Allow duplicate braces in allman style

[wilzbach]

It seems like I have tested Phobos yesterday with a wrong dscanner binary.
However, there's only one small need for amendment that I found.

Apparently e.g in std/experimental/allocator/common.d this brace style seems to be popular for static if:

        static if (hasMember!(A, "alignedAllocate"))
        {{
             auto b3 = a.alignedAllocate(1, 256);
             assert(b3.length <= 1);
             assert(b3.ptr.alignedAt(256));
             assert(a.alignedReallocate(b3, 2, 512));
             assert(b3.ptr.alignedAt(512));
             static if (hasMember!(A, "alignedDeallocate"))
             {
                 a.alignedDeallocate(b3);
             }
        }}

[wilzbach]

FYI in case you are interested: the PR at Phobos to prepare for enabling the allman check: dlang/phobos#5480

[ghost]

what's the point with this thing ? Don't know how to call it btw immediate scope ?

[wilzbach]

what's the point with this thing ?

It's a "pattern" that appears in Phobos ...

Don't know how to call it btw immediate scope ?

Good idea.

Btw we seem to have an issue with the non-DUB builds & dmd-nightly:

containers/src/containers/treemap.d(33): Error: field tree must be initialized in constructor
containers/src/containers/treemap.d(166): Error: template instance containers.treemap.TreeMap!(int, int, StatsCollector!(FreeList!(AllocatorList!(Factory, GCAllocator), 64LU, 64LU, cast(Flag)false), 262143LU, 0LU)*, "a < b", false, 64LU) error instantiating

I think DUB managed to evade this by not trying to build it and just passing the containers as input directory?

[wilzbach]

Btw we seem to have an issue with the non-DUB builds & dmd-nightly:

See also: https://github.com/economicmodeling/containers/issues/80

(it's a regression from dlang/dmd#6816, similarly to dlang-community/D-YAML#66)

[ghost]

based on the previous comment i guess this is ok to merge.