Subject: Re: [boost] Range Adapters (map_keys, map_values) cause undefined behavior (segv, etc) when applied to R-Values, especially in the context of BOOST_FOREACH
From: Nathan Ridge (zeratul976_at_[hidden])
Date: 2014-02-01 01:21:43
> From: dberris_at_[hidden]
> On Thu, Jan 30, 2014 at 12:41 PM, Nathan Ridge <zeratul976_at_[hidden]> wrote:
>> This is a known issue: see https://svn.boost.org/trac/boost/ticket/7630.
>> (That bug talks about C++11 range-based for loops, but the issue with
>> BOOST_FOREACH is the same).
>> I'm not sure what the status of this bug is. Jeffrey Yasskin posted a
>> patch to that bug which I reviewed over email, but I haven't heard
>> from him about it since.
> Why was the review not done on the bug?
Jeffrey asked me to review it over email and I responded in kind.
I expected that he would post a follow-up patch that addresses the
points made in the review shortly thereafter.
I posted my review comments in the bug now.
> I'm tracking that bug and
> would like to see it resolved, and it seems the patch is just fine.
> Are there any issues still that need to be addressed elsewhere? What
> can I do to help expedite this process?
I think that the approach is fine. The suggestions I make in my review
are fairly minor (nonetheless I would like to see them addressed).
We should also make sure Neil is on board with this approach. Since
this is a significant change, I would like to have his approval before
committing anything. (Now that we use Git should I be saying "before
pushing anything"? (-: ).
Finally, once we know the approach has consensus, someone needs to
replicate the patch for all the adaptors (Jeffrey's patch is for
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk