Merge branch 'master' into stable

close #325

.gitattributes

@@ -0,0 +1 @@
+NEWS.md merge=union

.travis.yml

@@ -6,9 +6,15 @@ os:
   - osx
 osx_image: xcode8
 julia:
-  - 0.5
+  - 0.6
 #  - nightly 0.6 supports depends on #170
 
+branches:
+  only:
+    - master
+    - stable
+    - /^v\d+\.\d+(\.\d+)?(-\S*)?$/  # for tagging
+
 # dependent apt packages
 addons:
   apt:
@@ -27,18 +33,23 @@ addons:
       - g++-4.8
 
 before_install:
-  - export TRAVIS=test/travis
-  - source $TRAVIS/setup_env.sh
+  - export TRAVIS_DIR=test/travis
+  - source ${TRAVIS_DIR}/setup_env.sh
 
 notifications:
   email: false
 
 script:
-  - source $TRAVIS/run_test.sh
+  # bump the time limit of no ouput
+  # the `travis_wait` wrapper can be removed once this issue fixed:
+  # https://github.com/JuliaLang/julia/pull/23601
+  - ${TRAVIS_DIR}/run_test.sh
 
 after_success:
-  - source $TRAVIS/run_coverage.sh
+  # See https://github.com/dmlc/MXNet.jl/pull/303#issuecomment-341171774
+  - julia -e 'using MXNet; mx._sig_checker()'
+
+  - source ${TRAVIS_DIR}/run_coverage.sh
   - echo $TRAVIS_JULIA_VERSION
   - julia -e 'Pkg.add("Documenter")'
   - julia -e 'cd(Pkg.dir("MXNet")); include(joinpath("docs", "make.jl"))'
-

NEWS.md

@@ -1,3 +1,160 @@
+# v0.3.0 (2017.11.16)
+
+* Update `libmxnet` to
+    * On Windows: v0.12.0.
+    (See https://github.com/apache/incubator-mxnet/releases/tag/0.12.0)
+
+    * On Linux/macOS: v0.12.1.
+    (See https://github.com/apache/incubator-mxnet/releases/tag/0.12.1)
+
+* Drop 0.5 support. ([#300][300])
+
+## New API
+
+### `SymbolicNode`
+
+* Debugging print support. ([#276][276])
+
+### `NDArray`
+
+* `deepcopy` for `NDArray` ([#273][273])
+
+* `scalar ./ NDArray` is available now. ([#292][292])
+
+* `fill` and `fill!` for `NDArray`. ([#297][297], [#311][311])
+
+  An API correspond to Python's `mx.nd.full()`
+
+    * `fill(x, dims, ctx=cpu())`
+    * `fill(x, dims...)`
+    * `fill!(arr::NDArray, x)`
+
+* Matrix (2D `NDArray`) multiplication is available now. ([#300][300])
+
+    ```julia
+    julia> x
+    1x2 mx.NDArray{Float64} @ CPU0:
+     1.0  2.0
+
+    julia> x' * x
+    2x2 mx.NDArray{Float64} @ CPU0:
+     1.0  2.0
+     2.0  4.0
+    ```
+
+* `NDArray` `getindex`/`setindex!` linear indexing support and `first` for
+  extracting scalar value. ([#294][294])
+
+  ```julia
+  julia> x = mx.zeros(2, 5)
+
+  julia> x[5] = 42  # do synchronization and set the value
+  ```
+
+  ```julia
+  julia> y = x[5]  # actually, getindex won't do synchronization, but REPL's showing did it for you
+  1 mx.NDArray{Float32} @ CPU0:
+   42.0
+
+  julia> first(y)  # do sync and get the value
+  42.0f0
+
+  julia> y[]  # this is available, also
+  42.0f0
+  ```
+* Elementwise power of `NDArray`. ([#293][293])
+
+    * `x.^2`
+    * `2.^x`
+    * `x.^y`
+    * where `x` and `y` are `NDArray`s.
+
+* Elementwise power of irrational and `NDArray`. ([#310][310])
+
+    * `e.^x`
+    * `x.^e`
+    * `π.^x`
+
+## API Changes
+
+### `SymbolicNode`
+
+* `reshape` of `SymbolicNode` shares the same interface with Base
+  and additional keyword argument. ([#279][279])
+
+    * `reshape(SymbolicNode, dim; reverse=false, name)`
+    * `reshape(SymbolicNode, dim...; reverse=false, name)`
+    * `Reshape` is deprecated.
+
+* `mx.forward(x)` will return `x.outputs` now. ([#312][312])
+
+### `NDArray`
+
+* `reshape` of `NDArray` shares the same interface with Base. ([#272][272])
+
+    * `reshape(NDArray, dim; reverse=false)`
+    * `reshape(NDArray, dim...; reverse=false)`
+    * `Reshape` is deprecated.
+
+* `srand!` deprecated, please use `srand`. ([#282][282])
+
+* `mean` and `sum` of `NDArray` share the same interface with Base
+  and fix the `axis` indexing. ([#303][303])
+
+    * This is a breaking change; no deprecated warning.
+    * Before: `mean(arr, axis=0)`
+    * After: `mean(arr, 1)`
+
+* `max` and `min` of `NDArray` renamed to `maximum` and `minimum` and share the
+  same interface with Base. The `axis` indexing is fixed, also. ([#303][303])
+
+    * This is a breaking change; no deprecated warning.
+    * Before: `mx.max(arr, axis=0)` or `mx.max_axis(arr, axis=0)`
+    * After: `maximum(arr, 1)`
+
+* `mx.transpose` for high dimension `NDArray` has been renamed to `permutedims`
+  and shares the same interface with Base. ([#303][303])
+
+    * This is a breaking changes; no deprecated warning.
+    * Before: `mx.transpose(A, axis=[2, 1, 3])`
+    * After: `permutedims(A, [2, 1, 3])`
+
+* `prod` of `NDArray` shares the same interface with Base and fix the `axis`
+  indexing. ([#303][303])
+
+    * This is a breaking change; no deprecated warning.
+    * Before: `prod(arr, axis=-1)`
+    * After: `prod(arr, 1)`
+
+## Bugfix
+
+* Broadcasting operation on same variable is back. ([#300][300], [#314][314])
+  ```julia
+  x = mx.NDArray(...)
+  x .* x
+  ```
+
+  ```julia
+  y = mx.Variable(:y)
+  y .* y
+  ```
+
+[272]: https://github.com/dmlc/MXNet.jl/pull/272
+[273]: https://github.com/dmlc/MXNet.jl/pull/273
+[276]: https://github.com/dmlc/MXNet.jl/pull/276
+[279]: https://github.com/dmlc/MXNet.jl/pull/279
+[282]: https://github.com/dmlc/MXNet.jl/pull/282
+[292]: https://github.com/dmlc/MXNet.jl/pull/292
+[293]: https://github.com/dmlc/MXNet.jl/pull/293
+[294]: https://github.com/dmlc/MXNet.jl/pull/294
+[297]: https://github.com/dmlc/MXNet.jl/pull/297
+[300]: https://github.com/dmlc/MXNet.jl/pull/300
+[303]: https://github.com/dmlc/MXNet.jl/pull/303
+[310]: https://github.com/dmlc/MXNet.jl/pull/310
+[311]: https://github.com/dmlc/MXNet.jl/pull/311
+[312]: https://github.com/dmlc/MXNet.jl/pull/312
+[314]: https://github.com/dmlc/MXNet.jl/pull/314
+
 # v0.2.2 (2017.05.14)
 * Updated supported version of MXNet to 0.9.4.
 * Improved build-system with support for auto-detecting GPU support.

README.md

@@ -4,9 +4,10 @@
 [![Windows Build](https://ci.appveyor.com/api/projects/status/re90njols2th2ide?svg=true)](https://ci.appveyor.com/project/pluskid/mxnet-jl)
 [![codecov.io](https://codecov.io/github/dmlc/MXNet.jl/coverage.svg?branch=master)](https://codecov.io/github/dmlc/MXNet.jl?branch=master)
 [![](https://img.shields.io/badge/docs-latest-blue.svg)](https://dmlc.github.io/MXNet.jl/latest)
-[![MXNet](http://pkg.julialang.org/badges/MXNet_0.4.svg)](http://pkg.julialang.org/?pkg=MXNet)
+[![](https://img.shields.io/badge/docs-stable-blue.svg)](https://dmlc.github.io/MXNet.jl/stable)
+[![MXNet](http://pkg.julialang.org/badges/MXNet_0.6.svg)](http://pkg.julialang.org/?pkg=MXNet)
 [![License](http://dmlc.github.io/img/apache2.svg)](LICENSE.md)
-[![Join the chat at https://gitter.im/dmlc/mxnet](https://badges.gitter.im/Join%20Chat.svg)](https://gitter.im/dmlc/mxnet?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
+[![Join the chat at https://gitter.im/dmlc/mxnet](https://badges.gitter.im/Join%20Chat.svg)](https://gitter.im/dmlc/mxnet)
 
 
 MXNet.jl is the [dmlc/mxnet](https://github.com/dmlc/mxnet) [Julia](http://julialang.org/) package. MXNet.jl brings flexible and efficient GPU computing and state-of-art deep learning to Julia. Some highlight of its features include:

REQUIRE

@@ -1,5 +1,6 @@
-julia 0.5
-Compat 0.9.5
+julia 0.6
 Formatting
 BinDeps
 JSON
+MacroTools
+TakingBroadcastSeriously

appveyor.yml

@@ -1,6 +1,6 @@
 environment:
   matrix:
-  - JULIA_URL: "https://julialang-s3.julialang.org/bin/winnt/x64/0.5/julia-0.5-latest-win64.exe"
+  - JULIA_URL: "https://julialang-s3.julialang.org/bin/winnt/x64/0.6/julia-0.6-latest-win64.exe"
 
 branches:
   only:
@@ -20,6 +20,7 @@ install:
         https://ci.appveyor.com/api/projects/$env:APPVEYOR_ACCOUNT_NAME/$env:APPVEYOR_PROJECT_SLUG/history?recordsNumber=50).builds | `
         Where-Object pullRequestId -eq $env:APPVEYOR_PULL_REQUEST_NUMBER)[0].buildNumber) { `
         throw "There are newer queued builds for this pull request, failing early." }
+
 # Download most recent Julia Windows binary
   - ps: (new-object net.webclient).DownloadFile(
         $env:JULIA_URL,
@@ -35,3 +36,4 @@ build_script:
 
 test_script:
   - C:\projects\julia\bin\julia --check-bounds=yes -e "Pkg.test(\"MXNet\")"
+