| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Make builds able to depend on external projects
Builds are now able to load external projects and depend on them
as if they were local submodules.
`import $file.external.path.build`
* Disambiguate "dest" for foreign modules.
* Calling modules loaded from external directories "Foreign" to avoid
conflicting with the already existing concept of "ExternalModule".
* Amended the way `dest` is computed for foreign modules
* Added tests to check that the source paths and dest are as expected
* Added a test to show that local modules do not conflict with foreign
modules when they are named the same
* WIP windows build fail
* Added bootstrapping step in CYGWIN CI job
* * Revert externalOutPath deletion
* Add documentation for foreign-modules
* reverting appveyor changes
* Disabling Foreign modules tests against Java9
See https://github.com/lihaoyi/mill/issues/302
|
| |
|
|
|
|
| |
because initializing it more than once within a single process causes crashes
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Moved the `System.getenv` side effect to the end of the world
* Adds a test to make sure that the `Map[String, String]` gets
propagated correctly
* Adds a `Ctx.Env` trait for consistency
https://github.com/lihaoyi/mill/issues/257
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Since Mill now executes in a long-lived JVM, the builds do not have a
chance to use environment variables as inputs. This propagates the
environment variables from the client all the way down to the context
available to the tasks as a `Map[String, String]` so that they can be
used as inputs should the user choose to do so.
https://github.com/lihaoyi/mill/issues/257
|
| |
| |
| |
| | |
- Make `Evaluator` robust against `null` results
|
|/ |
|
| |
|
| |
|
| |
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* fix class loading for inprocess in java 9
* always add rt jar in case of java 9
* move all custom classloading logic into mill.util.ClassLoader
* add comments explaining parent class loader changes and comment in tests pointing to original issue
|
|/ |
|
| |
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Upgraded Ammonite to 1.0.5-7-f032887 that includes a Java 9 performance optimization for mill's client-server mode (lihaoyi/Ammonite#773).
* Java 9 optimization for interactive mode by caching rt.jar.
* Refactored based on @lihaoyi's comments on #239.
* Removed client side rt.jar caching.
* Propagatio of home path from Cli.
* Propagation, propagation, propagation!
|
|/ |
|
|
|
|
|
|
| |
* Fixes for client/server mode on Java 9.
* Upgraded ammonite to 1.0.5-4-c0cdbaf.
|
|
|
|
|
|
|
|
|
|
| |
9 (#217)
* Append "exported rt.jar" when creating URLClassLoader.
* Disabled tests involving Scala 2.10 & 2.11 and nashorn.
* Use java.net.URLClassLoader.
|
|
|
|
| |
make it easier to see where your mill evaluation time is going
|
|
|
|
| |
execution plan of running the `foo.bar` task looks like without actually evaluating it.
|
|
|
|
| |
badly in URIs and other places
|
|
|
|
| |
Also re-write `Jvm.createAssembly` to allow incremental assembly construction. This should allow much faster assembly creation in the common case where upstream dependencies do not change
|
|
|
|
|
|
| |
necessary.
Now we only create them if a Task asks for `T.ctx().dest` or logs something. The vast majority of tasks just do plumbing and neither log output nor create files in `dest`, and this avoids the unnecessary overhead of creating all those un-used files, folders & stack trace
|
| |
|
| |
|
|
|
|
| |
are de-serialized
|
| |
|
|
|
|
|
|
|
|
| |
file-sockets
Seems to work well enough for interactive scala consoles, though still not Ammonite
Also Added ScalaModule#launcher and re-worked our build.sc file to use it
|
| |
|
| |
|
| |
|
|
|
|
| |
Added test cases for new "Did you mean...", "Try `mill.resolve ...`" hints
|
| |
|
|
|
|
| |
Often the main output stream already ends with a newline, so only print a newline in the case where it hasn't
|
| |
|
|
|
|
|
| |
- Add a version of `Resolve` that resolves names only, but works on entrypoints/etc. without needing arguments
- Fix tests to compile with new `multiSelect` parsing flag
|
|
|
|
| |
argument parsing
|
| |
|
| |
|
|
|
|
| |
GenIdea by making it take an Evaluator as an argument
|
| |
|
| |
|
| |
|
|\ |
|
| |
| |
| | |
this used to work but stopped, I guess, after some on segments
|
|/
|
|
| |
our custom releaseCI/releaseManual commands
|