- User Since
- Jun 5 2014, 10:16 AM (250 w, 3 d)
Wed, Mar 13
Sat, Mar 9
Mon, Mar 4
Jan 7 2019
Is the patch on gitlab too? (for CI)
Dec 21 2018
Dec 20 2018
@mpickering Great! This look very encouraging.
So I think that is all the packages?
This is looking pretty good. Besides Andrey's outstanding question, I made a suggestion and have one request to add a TODO.
Dec 17 2018
- address Andrey's feedback
Dec 15 2018
Some of your comments apply to the contents that was already there, and I don't feel like using this diff to freshen up the README entirely. I did remove things that were obviously outdated though. In particular, we haven't migrated some of the issues yet and we therefore don't have a trac ticket to point to yet.
Dec 14 2018
Github and Gitlab both render hadrian's hadrian/README.md as well as the more specific documents under hadrian/doc/, e.g https://github.com/ghc/ghc/tree/master/hadrian#hadrian or https://gitlab.staging.haskell.org/ghc/ghc/tree/master/hadrian#hadrian.
Dec 13 2018
Dec 11 2018
Added some documentation for those new targets and the error reporting @snowleopard requested for "stage 0 executables".
- document the simple package targets, handle stage0 exes
I just updated the patch with a new file, hadrian/doc/testsuite.md, with a bunch of illustrated examples for most of the testsuite options (all the important ones anyway). And hadrian's readme now links to it in the testing section.
- in-tree documentation for the testsuite rule and its options
Also, this should be documented too.
Hmm, right, well the quickstart page being on trac and with the move to gitlab, I guess it's not such a great idea to focus the documentation effort there anyway. It's just absolutely not the direction I've been taking in the past few months. I'll add a few words about what this patch allows in the README and then perhaps start thinking about enhancing the README and doc/ directory with the contents that I ended up putting in the quickstart page. This latter part would be in another patch though of course.
Dec 10 2018
This will do for fixing the build and will give us time to fix this all properly, including making use of platformSupportsSharedLibs and dynamicGhcPrograms in requiresDynamic.
I think there's a bit of work to do to make this work nicely under all circumstances, but that should otherwise do the trick.
Dec 7 2018
- oops, remove leftover
Dec 6 2018
The hadrian part looks good to me. I suggested a set of flags in the ticket, but making the choice through configure is sensible enough and would work for me if everyone's happy with that.
Alright, I think I addressed all of your feedback. How is this patch looking now?
- fix BuildPath docs
- add a note about object files representation
Dec 5 2018
- move common path parsing bits in Hadrian.BuildPath
Dec 4 2018
Thanks for the quick feedback!
- doc improvement
I'm importing the existing "infrastructure" for parsing from Rules.Library but I suspect a refactoring might be nice, putting the common utilities in a dedicated module. I'm not inspired for the name/place of that module, so I went for the shortest route to a working patch.
Dec 3 2018
I only see a tiny problem with your patch, it otherwise looks good to me.
Dec 2 2018
All good then! :-)
Looks good to me, but it'd be good to have @Phyx confirm that this doesn't cause any breakage on Windows. Tamar, any chance you can fire up a build in the upcoming week?
This breaks the default ./hadrian/build.sh on OSX. :(
Nov 29 2018
OK, it's indeed not that straightforward, we'll see what we can do about this in due time then. Thanks for the explanation, Ryan.
I'm marking this as accepted as this is a definite improvement.
Nov 28 2018
I don't see a way, currently.
Nov 27 2018
Nov 26 2018
Yes, it is a bit fragile, I'm sure we will have a better picture of how to refactor it all once things settle down a little bit. In the meantime: LGTM.
Nov 22 2018
- missing ~ in note
Done. I added one note about the way we generate bindists, and one note about wrapper scripts / relocatable GHCs, i.e about the way we consume them. Hopefully things are much clearer now.
- add notes about the bindists in general and the story for wrapper scripts
Wow, that's great!
It would certainly be ideal to eventually have the inverse functions live in the same place, to avoid "silent breakage" when definitions are changed in one place but not the other.
@Phyx Great news, sorry for the initial breakage =)
Nov 21 2018
Looks good to me!
Nov 19 2018
I confirm that with this patch, we can run the testsuite with quickest/quick/perf, building respectively the vanilla, vanilla+dynamic, vanilla+dynamic+profiling "ways" of the iserv program along the way.
Nov 17 2018
There's a diff that bumps the Cabal submodule off the 2.4 branch, to track master: https://phabricator.haskell.org/D5329
Nov 16 2018
Nov 15 2018
Nov 14 2018
Merged in 08cc9b6b2c7f7fdaaaf80361ab84a501f0a573c5. That somehow didn't close the diff automatically, and I can't mark the diff as closed myself.
Nov 12 2018
Modulo the typo in the comment, looks good to me.
Nov 5 2018
- add reference to integer library note in PrelNames
Nov 1 2018
(try pushing this patch to the staging repo)
Oct 31 2018
LGTM. It's a little annoying that you have to define makeRelativeNoSysLink and handle all those things in Hadrian itself. Perhaps could we at least throw that function's definition in a dedicated module (or in Utilities or something). And indeed what about Windows?