Boost logo

Boost :

Subject: [boost] A possible date for dropping c++03 support
From: Mike Dev (mike.dev_at_[hidden])
Date: 2018-08-31 03:25:53

Sorry for all the bad line-breaks in my previous message.
I don't have access to a properly configured email client right now.
This should be easier to read:

====== Original Message ==============

Please, let's not derail the discussion just after some consensus has emerged and just before it actually produces
actionable items. Let's not make the better the enemy of the good here.

I made the suggestion to commonly drop c++03 support in the hope that it would actually lead to some code
simplifications and reduced coupling between the boost libraries - even if only in their implementation (if someone
wanted to do a real c++14 reboot of their library, no one is/was stopping them anyway).

>From the survey, I'd say interest in actually applying c++11 features in libraries that a currently supporting 03 is
already relatively low, but big enough that I believe this will actually lead to some changes other than just changing
the default
test settings.
That is with a standard almost universally supported by compilers in use today. Even though the move from 11 to 14 is
much less dramatic, c++14 support is less wide spread and outside of the big three, there are many compilers where even
the most recent
version doesn't support c++14, so I just don't see people that are currently maintaining c++03 support in their
libraries suddenly starting to use c++14.

All this doesn't preclude a discussion about if and when to move to c++14 or a general policy about LTS versions and the
like, but please keep it separate unless you are really, really confident it will lead to something actionable earlier
than what is currently proposed.
I have lots of Ideas, how boost could be reformed and/or modernized but in the end it comes down to convince the
individual library maintainers to actually do something. I think the current proposal will at least spark some action.
Let the announcement to move to cmake be a warning sign about realistic expectations about the evolution of boost. Maybe
I'm aiming too low, but again, if you want more, please open a separate thread and most likely, whatever you come up
with there will not be mutually exclusive with what is proposed here.

Regarding the test settings: If c++11 is supposed to be the lowest official supported version then that is what should
be tested Anything else doesn't make sense (you can't claim to support c++11 if you don't test if it actually works
there). In any case, as was mentioned before, for the announcement itself we don't need give any specifics about
something that is going to happen 1-1,5 years from now. How about something like " c++03 mode/compilation will no longer
be tested / part off the general test/suit"?



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