Re: [boost] [intrusive] Possible bug in bstree_algorithms.hpp

Subject: Re: [boost] [intrusive] Possible bug in bstree_algorithms.hpp
From: Lars Hagström (lars_at_[hidden])
Date: 2014-09-07 13:38:53

Yeah, I think that seems a likely explanation.
I'm not sure that I'll be able to produce something minimal that exhibits
the problem. But the product I'm working on is open source, and the problem
occurs during the build process of the code, so depending on your
environment it should be fairly easy to replicate.

But before we go down that path I'll have a look at it and see if I can
find anything. Is there anything special that I should look at?

/Lars

On Sun, Sep 7, 2014 at 11:24 AM, Ion Gaztañaga <igaztanaga_at_[hidden]> wrote:

> El 06/09/2014 22:09, Lars Hagström escribió:
>
>> Hi,
>>
>> I've just upgraded to boost 1.56 and I've run into a problem. I get an
>> infinite loop inside insert_equal_lower_bound_check(...) in
>> bstree_algorithms.hpp.
>> I'm using that file indirectly through Boost.Interprocess.
>>
>> I've not been able to penetrate exactly what that function does or whether
>> there is indeed a bug in it. But I did look at the version that was in
>> 1.55
>> (called insert_equal_check_impl(...) there), and to me it looks like there
>> is at least one unexpected change:
>>
>
> I think the bug might be earlier, it seems that tree invariants were
> somehow broken and effects appear later in this function.
>
> Any chance to get a test case with the problem?
>
> Best,
>
> Ion
>
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/
> mailman/listinfo.cgi/boost
>