GHC issueshttps://gitlab.haskell.org/ghc/ghc/-/issues2024-03-21T09:39:17Zhttps://gitlab.haskell.org/ghc/ghc/-/issues/24548configure ignores LDFLAGS2024-03-21T09:39:17ZRodrigo Mesquitaconfigure ignores LDFLAGS## Summary
It seems that `./configure` ignores `LDFLAGS` (whereas `ghc-toolchain` doesn't).
## Steps to reproduce
`LDFLAGS="-Lsomethingbogus" ./configure`
You should get
```
There are some differences between the toolchain configure...## Summary
It seems that `./configure` ignores `LDFLAGS` (whereas `ghc-toolchain` doesn't).
## Steps to reproduce
`LDFLAGS="-Lsomethingbogus" ./configure`
You should get
```
There are some differences between the toolchain configured by "configure" (hadrian/cfg/default.target) and the toolchain configured by the "ghc-toolchain" program (hadrian/cfg/default.target.ghc-toolchain).
20c20
< { ccLinkProgram = Program { prgPath = "/usr/bin/gcc" , prgFlags = ["--target=arm64-apple-darwin","-Wl,-no_fixup_chains","-Wl,-no_warn_duplicate_libraries"] }
---
> { ccLinkProgram = Program { prgPath = "/usr/bin/gcc" , prgFlags = ["--target=arm64-apple-darwin","-Lsomethingbogus","-Wl,-no_fixup_chains","-Wl,-no_warn_duplicate_libraries"] }
```
## Expected behavior
`./configure`-based toolchain configuration should also consider `LDFLAGS` into the cclinker program.
I suppose this is not a priority because `ghc-toolchain` is already doing this right and we ultimately want to move toolchain configuration to it. It is yet another datapoint to hurry on the move (BTW, I think it has been two releases since ghc-toolchain is available opt-in... perhaps the time to make the move is drawing close?).
## Environment
* GHC version used: HEAD
Optional:
* Operating System: macOS
* System Architecture: aarch64https://gitlab.haskell.org/ghc/ghc/-/issues/24531Runtime failure on windows with zlib-0.7.0.0 when missing zlib library2024-03-12T15:28:23ZHannes SiebenhandlRuntime failure on windows with zlib-0.7.0.0 when missing zlib libraryFirst reported by https://github.com/haskell/zlib/issues/65
GHC doesn't fail compilation of a binary depending on `zlib-0.7.0.0`, although the linking of the zlib system depedency failed. The issue exists at least on GHC 9.2.8 through G...First reported by https://github.com/haskell/zlib/issues/65
GHC doesn't fail compilation of a binary depending on `zlib-0.7.0.0`, although the linking of the zlib system depedency failed. The issue exists at least on GHC 9.2.8 through GHC 9.6.
There is no message when running `cabal build`, but opening a repl session fails with the message:
```
ghc-9.4.8.exe: Could not load `zlib1.dll'. Reason: addDLL: zlib1.dll or dependencies not loaded. (Win32 error 126)
```
I was able to reproduce the issue on Win 10, GHC 9.4.8 and cabal 3.10.2.1, but was unable to reproduce the issue on any windows GitHub CI runners.
I briefly investigated, the main change seems to be that zlib haskell library switched to using `extra-libraries` on windows, which should alwasy exist as GHC ships with zlib on windows. Likewise, I verified the library exists, too.Ben GamariBen Gamarihttps://gitlab.haskell.org/ghc/ghc/-/issues/24356Install bash completion file2024-03-05T09:53:22ZBryan Rbryan@haskell.foundationInstall bash completion file#20802 included ghc.bash in the bindist, but it might be nice to actually install it with `make install`, too. At the very least, that would make the script available to GHCup users, though they would still need to manually source it in ...#20802 included ghc.bash in the bindist, but it might be nice to actually install it with `make install`, too. At the very least, that would make the script available to GHCup users, though they would still need to manually source it in their rc files. (Plus, they would need to know where to find it.)https://gitlab.haskell.org/ghc/ghc/-/issues/24276The global package db directory of GHC has duplicate files (*.copy)2024-01-09T15:13:57ZOleg GrenrusThe global package db directory of GHC has duplicate files (*.copy)```
/opt/ghc/9.8.1/lib/ghc-9.8.1/lib/package.conf.d % ll
total 1272
-rw-r--r-- 1 phadej phadej 1788 loka 10 10:56 array-0.5.6.0-88aa.conf
-rw-rw-r-- 1 phadej phadej 1788 loka 10 10:56 array-0.5.6.0-88aa.conf....```
/opt/ghc/9.8.1/lib/ghc-9.8.1/lib/package.conf.d % ll
total 1272
-rw-r--r-- 1 phadej phadej 1788 loka 10 10:56 array-0.5.6.0-88aa.conf
-rw-rw-r-- 1 phadej phadej 1788 loka 10 10:56 array-0.5.6.0-88aa.conf.copy
```
I checked versions 9.4.6, 9.6.2 and 9.8.1. All have these `*.copy` files. The 9.2.8 doesn't have.Matthew PickeringMatthew Pickeringhttps://gitlab.haskell.org/ghc/ghc/-/issues/242009.2.8 armv7 bindist is broken2023-11-22T04:02:21ZJulian Ospald9.2.8 armv7 bindist is broken`make install` of bindist https://downloads.haskell.org/~ghc/9.2.8/ghc-9.2.8-armv7-deb10-linux.tar.xz fails with
```
/usr/bin/install -c -m 755 -d "/root/.ghcup/tmp/ghcup-c90a8b5e/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8"
for i in driver/g...`make install` of bindist https://downloads.haskell.org/~ghc/9.2.8/ghc-9.2.8-armv7-deb10-linux.tar.xz fails with
```
/usr/bin/install -c -m 755 -d "/root/.ghcup/tmp/ghcup-c90a8b5e/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8"
for i in driver/ghc-usage.txt driver/ghci-usage.txt includes/dist/build/settings llvm-targets llvm-passes; do case $i in *.a) /usr/bin/install -c
-m 644 $i "/root/.ghcup/tmp/ghcup-c90a8b5e/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8"; true "/root/.ghcup/tmp/ghcup-c90a8b5e/root/.ghcup/ghc/9.2.8/lib
/ghc-9.2.8"/`basename $i` ;; *.dll) /usr/bin/install -c -m 755 $i "/root/.ghcup/tmp/ghcup-c90a8b5e/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8" ; : "/ro
ot/.ghcup/tmp/ghcup-c90a8b5e/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8"/`basename $i` ;; *.so) /usr/bin/install -c -m 755 $i "/root/.ghcup/tmp/ghcup-c9
0a8b5e/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8" ;; *.dylib) /usr/bin/install -c -m 755 $i "/root/.ghcup/tmp/ghcup-c90a8b5e/root/.ghcup/ghc/9.2.8/lib/
ghc-9.2.8";; *) /usr/bin/install -c -m 644 $i "/root/.ghcup/tmp/ghcup-c90a8b5e/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8"; esac; done
gcc -E -undef -traditional -P -DINSTALLING -DLIB_DIR='"/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8"' -DINCLUDE_DIR='"/root/.ghcup/ghc/9.2.8/lib/ghc-9.2.8/
include"' -DFFI_INCLUDE_DIR= -DFFI_LIB_DIR= '-DFFI_LIB="Cffi"' -DLIBDW_INCLUDE_DIR= -DLIBDW_LIB_DIR= -Iincludes -Iincludes/dist -Iincludes/dist-der
ivedconstants/header -Iincludes/dist-ghcconstants/header -Iincludes/dist-install/build -x c rts/package.conf.in -o rts/dist/package.conf.install.ra
w
grep -v '^#pragma GCC' rts/dist/package.conf.install.raw | sed -e 's/""//g' -e 's/:[ ]*,/: /g' >rts/dist/package.conf.install
make[1]: *** No rule to make target 'utils/unlit/dist-install/build/tmp/unlit', needed by 'install_libexecs'. Stop.
make: *** [Makefile:51: install] Error 2
```
----
Building 9.2.8 from source and trying to package/install it results in linking errors, because this hsc2hs bug has not been backported: https://gitlab.haskell.org/ghc/ghc/-/commit/99623358754d812b8b4bdfcdc57190d38617b9cc
I fixed all of that downstream: https://downloads.haskell.org/~ghcup/unofficial-bindists/ghc/9.2.8/ghc-9.2.8-armv7-deb10-linux-gnueabihf.tar.xzhttps://gitlab.haskell.org/ghc/ghc/-/issues/24129GHC doc build broken by sphinx 72023-11-01T08:10:39ZsterniGHC doc build broken by sphinx 7## Summary
Building the GHC documentation fails with
```
Theme error:
An error happened in rendering the page 9.0.1-notes.
Reason: UndefinedError("'style' is undefined")
```
After upgrading to sphinx 7.2.6.
## Steps to reproduce
Bui...## Summary
Building the GHC documentation fails with
```
Theme error:
An error happened in rendering the page 9.0.1-notes.
Reason: UndefinedError("'style' is undefined")
```
After upgrading to sphinx 7.2.6.
## Steps to reproduce
Build GHC with sphinx 7.2.6, e.g. by building any `ghc*` (non `Binary` ofc) attribute under `haskell.compiler` on nixpkgs `6515b811b82a83f551988dd25ec6ef587fa93ef0`.
Example logs:
- https://hydra.nixos.org/build/239488473/nixlog/1
- https://hydra.nixos.org/build/239551789/nixlog/1
## Expected behavior
Build succeeds.
## Environment
* GHC version used: 9.0.2, 9.4.7
Optional:
* Operating System: NixOS
* System Architecture: x86_64-linux, aarch64-darwinMatthew PickeringMatthew Pickeringhttps://gitlab.haskell.org/ghc/ghc/-/issues/24122Boot library version history table is missing 9.4.x entries for x>42023-10-31T15:04:20ZBryan Rbryan@haskell.foundationBoot library version history table is missing 9.4.x entries for x>4https://gitlab.haskell.org/ghc/ghc/-/wikis/commentary/libraries/version-history only goes up to 9.4.4.https://gitlab.haskell.org/ghc/ghc/-/wikis/commentary/libraries/version-history only goes up to 9.4.4.ZubinZubinhttps://gitlab.haskell.org/ghc/ghc/-/issues/241219.8.1 hadrian installs libraries LICENSE files into /usr/lib64/ghc-9.8.1/shar...2023-10-31T15:02:42ZJens Petersen9.8.1 hadrian installs libraries LICENSE files into /usr/lib64/ghc-9.8.1/share/doc/x86_64-linux-ghc-9.8.1/*/## Summary
## Steps to reproduce
The general Fedora steps are in ghc9.8.spec here:
https://src.fedoraproject.org/rpms/ghc9.8/tree/rawhide
But currently it works around this problem by just removing the said files at the end.
So curren...## Summary
## Steps to reproduce
The general Fedora steps are in ghc9.8.spec here:
https://src.fedoraproject.org/rpms/ghc9.8/tree/rawhide
But currently it works around this problem by just removing the said files at the end.
So currently it is not a high priority - though the file path seems unsatisfactory to me.
## Expected behavior
The LICENSE files are platform dependent so they could be installed in `/usr/lib64/ghc-9.8.1/share/doc/*/` at least.
Normally they should live around `/usr/share/{doc,licenses}/ghc-9.8.1/*/` instead.
