Merge branch 'master' into release/graal-vm/1.0

Author: chrisseaton

.gitignore

@@ -29,6 +29,7 @@ lib/mri/openssl.su
 lib/mri/zlib.su
 lib/mri/psych.su
 lib/mri/syslog.su
+lib/mri/nkf.su
 lib/mri/etc.su
 lib/mri/rbconfig/sizeof.su
 /.ext

CHANGELOG.md

@@ -1,4 +1,37 @@
-# 1.0 RC 9
+# 1.0 RC 10
+
+New features:
+
+* The `nkf` and `kconv` standard libraries were added (#1439).
+* `Mutex` and `ConditionVariable` have a new fast path for acquiring locks
+  that are unlocked.
+* `Queue` and `SizedQueue`, `#close` and `#closed?`, have been implemented.
+* `Kernel#clone(freeze)` has been implemented (#1454).
+* `Warning.warn` has been implemented (#1470).
+* `Thread.report_on_exception` has been implemented (#1476).
+
+Bug fixes:
+
+* Added `rb_eEncodingError` for C extensions (#1437).
+* Fixed race condition when creating threads (#1445).
+* Handle `exception: false` for IO#write_nonblock (#1457).
+* `File.expand_path` now raises an exception for a non-absolute user-home.
+* `ArgumentError` messages now better match MRI (#1467).
+* Added support for `:float_millisecond`, `:millisecond`, and
+  `:second` time units to `Process.clock_gettime` (#1468).
+* Fixed backtrace of re-raised exceptions (#1459).
+* Updated an exception message in Psych related to loading a non-existing
+  class so that it now matches MRI.
+* Fixed a JRuby-style Java interop compatibility issue seen in `test-unit`.
+* Fixed problem with calling `warn` if `$stderr` has been reassigned.
+* Fixed definition of `RB_ENCODING_GET_INLINED` (#1440).
+
+Changes:
+
+* Timezone messages are now logged at `CONFIG` level, use `-Xlog=CONFIG` to
+  debug if the timezone is incorrectly shown as `UTC`.
+
+# 1.0 RC 9, 5 November 2018
 
 Security:
 
@@ -10,7 +43,7 @@ New features:
 
 * LLVM for Oracle Linux 7 can now be installed without building from source.
 
-Bug fixes
+Bug fixes:
 
 * Times can now be created with UTC offsets in `+/-HH:MM:SS` format.
 * `Proc#to_s` now has `ASCII-8BIT` as its encoding instead of the
@@ -34,7 +67,7 @@ Changes:
 * Detect when the system libssl is incompatible more accurately and
   add instructions on how to recompile the extension.
 
-# 1.0 RC 8, October 2018
+# 1.0 RC 8, 19 October 2018
 
 New features:
 
@@ -52,11 +85,11 @@ Changes:
 * JCodings has been updated from 1.0.30 to 1.0.40.
 * Joni has been updated from 2.1.16 to 2.1.25.
 
-Performance
+Performance:
 
 * Performance of setting the last exception on a thread has now been improved.
 
-# 1.0 RC 7, October 2018
+# 1.0 RC 7, 3 October 2018
 
 New features:
 
@@ -98,7 +131,7 @@ Changes:
 * Reimplemented exceptions and other APIs that can return a backtrace to use
   Truffle's lazy stacktraces API.
 
-# 1.0 RC 6, September 2018
+# 1.0 RC 6, 3 September 2018
 
 New features:
 
@@ -177,7 +210,7 @@ Changes:
 * The `-Xsync.stdio` option has been removed - use standard Ruby
   `STDOUT.sync = true` in your program instead.
 
-# 1.0 RC 5, August 2018
+# 1.0 RC 5, 3 August 2018
 
 New features:
 

README.md

@@ -24,12 +24,11 @@ There are three ways to install TruffleRuby:
   as a simple binary tarball. This distribution can also be used for
   [testing TruffleRuby in CI](doc/user/standalone-distribution.md).
 
-You can use `gem` to install Gems as normal. TruffleRuby currently requires
-Bundler version `1.16.x`.
+You can use `gem` to install Gems as normal.
 
 You can also build TruffleRuby from source, see the
 [building instructions](doc/contributor/workflow.md), and using
-[Docker](doc/contributor/docker.md)
+[Docker](doc/contributor/docker.md).
 
 ## Aim
 
@@ -45,12 +44,12 @@ TruffleRuby aims to:
 
 ## TruffleRuby configurations
 
-There are two main configurations of TruffleRuby - *native* and *JVM*. It's
+There are two main configurations of TruffleRuby: *native* and *JVM*. It's
 important to understand the different configurations of TruffleRuby, as each has
 different capabilities and performance characteristics. You should pick the
 execution mode that is appropriate for your application.
 
-When distributed as part of GraalVM, TruffleRuby by default runs in the *native*
+TruffleRuby by default runs in the *native*
 configuration. In this configuration, TruffleRuby is ahead-of-time compiled to a
 standalone native executable. This means that you don't need a JVM installed on
 your system to use it. The advantage of the native configuration is that it
@@ -78,7 +77,7 @@ highest possible performance you probably want the *JVM* configuration, by using
 `--jvm`.
 
 At runtime you can tell if you are using the native configuration using
-`TruffleRuby.native?`
+`TruffleRuby.native?`.
 
 You won't encounter it when using TruffleRuby from the GraalVM, but there is
 also another configuration which is TruffleRuby running on the JVM but with the
@@ -128,7 +127,7 @@ You will find that many C extensions will not work without modification.
 ## Migration from MRI
 
 TruffleRuby should in most cases work as a drop-in replacement for MRI, but you
-should read about out [compatibility](doc/user/compatibility.md).
+should read about our [compatibility](doc/user/compatibility.md).
 
 ## Migration from JRuby
 

ci.jsonnet

@@ -289,7 +289,7 @@ local part_definitions = {
       downloads+: {
         JAVA_HOME: {
           name: "labsjdk",
-          version: "8u172-jvmci-0.48",
+          version: "8u192-jvmci-0.52",
           platformspecific: true,
         },
       },
@@ -303,7 +303,7 @@ local part_definitions = {
       downloads+: {
         JAVA_HOME: {
           name: "openjdk",
-          version: "8u172-jvmci-0.48",
+          version: "8u192-jvmci-0.52",
           platformspecific: true,
         },
       },

doc/contributor/third-party-assumptions.md

@@ -1,13 +1,17 @@
 # Assumptions Third Party Code Makes About TruffleRuby
 
-When third-party code starts to make assumptions about the way TruffleRuby does
-something, such as a `TruffleRuby` method, or the layout of our distribution, we
-record it here so that we know who is using it if we need to change something
-about it.
+Third-party code should never use any APIs apart from standard Ruby, and those
+listed in [TruffleRuby additions]('../user/truffleruby-additions.md').
+
+When third-party code uses the TruffleRuby additions, it should be listed here
+so that we know why non-standard APIs have been added.
 
 ## TruffleRuby Extensions in Gems
 
-* `concurrent-ruby` <= 1.0.5 uses: `Truffle::System.full_memory_barrier`,
-  `Truffle::Primitive.logical_processors`
-* `concurrent-ruby` >= 1.1.0 uses: `TruffleRuby.full_memory_barrier`, 
-  `Truffle::AtomicReference`, `Truffle::System.synchronized`
+* `concurrent-ruby` < 1.1.0 used `Truffle::System.full_memory_barrier`,
+  `Truffle::Primitive.logical_processors`,  `Truffle::AtomicReference`, and
+  `Truffle::System.synchronized`. These are all still available in order to
+  support these older versions of `concurrent-ruby`.
+
+* `concurrent-ruby` >= 1.1.0 uses `TruffleRuby.full_memory_barrier`,
+  `TruffleRuby.synchronized`, and `TruffleRuby::AtomicReference`, 

doc/legal/legal.md

@@ -22,7 +22,7 @@ version 2.4.4, including:
 
 * the standard library in `lib/mri`, 
 * Ruby C extension API in `lib/cext/include` and `src/main/c/cext`, 
-* C extensions in `src/main/c/{etc,openssl,psych,rbconfig-sizeof,syslog,zlib}`
+* C extensions in `src/main/c/{etc,nkf,openssl,psych,rbconfig-sizeof,syslog,zlib}`
 
 MRI is copyright Yukihiro Matsumoto. It is made available under the terms of the
 2-clause BSD licence `ruby-bsdl.txt`, or a custom licence `ruby-licence.txt`.

doc/user/compatibility.md

@@ -83,7 +83,7 @@ currently using.
 #### Some classes marked as internal will be different
 
 MRI provides some classes that are described in the documentation as being only
-available on C Ruby (MRI). We implement these classes if it's practical to do
+available on MRI (C Ruby). We implement these classes if it's practical to do
 so, but this isn't always the case. For example `RubyVM` is not available.
 
 ## Features with subtle differences
@@ -128,6 +128,11 @@ Also, IO buffer drains, writes on IO objects with `sync` set, and
 `write_nonblock`, will not retry the write on `EAGAIN` and `EWOULDBLOCK`, as the
 streams do not provide a way to detect this.
 
+#### Error messages
+
+Error message strings will sometimes differ from MRI, as these are not generally
+covered by the Ruby Specification suite or tests.
+
 ## Features with very low performance
 
 #### Keyword arguments

lib/cext/include/ruby/encoding.h

@@ -40,8 +40,7 @@ enum ruby_encoding_consts {
 #define RB_ENCODING_SET_INLINED(obj,i) RB_ENCODING_SET(obj,i)
 #define RB_ENCODING_SET(obj,i) rb_enc_set_index((obj), (i))
 
-#define RB_ENCODING_GET_INLINED(obj) \
-    (int)((RBASIC(obj)->flags & RUBY_ENCODING_MASK)>>RUBY_ENCODING_SHIFT)
+#define RB_ENCODING_GET_INLINED(obj) RB_ENCODING_GET(obj)
 #define RB_ENCODING_GET(obj) rb_enc_get_index(obj)
 
 #define RB_ENCODING_IS_ASCII8BIT(obj) (RB_ENCODING_GET_INLINED(obj) == 0)

lib/cext/include/truffleruby/constants.h

@@ -9,6 +9,7 @@ VALUE rb_tr_get_Class(void);
 VALUE rb_tr_get_Comparable(void);
 VALUE rb_tr_get_Data(void);
 VALUE rb_tr_get_Encoding(void);
+VALUE rb_tr_get_EncodingError(void);
 VALUE rb_tr_get_Enumerable(void);
 VALUE rb_tr_get_FalseClass(void);
 VALUE rb_tr_get_File(void);
@@ -79,6 +80,7 @@ VALUE rb_tr_get_default_rs(void);
 #define rb_mComparable rb_tr_get_Comparable()
 #define rb_cData rb_tr_get_Data()
 #define rb_cEncoding rb_tr_get_Encoding()
+#define rb_eEncodingError rb_tr_get_EncodingError()
 #define rb_mEnumerable rb_tr_get_Enumerable()
 #define rb_cFalseClass rb_tr_get_FalseClass()
 #define rb_cFile rb_tr_get_File()