Getting ready for a GHC 8.8.2 port? (no rush yet)

Viktor Dukhovni ietf-dane at dukhovni.org
Tue Jan 28 06:48:37 UTC 2020


On Tue, Jan 28, 2020 at 10:34:42AM +0400, Gleb Popov wrote:

> On Tue, Jan 28, 2020 at 9:03 AM Viktor Dukhovni <ietf-dane at dukhovni.org>
> wrote:
> 
> > While presently I'm using my own build of GHC 8.8.2 from git, and thus
> > can wait indefinitely for a FreeBSD port of same, I should note that
> > stackage nightly snapshots are based on GHC 8.8.2 as of 2020-01-24, and
> > there will presumably be a GHC 8.8.2 LTS at some point in the not too
> > distant future.
> >
> > So it may soonish be a good time to consider releasing a GHC 8.8 port.
> >
> 
> I'm slowly working on this, however it requires updating bootstrap compiler
> - I get a strange hang when building 8.8 using 8.4.3. And newer bootstrap
> have their own problems, so I'm figuring them out.

My build includes four post-release commits:

    $ git log --oneline ghc-8.8.2-release..HEAD
    0c81de31db (HEAD -> build-8.8) Remove GHC_STAGE guards from MachDeps
    707ac619a5 TcIface: Fix inverted logic in typechecking of source ticks
    8b193877fe Fix overflow.
    4d59e883b4 Make CorePrep.tryEtaReducePrep and CoreUtils.tryEtaReduce line up

the "HEAD" commit could perhaps help.  My build.mk file is basically
like the sample one with:

    $ diff -U0 mk/build.mk.sample ../ghc-build/build.mk
    --- mk/build.mk.sample  2020-01-16 01:19:52.708063000 -0500
    +++ ../ghc-build/build.mk       2019-09-29 22:40:51.355059000 -0400
    @@ -23 +23 @@
    -#BuildFlavour = quick
    +BuildFlavour = quick
    @@ -77 +77 @@
    -#V=0
    +V=0
    @@ -104 +104 @@
    -#libraries/integer-gmp_CONFIGURE_OPTS += --configure-option=--with-intree-gmp
    +libraries/integer-gmp_CONFIGURE_OPTS += --configure-option=--with-intree-gmp
    @@ -111 +111,3 @@
    -STRIP_CMD = :
    +#STRIP_CMD = :
    +
    +HaveDtrace = NO

-- 
    Viktor.


More information about the freebsd-haskell mailing list