Boost logo

Boost :

Subject: Re: [boost] "peer reviewed" - Rights and responsibilities of maintainers
From: Robert Ramey (ramey_at_[hidden])
Date: 2018-10-16 17:03:21

On 10/16/18 8:15 AM, Alexander Grund via Boost wrote:

> But Robert insists that #111 "does not show anything at all. It's ill
> conceived."
He does not acknowledge that Boost.Serialization makes the
> application crash
and does not bring any argument why it would be "ill
> conceived".
I do not understand this. It obviously crashes. This is a
> fact easily reproducible by anyone on linux.

If "It" refers to the test_dll_exported on clang - True
else False

If the test misuses C++
> (UB, ...) or the library, then where and how? If not,

If the test refers to your recent test - you test doesn't test anything.

If the test refers to the serialization library test_export_dll on clang
(with which version of standard library - who knows?) there is a
problem. That is why I made such a test after all. So True

why is it seen as
> "ill conceived"?

It's based on a mis-understanding of how the serialization uses the
singleton and of how the singleton works.
> > My version of the patch captures the essence of the PR while
> retaining the original intent of the code.
> This is simply wrong.


If it would #110 and #111 would pass.


It fails to
> fix the `is_destroyed` function or the core problem with destruction order

I see no evidence of this.
> > It's much simpler and alters many less files.

Verifiably true by inspection.
> My fix alters 1 file to fix `is_destroyed` and 4 more to remove a single
> line with an assertion which doesn't hold as shown in #111. Most of the
> changes are added reverts to an earlier, working version and comments on
> why and how stuff works, so it doesn't get accidentally broken by
> optimization/reduction efforts as happened in 1.65
> > I have no idea why I am being criticized for making this patch.
> Because a) it combines unrelated changes,

If a) refers to my patch -> Verifiably True by inspection

hiding the fix and b) pushing
> it w/o review by the original patch author who might have reasonable
> arguments why it isn't enough or "the essence of the PR". I learned it
> is your right to do that.

For good reason.

But then again it is my right to to tell you
> that your claim is wrong.


As I as the author of the patch might
> understand the patch better, it is quite possible, that I'm right, isn't
> it?

It's possible - but after having considered that possibility, I've
concluded that it's False

Robert Ramey
> _______________________________________________
> Unsubscribe & other changes:

Boost list run by bdawes at, gregod at, cpdaniel at, john at