change, I can report that the unordered containers seem to work fine with Interprocess. What is the status of the unordered containers library? I don't see it listed on the review schedule. Is it planned to be submitted for formal review any time soon?
Aaron W. LaFramboise wrote:
The 18.03.2007 20:42 upload of unordered.tar.gz in the vault, which as far as I know is the most recent copy of the proposed Boost unordered containers library, is no longer compatible with Ion Gaztañaga's Boost Interprocess. See <http://lists.boost.org/boost-users/2007/06/28971.php> for the details of this change. I've seen the proposal to remove destroy, construct and address, and I think it's a good idea but it seems premature to remove them before the change has even been added to a draft standard. Other than this recent change, I can report that the unordered containers seem to work fine with Interprocess.
What is the status of the unordered containers library? I don't see it listed on the review schedule. Is it planned to be submitted for formal review any time soon? I'll submit it soon, I'm just having some problems running the unit tests with the intel (which might be a problem with my setup) and g++ 3.3 (does anyone still use that?). Once I get them fixed I'm going to submit the library. I haven't had any time this week to work on this, but I should on the weekend.
Daniel
Daniel James wrote:
I've seen the proposal to remove destroy, construct and address, and I think it's a good idea but it seems premature to remove them before the change has even been added to a draft standard.
Yes, it was a bit premature, but I didn't know anyone was using unordered+interprocess. I can add construct/destroy again to the allocators to make them compatible (I'm preparing a new version for the next week), but avoiding construct/destroy in unordered containers might be good idea to clean up a bit the implementation once the removal is added to the draft. Regards, Ion
participants (3)
-
Aaron W. LaFramboise -
Daniel James -
Ion Gaztañaga