## Environment
* GHC version used: 9.8.1
Optional:
* Operating System: Fedora Linux
* System Architecture: x86_649.8.2Ben GamariBen Gamarihttps://gitlab.haskell.org/ghc/ghc/-/issues/24093user_guide.pdf not present in doc tarball2024-03-17T23:54:24ZBen Gamariuser_guide.pdf not present in doc tarballIn 9.8.1 the PDF version of the user-guide did not make it into the doc-tarball which gets uploaded to `downloads.haskell.org`.In 9.8.1 the PDF version of the user-guide did not make it into the doc-tarball which gets uploaded to `downloads.haskell.org`.9.10.1Ben GamariBen Gamarihttps://gitlab.haskell.org/ghc/ghc/-/issues/24028Missing users_guide for the Windows distributions of GHC 9.6.*2024-01-09T10:32:10ZHai / @BestYeenMissing users_guide for the Windows distributions of GHC 9.6.*## Summary
In the past, the users_guide doc directory was shipped with the Windows distro of GHC, and that was very handy.
At the moment, it's not shipped anymore, but there's still a link to it in the shipped docs, which is now dead.
...## Summary
In the past, the users_guide doc directory was shipped with the Windows distro of GHC, and that was very handy.
At the moment, it's not shipped anymore, but there's still a link to it in the shipped docs, which is now dead.
A short discussion about it is here: https://discourse.haskell.org/t/ghc-9-6-3-is-now-released/7682
I'm using this documentation a lot when working with Haskell, and I prefer the docs I use to be easy to obtain for offline use.
Please include it again. It's extremely useful to have. Sorry for being a Windows user in this case. :)9.8.2Matthew PickeringMatthew Pickeringhttps://gitlab.haskell.org/ghc/ghc/-/issues/24007relpath.sh script produces wrong results2023-10-03T21:40:13ZIlias Tsitsimpisrelpath.sh script produces wrong results## Summary
The mk/relpath.sh script is supposed to be a shell implementation of realpath, but produces wrong results.
## Steps to reproduce
```
$ ./mk/relpath.sh /usr/lib/ghc/lib /usr/lib/ghc-doc
..-doc
```
## Expected behavior
```
...## Summary
The mk/relpath.sh script is supposed to be a shell implementation of realpath, but produces wrong results.
## Steps to reproduce
```
$ ./mk/relpath.sh /usr/lib/ghc/lib /usr/lib/ghc-doc
..-doc
```
## Expected behavior
```
$ realpath --relative-to=/usr/lib/ghc/lib /usr/lib/ghc-doc
../../ghc-doc
```9.8.2Ben GamariBen Gamarihttps://gitlab.haskell.org/ghc/ghc/-/issues/23871Testing ghc-toolchain in packaging environments2024-03-21T09:52:05ZMatthew PickeringTesting ghc-toolchain in packaging environmentsWe want to test that ghc-toolchain works in as many situations as possible. This will require building ghc with the `--enable-strict-ghc-toolchain-check` configure flag and verifying that the output produced by ghc-toolchain is the same ...We want to test that ghc-toolchain works in as many situations as possible. This will require building ghc with the `--enable-strict-ghc-toolchain-check` configure flag and verifying that the output produced by ghc-toolchain is the same as the configure script.
Here are a list of potential places where we will want to test this:
- [x] GHC CI itself ( !10976 )
- [ ] Fedora packaging (cc @juhp)
- [ ] Debian packaging
- [ ] Homebrew packaging
- [ ] Haskell.nix packaging
- [ ] nixpkgs packaging
Perhaps people have some other ideas about places where GHC is built from source we should also be testing.Matthew PickeringMatthew Pickeringhttps://gitlab.haskell.org/ghc/ghc/-/issues/23854ghc-toolchain binary not in cross-compiler bindists2023-08-22T14:58:04ZMatthew Pickeringghc-toolchain binary not in cross-compiler bindistsSee: https://gitlab.haskell.org/ghc/ghc/-/jobs/1641836
```
./configure: line 12232: bin/ghc-toolchain-bin: No such file or directory
./configure: line 12401: bin/ghc-toolchain-bin: No such file or directory
./configure: line 12410: bin/...See: https://gitlab.haskell.org/ghc/ghc/-/jobs/1641836
```
./configure: line 12232: bin/ghc-toolchain-bin: No such file or directory
./configure: line 12401: bin/ghc-toolchain-bin: No such file or directory
./configure: line 12410: bin/ghc-toolchain-bin: No such file or directory
./configure: line 12411: bin/ghc-toolchain-bin: No such file or directory
```Matthew PickeringMatthew Pickeringhttps://gitlab.haskell.org/ghc/ghc/-/issues/23828Provide dynamically linked alpine binary distributions.2023-08-15T13:43:14ZOleg GrenrusProvide dynamically linked alpine binary distributions.My primary use for Alpine is to produce staticly linked linux binaries. But I don't care how GHC itself is linked, as long as it can produce (both C and Hs) static binaries.
E.g. I don't need https://gitlab.haskell.org/ghc/ghc/-/issues/...My primary use for Alpine is to produce staticly linked linux binaries. But I don't care how GHC itself is linked, as long as it can produce (both C and Hs) static binaries.
E.g. I don't need https://gitlab.haskell.org/ghc/ghc/-/issues/23043 to be resolved, which as far as I understand is a hard issue, but apparently building dynamically linked alpine distibution is not impossible (@maerwald comment https://gitlab.haskell.org/ghc/ghc/-/issues/23043#note_513809).
I'd welcome the official dynamic alpine bindists.https://gitlab.haskell.org/ghc/ghc/-/issues/23714Release ghc-toolchain and ghc-platform on hackage2023-08-03T18:29:29ZRodrigo MesquitaRelease ghc-toolchain and ghc-platform on hackageWe should release the two recently-introduced packages on hackage
* We should, when !9263, !10903, and the fixes in !10901 land, upload ghc-toolchain to hackage.
* ghc-platform has already landed on master - we should upload it to hackag...We should release the two recently-introduced packages on hackage
* We should, when !9263, !10903, and the fixes in !10901 land, upload ghc-toolchain to hackage.
* ghc-platform has already landed on master - we should upload it to hackage.
See also the related comment in https://gitlab.haskell.org/ghc/ghc/-/merge_requests/10696#note_513029 asking for ghc-platform to be uploaded ASAP.9.10.1https://gitlab.haskell.org/ghc/ghc/-/issues/23615Remove x86_64 deb9 bindists2023-08-24T19:20:19ZMatthew PickeringRemove x86_64 deb9 bindistsUbuntu 18 and Linux Mint 19 are now EOL so debian 9 bindists are on their last legs (deb9 is also now EOL).
If something happens to the deb9 build we should just retire these bindists rather than attempt to fix it again.
cc @maerwaldUbuntu 18 and Linux Mint 19 are now EOL so debian 9 bindists are on their last legs (deb9 is also now EOL).
If something happens to the deb9 build we should just retire these bindists rather than attempt to fix it again.
cc @maerwaldhttps://gitlab.haskell.org/ghc/ghc/-/issues/23608$tooldir variable in settings should not exist2023-08-23T17:44:01ZMatthew Pickering$tooldir variable in settings should not existAfter a call with @alt-romes we discussed why 5d76846405240c051b00cddcda9d8d02c880968e introduced support for the `$tooldir` variable in a settings file.
It seems that this support for finding the tooldir (see `findToolDir` which looks...After a call with @alt-romes we discussed why 5d76846405240c051b00cddcda9d8d02c880968e introduced support for the `$tooldir` variable in a settings file.
It seems that this support for finding the tooldir (see `findToolDir` which looks upwards two directories) was implemented as workaround for the fact that the bundled mingw toolchain exists in two diffent places during a build and in the bindist.
1. During a build `mingw` exists at `$topdir/../../mingw` because the bundled toolchain is placed in `_build/mingw` rather than `_build/stageN/mingw`, and topdir is `_build/stageN/lib`.
2. In a bindist the `mingw` folder is in `$topdir/../mingw`
It seems that a better solution is to have a different settings file for build time and install time
* The settings file during a build points to `$topdir/../../mingw`
* The settings file in the bindist points to `$topdir/../mingw`
This situation mirrors the same situation as with linux builds where we have one settings file which is used during the build process and then a generate a different settings file at install time. The difference is that on windows we can generate this "install time" settings file at build time because it's the same for all installations.
cc @alt-romeshttps://gitlab.haskell.org/ghc/ghc/-/issues/23601ghcup-metadata: In nightly pipelines, still generate the metadata even if som...2023-09-25T11:13:57ZMatthew Pickeringghcup-metadata: In nightly pipelines, still generate the metadata even if some of the jobs failAt the moment if one of the jobs fails (ie Centos7) then the metadata generation script fails to run. We should instead for the nightly builds just omit that part of the metadata if the relevant job doesn't exist, so the nightly can stil...At the moment if one of the jobs fails (ie Centos7) then the metadata generation script fails to run. We should instead for the nightly builds just omit that part of the metadata if the relevant job doesn't exist, so the nightly can still be uploaded for all the platforms where things do work.
This would have fall-out effects later on in the pipeline when the metadata we generated is tested (`ghcup-ci`) but if this situation happens then the pipeline will already be failing.
OTOH, this might just be a bad idea because if you are trying to rely on the nightly pipelines it could be a bit unpredictable which bindist ghcup will try and install as if your "normal" bindist isn't available it will try the fallback (rocky8) which may or may not work for your platform.
We should at least requires that certain platforms pass CI (these ones are tested on normal validate pipelines anyway):
* darwin (x86/aarch64)
* windows
* fedora33
* deb10
* rocky8 (the fallback)
cc @maerwald who requested this.Matthew PickeringMatthew Pickeringhttps://gitlab.haskell.org/ghc/ghc/-/issues/23594GHC 9.6 bindists for Aarch64 linux have a broken install of directory package2023-10-10T15:55:22ZAlex MasonGHC 9.6 bindists for Aarch64 linux have a broken install of directory package## Summary
Installing GHC 9.6.1 and 9.6.2 on Aarch64 ubuntu using GHCup leads to a broken installation of the directory package.
## Steps to reproduce
1. Install ghcup
2. `ghcup install ghc 9.6.2` (or 9.6.1)
3. `ghcup set ghc 9.6.2`
4...## Summary
Installing GHC 9.6.1 and 9.6.2 on Aarch64 ubuntu using GHCup leads to a broken installation of the directory package.
## Steps to reproduce
1. Install ghcup
2. `ghcup install ghc 9.6.2` (or 9.6.1)
3. `ghcup set ghc 9.6.2`
4. `ghc-pkg check --verbose`
```shell
> ghc-pkg check --verbose
GHC package manager version 9.6.2
Timestamp 2023-07-02 11:06:19.951587334 UTC for /data/home/ubuntu/.ghcup/ghc/9.6.2/lib/ghc-9.6.2/lib/package.conf.d/package.cache
