diff options
| author | Herbert Valerio Riedel <hvr@gnu.org> | 2017-09-20 15:16:24 +0200 |
|---|---|---|
| committer | Herbert Valerio Riedel <hvr@gnu.org> | 2017-09-20 15:31:50 +0200 |
| commit | 02ff70563e490d2a7f3141eab7229803c523da57 (patch) | |
| tree | 339c80a390d95e7688ddc7df9987f436000e9d0f /ghc.mk | |
| parent | 11d9615e9f751d6ed084f1cb20c24ad6b408230e (diff) | |
| download | haskell-02ff70563e490d2a7f3141eab7229803c523da57.tar.gz | |
Add 'stm' package to the global package database
This is a preparation for `haskeline` picking up a dependency on `stm`
real soon now. See https://github.com/judah/haskeline/pull/61 for details.
If we figure out a way to not bundle the libraries depended upon by the
GHCi executable in the global package database (see #8919 for the original
reason why we had to start bundling terminfo/haskeline in the first place)
we can get rid of `stm` again...
On the bright side, we were able to avoid uploading new `stm` releases for
over two years already, so it shouldn't cause too much trouble if GHC imposes
a strong preference on the `stm` package's version (this most likely will
mostly affect Linux distributions & similiar).
While at it, this also update the stm submodule to include relaxed
bounds to allow the upcoming base-4.11 version.
Diffstat (limited to 'ghc.mk')
| -rw-r--r-- | ghc.mk | 1 |
1 files changed, 1 insertions, 0 deletions
@@ -470,6 +470,7 @@ else libraries/haskeline_CONFIGURE_OPTS += --flags=-terminfo endif +PACKAGES_STAGE1 += stm PACKAGES_STAGE1 += haskeline PACKAGES_STAGE1 += ghci |