Re: [Boost-maint] [boost-maint][concept_check] Pull request

Subject: Re: [Boost-maint] [boost-maint][concept_check] Pull request
From: Daniel James (dnljms_at_[hidden])
Date: 2014-02-19 06:25:29

On 19 February 2014 10:09, Ahmed Charles <acharles_at_[hidden]> wrote:
> ----------------------------------------
>> Date: Wed, 19 Feb 2014 09:24:16 +0000
>> From: dnljms_at_[hidden]
>> To: boost-maint_at_[hidden]
>> Subject: Re: [Boost-maint] [boost-maint][concept_check] Pull request
>>
>> On 19 February 2014 09:08, Ahmed Charles <acharles_at_[hidden]> wrote:
>>>> Date: Wed, 19 Feb 2014 08:50:14 +0000
>>>> From: dnljms_at_[hidden]
>>>> To: boost-maint_at_[hidden]
>>>> Subject: Re: [Boost-maint] [boost-maint][concept_check] Pull request
>>>>
>>>> On 19 February 2014 06:46, Marshall Clow <mclow.lists_at_[hidden]> wrote:
>>>>>
>>>>> I think we (Boost) are committed to merging Stephen's changes to master.
>>>>
>>>> We're not really, he made them after being asked not to. It's up to
>>>> individual libraries' maintainers whether they want to use them. IMO
>>>> what we should do with modules where no one is dealing with changes on
>>>> develop (all of them, not just Stephen Kelly's) is revert them in
>>>> develop and create feature branches so that we can get master and
>>>> develop in sync, but keep a record of the outstanding changes.
>>>
>>> I'd rather take a less heavy handed approach and determine if the changes in develop are easily merged to master and avoid having the changes potentially get lost.
>>
>> It's not just a case of whether they can be merged or not. Some are
>> large and potentially disruptive to fairly arcane but stable code.
>> Some changes remove headers which is problematic, even if they are in
>> 'detail' directories. And there are people still using old compilers
>> who rely on the support for them.
>
> I was talking about all changes on develop that aren't on master, did you think I was talking about Stephen's changes exclusively? Just want to make sure we're talking about the same thing.

I'm talking about both at once, so perhaps a bit muddled. We've had
these issues for a long time. I've often thought of suggesting
something similar in the past.