Regression in 2.30 for MinGW constructors/destructors

[Martin Storsjö]

On Mon, 9 Jul 2018, Nick Clifton wrote:

> Hi JonY,
>
>>> It seems like the following commit introduced a regression for
>>> constructors for MinGW:
>>>
>>> commit ca6f2be7f6bc638fd4fad48def1fae4ae4d7906e
>>> Author: Nick Clifton <nickc@redhat.com>
>>> Date:Â Â  Thu Aug 10 13:17:12 2017 +0100
>>>
>>> Â Â Â  Add support for PE startup files that want to define their own
>>> constructor and desctructor sequencing.
>>>
>>> Â Â Â Â Â Â Â  * scripttempl/pep.sc: Enclose __CTOR_LIST__, ___CTOR_LIST__,
>>> Â Â Â Â Â Â Â  __DTOR_LIST__ and ___DTOR_LIST__ in PROVIDE statements so that
>>> Â Â Â Â Â Â Â  they can be overrridden by crt input files.
>>> Â Â Â Â Â Â Â  * scripttempl/pe.sc: Likewise.
>>>
>>> After this commit, setups with existing versions of mingw-w64 no longer
>>> run constructors.
>
>> Nick, any ideas?
>
> Could this PR be the key:
>
>  https://sourceware.org/bugzilla/show_bug.cgi?id=22762
>
> If so then the upcoming 2.31 release will contain the fix that you need.
> Or you could use the 2.30 branch sources.  But not the 2.30 release
> tarball, as that was created just before the patch went into the branch.

Thanks; this looks indeed like the exact issue, and sorry for the 
duplicate report.

I'll add a few further thoughts/comments on the matter on that PR.

// Martin