using cache: /data/home/ubuntu/.ghcup/ghc/9.6.2/lib/ghc-9.6.2/lib/package.conf.d/package.cache
db stack: ["/data/home/ubuntu/.ghc/aarch64-linux-9.6.2/package.conf.d","/data/home/ubuntu/.ghcup/ghc/9.6.2/lib/ghc-9.6.2/lib/package.conf.d"]
flag db stack: ["/data/home/ubuntu/.ghc/aarch64-linux-9.6.2/package.conf.d","/data/home/ubuntu/.ghcup/ghc/9.6.2/lib/ghc-9.6.2/lib/package.conf.d"]
Warning: include-dirs: /data/home/ubuntu/.ghcup/ghc/9.6.2/lib/ghc-9.6.2/lib/../lib/aarch64-linux-ghc-9.6.2/directory-1.3.8.1/include doesn't exist or isn't a directory
```
```shell
> tree $HOME/.ghcup/ghc/9.6.2/lib/ghc-9.6.2/lib/../lib/aarch64-linux-ghc-9.6.2/directory-1.3.8.1/
/data/home/ubuntu/.ghcup/ghc/9.6.2/lib/ghc-9.6.2/lib/../lib/aarch64-linux-ghc-9.6.2/directory-1.3.8.1/
├── System
│ ├── Directory
│ │ ├── Internal
│ │ │ ├── C_utimensat.dyn_hi
│ │ │ ├── C_utimensat.hi
│ │ │ ├── C_utimensat.p_hi
│ │ │ ├── Common.dyn_hi
│ │ │ ├── Common.hi
│ │ │ ├── Common.p_hi
│ │ │ ├── Config.dyn_hi
│ │ │ ├── Config.hi
│ │ │ ├── Config.p_hi
│ │ │ ├── Posix.dyn_hi
│ │ │ ├── Posix.hi
│ │ │ ├── Posix.p_hi
│ │ │ ├── Prelude.dyn_hi
│ │ │ ├── Prelude.hi
│ │ │ ├── Prelude.p_hi
│ │ │ ├── Windows.dyn_hi
│ │ │ ├── Windows.hi
│ │ │ └── Windows.p_hi
│ │ ├── Internal.dyn_hi
│ │ ├── Internal.hi
│ │ ├── Internal.p_hi
│ │ ├── OsPath.dyn_hi
│ │ ├── OsPath.hi
│ │ └── OsPath.p_hi
│ ├── Directory.dyn_hi
│ ├── Directory.hi
│ └── Directory.p_hi
├── libHSdirectory-1.3.8.1.a
└── libHSdirectory-1.3.8.1_p.a
3 directories, 29 files
```
```shell
> cat $HOME/.ghcup/ghc/9.6.2/lib/ghc-9.6.2/lib/package.conf.d/directory-1.3.8.1.conf
...
data-dir:
${pkgroot}/../share/aarch64-linux-ghc-9.6.2/directory-1.3.8.1
hs-libraries: HSdirectory-1.3.8.1
include-dirs:
${pkgroot}/../lib/aarch64-linux-ghc-9.6.2/directory-1.3.8.1/include
depends:
base-4.18.0.0 filepath-1.4.100.1 time-1.12.2 unix-2.8.1.0
...
```
## Expected behavior
I assume the include directory should exist, or the package config shouldn't include the above include-dirs directive:
```shell
> cat $HOME/.ghcup/ghc/9.4.5/lib/ghc-9.4.5/lib/package.conf.d/directory-1.3.7.1.conf
...
data-dir:
${pkgroot}/../share/aarch64-linux-ghc-9.4.5/directory-1.3.7.1
hs-libraries: HSdirectory-1.3.7.1
depends: base-4.17.1.0 filepath-1.4.2.2 time-1.12.2 unix-2.7.3
...
```
## Environment
```shell
> uname -a
Linux ip-172-31-17-153 5.19.0-1025-aws #26~22.04.1-Ubuntu SMP Mon Apr 24 01:58:03 UTC 2023 aarch64 aarch64 aarch64 GNU/Linux
> ghcup --version
The GHCup Haskell installer, version 0.1.19.4
> ghc --version
The Glorious Glasgow Haskell Compilation System, version 9.6.2
```9.8.2Ben GamariZubinBen Gamarihttps://gitlab.haskell.org/ghc/ghc/-/issues/23519RISC-V binary packages2023-06-20T14:37:11ZFrancesco Gazzettafgaz@fgaz.meRISC-V binary packagesAffordable and powerful enough RISC-V machines are starting to pop up. Having official RISC-V binary tarballs would make native bootstrapping easier.Affordable and powerful enough RISC-V machines are starting to pop up. Having official RISC-V binary tarballs would make native bootstrapping easier.