commit db98782e890b3cebf75a19b2da3a1bdcbee17fa4
Author: hololeap <hololeap@users.noreply.github.com>
Date: Fri May 12 06:28:08 2023 -0600

*/*: Remove [profile?] from ghc RDEPEND

This reverts commit b2f01bbf1333c88785109bebf75d49a3bd623727.

Adding [profile?] universally causes issues with exe-only packages with
no profile USE flag. This is better handled in the eclass.

Signed-off-by: hololeap <hololeap@users.noreply.github.com>

commit a2f524e74d7e5e5000c1d057a74dcd7f4af7b43c
Author: hololeap <hololeap@users.noreply.github.com>
Date: Sat May 6 17:04:20 2023 -0600

*/*: Remove ~x86 (and related) from KEYWORDS

x86 is mostly untested, and even building a subset of ::haskell
revealed failing tests which pass on amd64.

With the new behavior of hackport to only add ARCH to KEYWORDS, there
will be CI failures if a package has ~x86 removed but not from its
revdep tree as well. The only sensible mitigation is to remove ~x86
tree-wide until we have a dedicated tester willing to fix current bugs.

See: https://github.com/snoyberg/mono-traversable/issues/210
Signed-off-by: hololeap <hololeap@users.noreply.github.com>

commit b2f01bbf1333c88785109bebf75d49a3bd623727
Author: hololeap <hololeap@users.noreply.github.com>
Date: Fri May 5 15:53:40 2023 -0600

*/*: Add [profile?] to ghc RDEPEND

Any package that depends on 'base' will require the 'base' profiling
libs when profiling is enabled. Because virtually every package depends
on 'base', it makes sense to add [profile?] universally to
'dev-lang/ghc' in RDEPEND.

Reported-by: Filip Kszczot <filip@kszczot.pl>
Signed-off-by: hololeap <hololeap@users.noreply.github.com>

commit 0821308e1cfd9a96f9dbad241ec11873d61a6d06
Author: Miezhiko <Miezhiko@gmail.com>
Date: Sun Apr 16 13:16:34 2023 -0600

dev-haskell/di: new package, add 1.3

Signed-off-by: Miezhiko <Miezhiko@gmail.com>