add new sbt+zinc, remove old sbt+zinc #264
Comments
|
I'm going to be working on several aspects from the modules splitting of sbt, so I can do this. |
|
sbt/zinc depends on a number of dependencies including more sbt modules. We intend to setup a dbuild for sbt modules (and later plugins) for the purpose ensuring that the latest sbt continues to work (as opposed to the community build which is ensuring the libraries continue to work with changes to Scala). Once we have that we can start more aggressively making changes across modules in preparation for sbt 1.0, but that might leave moments of cross-module breakages. As such we wouldn't want to the community build to suffer, so this issue might be open for a while.. |
|
we also discussed this at #290. summary: tracking sbt 1.0 master is too fragile for now, because of module interdependencies, but we could build from a tag, and then try to remember to push that tag forward from time to time in particular, it sounds like we can go ahead and add sbt/sbt.git#v1.0.0-M4, along with its modules including sbt/zinc — the tags for the individual modules can be determined by looking at the version numbers in sbt's |
SethTisue
changed the title
|
@dwijnand what do you make of these errors: at https://scala-ci.typesafe.com/job/scala-2.12.x-integrate-community-build/629/consoleFull ? the changes I was testing are SethTisue@93f4fe2 it seems like dbuild must be using an old sbt somehow, but I've checked carefully that we're specifying 0.13.12 everywhere, and I see no other evidence in the log that an older version is being used |
|
It's the very confusing sbt/sbt#1696 which I'm sure you've found. Haven't looked into why or how it's happening. |
yeah I couldn't tell if it was a real issue or just people confused about what version they're running |
there's no great benefit to compiling such an old version of sbt on such a new version of Scala, and it's fragile. we'll re-add sbt its modules (including the new zinc) when the time comes; it's scala#264
there's no great benefit to compiling such an old version of sbt on such a new version of Scala, and it's fragile. we'll re-add sbt its modules (including the new zinc) when the time comes; it's scala#264
much of which was commented out anyway there's no great benefit to compiling such an old version of sbt on such a new version of Scala, and doing so is fragile. fear not, we'll re-add sbt and its modules (including the new zinc) when the time comes; it's scala#264
much of which was commented out anyway there's no great benefit to compiling such an old version of sbt on such a new version of Scala, and doing so is fragile. fear not, we'll re-add sbt and its modules (including the new zinc) when the time comes; it's scala#264
|
sbt 1.0 is now a lot closer to being a reality, so it's time or almost time to proceed with this. I asked @dwijnand about it just now and he agreed. as we already said earlier, it will probably be necessary to use tags rather than track branches. |
fixes scala#264
|
PR: #581. interested parties please subscribe |
typesafehub/zinc is obsolete and we don't need to be building it anymore
and we should be building sbt/zinc, where the action is these days
The text was updated successfully, but these errors were encountered: