Date: Fri, 12 Apr 2019 13:12:26 -0400
On 4/12/19 11:09 AM, Ben Boeckel via Modules wrote:
> On Fri, Apr 12, 2019 at 15:25:02 +0200, Boris Kolpackov wrote:
>>> Do you have new data that would invalidate that recommendation?
>> Google's use of Clang modules, which, if I understand correctly,
>> are pretty much header units.
> AFAIK, this is implicit module stuff. This is understood by those who
> went down this route to be a sub-optimal solution and explicit modules
> are just better.
I don't think one if objectively better than the other; there are
reasonable trade offs. Explicit module management can out-perform
implicit module management, but comes with the cost of, obviously,
having to actually explicitly mange the modules. For many projects,
implicit modules performs well enough and are easily deployed without
the additional build system costs. I don't think we should be
dismissive of either strategy.
Tom.
> On Fri, Apr 12, 2019 at 15:25:02 +0200, Boris Kolpackov wrote:
>>> Do you have new data that would invalidate that recommendation?
>> Google's use of Clang modules, which, if I understand correctly,
>> are pretty much header units.
> AFAIK, this is implicit module stuff. This is understood by those who
> went down this route to be a sub-optimal solution and explicit modules
> are just better.
I don't think one if objectively better than the other; there are
reasonable trade offs. Explicit module management can out-perform
implicit module management, but comes with the cost of, obviously,
having to actually explicitly mange the modules. For many projects,
implicit modules performs well enough and are easily deployed without
the additional build system costs. I don't think we should be
dismissive of either strategy.
Tom.
Received on 2019-04-12 19:12:30