You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Aug 2, 2020. It is now read-only.
(I haven't found a better title so far, I'm all ears if you have a better one.)
The point of this issue is to collect/track all the "tension points" between the way hadrian does things and the way the make build system and all its infrastructure (general ghc usage, tests, nofib, ...) work. In particular, we have seen some issues due to moving some binaries out of lib/bin/ to bin/ (which is why we currently have exceptions in place for unlit and touchy, because we'd otherwise have to patch GHC to tell it to look for those binaries under bin/ when it doesn't see them under lib/bin/).
In #592 (comment) I summed up why I think it's fine that we do things differently, and nobody has yelled at me so far so at least a few people agree with me on this, or are at least open to seeing where I'm offering to take this.
Please feel free to mention all those situations where something in the existing build system or in GHC itself has some expectation that hadrian does not match. We can then start addressing those issues one by one in a principled way and discuss the general direction we want to take with respect to layout of binaries, binary distributions, etc.
The text was updated successfully, but these errors were encountered:
snowleopard
changed the title
bin/ vs lib/bin/ and other mismatches between hadrian and make
bin/ vs bin/lib/ and other mismatches between hadrian and make
May 16, 2018
snowleopard
changed the title
bin/ vs bin/lib/ and other mismatches between hadrian and make
bin/ vs lib/bin/ and other mismatches between hadrian and make
May 17, 2018
(I haven't found a better title so far, I'm all ears if you have a better one.)
The point of this issue is to collect/track all the "tension points" between the way hadrian does things and the way the make build system and all its infrastructure (general ghc usage, tests, nofib, ...) work. In particular, we have seen some issues due to moving some binaries out of
lib/bin/
tobin/
(which is why we currently have exceptions in place forunlit
andtouchy
, because we'd otherwise have to patch GHC to tell it to look for those binaries underbin/
when it doesn't see them underlib/bin/
).In #592 (comment) I summed up why I think it's fine that we do things differently, and nobody has yelled at me so far so at least a few people agree with me on this, or are at least open to seeing where I'm offering to take this.
Please feel free to mention all those situations where something in the existing build system or in GHC itself has some expectation that hadrian does not match. We can then start addressing those issues one by one in a principled way and discuss the general direction we want to take with respect to layout of binaries, binary distributions, etc.
The text was updated successfully, but these errors were encountered: