1 % Building OpenJDK
   3 ## TL;DR (Instructions for the Impatient)
   5 If you are eager to try out building OpenJDK, these simple steps works most of
   6 the time. They assume that you have installed Mercurial (and Cygwin if running
   7 on Windows) and cloned the top-level OpenJDK repository that you want to build.
   9  1. [Get the complete source code](#getting-the-source-code): \
  10     `bash get_source.sh`
  12  2. [Run configure](#running-configure): \
  13     `bash configure`
  15     If `configure` fails due to missing dependencies (to either the
  16     [toolchain](#native-compiler-toolchain-requirements), [external libraries](
  17     #external-library-requirements) or the [boot JDK](#boot-jdk-requirements)),
  18     most of the time it prints a suggestion on how to resolve the situation on
  19     your platform. Follow the instructions, and try running `bash configure`
  20     again.
  22  3. [Run make](#running-make): \
  23     `make images`
  25  4. Verify your newly built JDK: \
  26     `./build/*/images/jdk/bin/java -version`
  28  5. [Run basic tests](##running-tests): \
  29     `make run-test-tier1`
  31 If any of these steps failed, or if you want to know more about build
  32 requirements or build functionality, please continue reading this document.
  34 ## Introduction
  36 OpenJDK is a complex software project. Building it requires a certain amount of
  37 technical expertise, a fair number of dependencies on external software, and
  38 reasonably powerful hardware.
  40 If you just want to use OpenJDK and not build it yourself, this document is not
  41 for you. See for instance [OpenJDK installation](
  42 http://openjdk.java.net/install) for some methods of installing a prebuilt
  43 OpenJDK.
  45 ## Getting the Source Code
  47 OpenJDK uses [Mercurial](http://www.mercurial-scm.org) for source control. The
  48 source code is contained not in a single Mercurial repository, but in a tree
  49 ("forest") of interrelated repositories. You will need to check out all of the
  50 repositories to be able to build OpenJDK. To assist you in dealing with this
  51 somewhat unusual arrangement, there are multiple tools available, which are
  52 explained below.
  54 In any case, make sure you are getting the correct version. At the [OpenJDK
  55 Mercurial server](http://hg.openjdk.java.net/) you can see a list of all
  56 available forests. If you want to build an older version, e.g. JDK 8, it is
  57 recommended that you get the `jdk8u` forest, which contains incremental
  58 updates, instead of the `jdk8` forest, which was frozen at JDK 8 GA.
  60 If you are new to Mercurial, a good place to start is the [Mercurial Beginner's
  61 Guide](http://www.mercurial-scm.org/guide). The rest of this document assumes a
  62 working knowledge of Mercurial.
  64 ### Special Considerations
  66 For a smooth building experience, it is recommended that you follow these rules
  67 on where and how to check out the source code.
  69   * Do not check out the source code in a path which contains spaces. Chances
  70     are the build will not work. This is most likely to be an issue on Windows
  71     systems.
  73   * Do not check out the source code in a path which has a very long name or is
  74     nested many levels deep. Chances are you will hit an OS limitation during
  75     the build.
  77   * Put the source code on a local disk, not a network share. If possible, use
  78     an SSD. The build process is very disk intensive, and having slow disk
  79     access will significantly increase build times. If you need to use a
  80     network share for the source code, see below for suggestions on how to keep
  81     the build artifacts on a local disk.
  83   * On Windows, extra care must be taken to make sure the [Cygwin](#cygwin)
  84     environment is consistent. It is recommended that you follow this
  85     procedure:
  87       * Create the directory that is going to contain the top directory of the
  88         OpenJDK clone by using the `mkdir` command in the Cygwin bash shell.
  89         That is, do *not* create it using Windows Explorer. This will ensure
  90         that it will have proper Cygwin attributes, and that it's children will
  91         inherit those attributes.
  93       * Do not put the OpenJDK clone in a path under your Cygwin home
  94         directory. This is especially important if your user name contains
  95         spaces and/or mixed upper and lower case letters.
  97       * Clone the OpenJDK repository using the Cygwin command line `hg` client
  98         as instructed in this document. That is, do *not* use another Mercurial
  99         client such as TortoiseHg.
 101     Failure to follow this procedure might result in hard-to-debug build
 102     problems.
 104 ### Using get\_source.sh
 106 The simplest way to get the entire forest is probably to clone the top-level
 107 repository and then run the `get_source.sh` script, like this:
 109 ```
 110 hg clone http://hg.openjdk.java.net/jdk9/jdk9
 111 cd jdk9
 112 bash get_source.sh
 113 ```
 115 The first time this is run, it will clone all the sub-repositories. Any
 116 subsequent execution of the script will update all sub-repositories to the
 117 latest revision.
 119 ### Using hgforest.sh
 121 The `hgforest.sh` script is more expressive than `get_source.sh`. It takes any
 122 number of arguments, and runs `hg` with those arguments on each sub-repository
 123 in the forest. The `get_source.sh` script is basically a simple wrapper that
 124 runs either `hgforest.sh clone` or `hgforest.sh pull -u`.
 126   * Cloning the forest:
 127     ```
 128     hg clone http://hg.openjdk.java.net/jdk9/jdk9
 129     cd jdk9
 130     bash common/bin/hgforest.sh clone
 131     ```
 133   * Pulling and updating the forest:
 134     ```
 135     bash common/bin/hgforest.sh pull -u
 136     ```
 138   * Merging over the entire forest:
 139     ```
 140     bash common/bin/hgforest.sh merge
 141     ```
 143 ### Using the Trees Extension
 145 The trees extension is a Mercurial add-on that helps you deal with the forest.
 146 More information is available on the [Code Tools trees page](
 147 http://openjdk.java.net/projects/code-tools/trees).
 149 #### Installing the Extension
 151 Install the extension by cloning `http://hg.openjdk.java.net/code-tools/trees`
 152 and updating your `.hgrc` file. Here's one way to do this:
 154 ```
 155 cd ~
 156 mkdir hg-ext
 157 cd hg-ext
 158 hg clone http://hg.openjdk.java.net/code-tools/trees
 159 cat << EOT >> ~/.hgrc
 160 [extensions]
 161 trees=~/hg-ext/trees/trees.py
 162 EOT
 163 ```
 165 #### Initializing the Tree
 167 The trees extension needs to know the structure of the forest. If you have
 168 already cloned the entire forest using another method, you can initialize the
 169 forest like this:
 171 ```
 172 hg tconf --set --walk --depth
 173 ```
 175 Or you can clone the entire forest at once, if you substitute `clone` with
 176 `tclone` when cloning the top-level repository, e.g. like this:
 178 ```
 179 hg tclone http://hg.openjdk.java.net/jdk9/jdk9
 180 ```
 182 In this case, the forest will be properly initialized from the start.
 184 #### Other Operations
 186 The trees extensions supplement many common operations with a trees version by
 187 prefixing a `t` to the normal Mercurial command, e.g. `tcommit`, `tstatus` or
 188 `tmerge`. For instance, to update the entire forest:
 190 ```
 191 hg tpull -u
 192 ```
 194 ## Build Hardware Requirements
 196 OpenJDK is a massive project, and require machines ranging from decent to
 197 powerful to be able to build in a reasonable amount of time, or to be able to
 198 complete a build at all.
 200 We *strongly* recommend usage of an SSD disk for the build, since disk speed is
 201 one of the limiting factors for build performance.
 203 ### Building on x86
 205 At a minimum, a machine with 2-4 cores is advisable, as well as 2-4 GB of RAM.
 206 (The more cores to use, the more memory you need.) At least 6 GB of free disk
 207 space is required (8 GB minimum for building on Solaris).
 209 Even for 32-bit builds, it is recommended to use a 64-bit build machine, and
 210 instead create a 32-bit target using `--with-target-bits=32`.
 212 ### Building on sparc
 214 At a minimum, a machine with 4 cores is advisable, as well as 4 GB of RAM. (The
 215 more cores to use, the more memory you need.) At least 8 GB of free disk space
 216 is required.
 218 ### Building on arm/aarch64
 220 This is not recommended. Instead, see the section on [Cross-compiling](
 221 #cross-compiling).
 223 ## Operating System Requirements
 225 The mainline OpenJDK project supports Linux, Solaris, macOS, AIX and Windows.
 226 Support for other operating system, e.g. BSD, exists in separate "port"
 227 projects.
 229 In general, OpenJDK can be built on a wide range of versions of these operating
 230 systems, but the further you deviate from what is tested on a daily basis, the
 231 more likely you are to run into problems.
 233 This table lists the OS versions used by Oracle when building JDK 9. Such
 234 information is always subject to change, but this table is up to date at the
 235 time of writing.
 237  Operating system   Vendor/version used
 238  -----------------  -------------------------------------------------------
 239  Linux              Oracle Enterprise Linux 6.4 / 7.1 (using kernel 3.8.13)
 240  Solaris            Solaris 11.1 SRU 21.4.1 / 11.2 SRU 5.5
 241  macOS              Mac OS X 10.9 (Mavericks) / 10.10 (Yosemite)
 242  Windows            Windows Server 2012 R2
 244 The double version numbers for Linux, Solaris and macOS is due to the hybrid
 245 model used at Oracle, where header files and external libraries from an older
 246 version is used when building on a more modern version of the OS.
 248 The Build Group has a wiki page with [Supported Build Platforms](
 249 https://wiki.openjdk.java.net/display/Build/Supported+Build+Platforms). From
 250 time to time, this is updated by the community to list successes or failures of
 251 building on different platforms.
 253 ### Windows
 255 Windows XP is not a supported platform, but all newer Windows should be able to
 256 build OpenJDK.
 258 On Windows, it is important that you pay attention to the instructions in the
 259 [Special Considerations](#special-considerations).
 261 Windows is the only non-POSIX OS supported by OpenJDK, and as such, requires
 262 some extra care. A POSIX support layer is required to build on Windows. For
 263 OpenJDK 9, the only supported such layer is Cygwin. (Msys is no longer
 264 supported due to a too old bash; msys2 and the new Windows Subsystem for Linux
 265 (WSL) would likely be possible to support in a future version but that would
 266 require a community effort to implement.)
 268 Internally in the build system, all paths are represented as Unix-style paths,
 269 e.g. `/cygdrive/c/hg/jdk9/Makefile` rather than `C:\hg\jdk9\Makefile`. This
 270 rule also applies to input to the build system, e.g. in arguments to
 271 `configure`. So, use `--with-freetype=/cygdrive/c/freetype` rather than
 272 `--with-freetype=c:\freetype`. For details on this conversion, see the section
 273 on [Fixpath](#fixpath).
 275 #### Cygwin
 277 A functioning [Cygwin](http://www.cygwin.com/) environment is thus required for
 278 building OpenJDK on Windows. If you have a 64-bit OS, we strongly recommend
 279 using the 64-bit version of Cygwin.
 281 **Note:** Cygwin has a model of continuously updating all packages without any
 282 easy way to install or revert to a specific version of a package. This means
 283 that whenever you add or update a package in Cygwin, you might (inadvertently)
 284 update tools that are used by the OpenJDK build process, and that can cause
 285 unexpected build problems.
 287 OpenJDK requires GNU Make 4.0 or greater on Windows. This is usually not a
 288 problem, since Cygwin currently only distributes GNU Make at a version above
 289 4.0.
 291 Apart from the basic Cygwin installation, the following packages must also be
 292 installed:
 294   * `make`
 295   * `zip`
 296   * `unzip`
 298 Often, you can install these packages using the following command line:
 299 ```
 300 <path to Cygwin setup>/setup-x86_64 -q -P make -P unzip -P zip
 301 ```
 303 Unfortunately, Cygwin can be unreliable in certain circumstances. If you
 304 experience build tool crashes or strange issues when building on Windows,
 305 please check the Cygwin FAQ on the ["BLODA" list](
 306 https://cygwin.com/faq/faq.html#faq.using.bloda) and the section on [fork()
 307 failures](https://cygwin.com/faq/faq.html#faq.using.fixing-fork-failures).
 309 ### Solaris
 311 See `make/devkit/solaris11.1-package-list.txt` for a list of recommended
 312 packages to install when building on Solaris. The versions specified in this
 313 list is the versions used by the daily builds at Oracle, and is likely to work
 314 properly.
 316 Older versions of Solaris shipped a broken version of `objcopy`. At least
 317 version 2.21.1 is needed, which is provided by Solaris 11 Update 1. Objcopy is
 318 needed if you want to have external debug symbols. Please make sure you are
 319 using at least version 2.21.1 of objcopy, or that you disable external debug
 320 symbols.
 322 ### macOS
 324 Apple is using a quite aggressive scheme of pushing OS updates, and coupling
 325 these updates with required updates of Xcode. Unfortunately, this makes it
 326 difficult for a project like OpenJDK to keep pace with a continuously updated
 327 machine running macOS. See the section on [Apple Xcode](#apple-xcode) on some
 328 strategies to deal with this.
 330 It is recommended that you use at least Mac OS X 10.9 (Mavericks). At the time
 331 of writing, OpenJDK has been successfully compiled on macOS versions up to
 332 10.12.5 (Sierra), using XCode 8.3.2 and `--disable-warnings-as-errors`.
 334 The standard macOS environment contains the basic tooling needed to build, but
 335 for external libraries a package manager is recommended. OpenJDK uses
 336 [homebrew](https://brew.sh/) in the examples, but feel free to use whatever
 337 manager you want (or none).
 339 ### Linux
 341 It is often not much problem to build OpenJDK on Linux. The only general advice
 342 is to try to use the compilers, external libraries and header files as provided
 343 by your distribution.
 345 The basic tooling is provided as part of the core operating system, but you
 346 will most likely need to install developer packages.
 348 For apt-based distributions (Debian, Ubuntu, etc), try this:
 349 ```
 350 sudo apt-get install build-essential
 351 ```
 353 For rpm-based distributions (Fedora, Red Hat, etc), try this:
 354 ```
 355 sudo yum groupinstall "Development Tools"
 356 ```
 358 ### AIX
 360 The regular builds by SAP is using AIX version 7.1, but AIX 5.3 is also
 361 supported. See the [OpenJDK PowerPC Port Status Page](
 362 http://cr.openjdk.java.net/~simonis/ppc-aix-port) for details.
 364 ## Native Compiler (Toolchain) Requirements
 366 Large portions of OpenJDK consists of native code, that needs to be compiled to
 367 be able to run on the target platform. In theory, toolchain and operating
 368 system should be independent factors, but in practice there's more or less a
 369 one-to-one correlation between target operating system and toolchain.
 371  Operating system   Supported toolchain
 372  ------------------ -------------------------
 373  Linux              gcc, clang
 374  macOS              Apple Xcode (using clang)
 375  Solaris            Oracle Solaris Studio
 376  AIX                IBM XL C/C++
 377  Windows            Microsoft Visual Studio
 379 Please see the individual sections on the toolchains for version
 380 recommendations. As a reference, these versions of the toolchains are used, at
 381 the time of writing, by Oracle for the daily builds of OpenJDK. It should be
 382 possible to compile OpenJDK with both older and newer versions, but the closer
 383 you stay to this list, the more likely you are to compile successfully without
 384 issues.
 386  Operating system   Toolchain version
 387  ------------------ -------------------------------------------------------
 388  Linux              gcc 4.9.2
 389  macOS              Apple Xcode 6.3 (using clang 6.1.0)
 390  Solaris            Oracle Solaris Studio 12.4 (with compiler version 5.13)
 391  Windows            Microsoft Visual Studio 2013 update 4
 393 ### gcc
 395 The minimum accepted version of gcc is 4.7. Older versions will generate a warning 
 396 by `configure` and are unlikely to work.
 398 OpenJDK 9 includes patches that should allow gcc 6 to compile, but this should
 399 be considered experimental.
 401 In general, any version between these two should be usable.
 403 ### clang
 405 The minimum accepted version of clang is 3.2. Older versions will not be
 406 accepted by `configure`.
 408 To use clang instead of gcc on Linux, use `--with-toolchain-type=clang`.
 410 ### Apple Xcode
 412 The oldest supported version of Xcode is 5.
 414 You will need the Xcode command lines developers tools to be able to build
 415 OpenJDK. (Actually, *only* the command lines tools are needed, not the IDE.)
 416 The simplest way to install these is to run:
 417 ```
 418 xcode-select --install
 419 ```
 421 It is advisable to keep an older version of Xcode for building OpenJDK when
 422 updating Xcode. This [blog page](
 423 http://iosdevelopertips.com/xcode/install-multiple-versions-of-xcode.html) has
 424 good suggestions on managing multiple Xcode versions. To use a specific version
 425 of Xcode, use `xcode-select -s` before running `configure`, or use
 426 `--with-toolchain-path` to point to the version of Xcode to use, e.g.
 427 `configure --with-toolchain-path=/Applications/Xcode5.app/Contents/Developer/usr/bin`
 429 If you have recently (inadvertently) updated your OS and/or Xcode version, and
 430 OpenJDK can no longer be built, please see the section on [Problems with the
 431 Build Environment](#problems-with-the-build-environment), and [Getting
 432 Help](#getting-help) to find out if there are any recent, non-merged patches
 433 available for this update.
 435 ### Oracle Solaris Studio
 437 The minimum accepted version of the Solaris Studio compilers is 5.13
 438 (corresponding to Solaris Studio 12.4). Older versions will not be accepted by
 439 configure.
 441 The Solaris Studio installation should contain at least these packages:
 443  Package                                            Version
 444  -------------------------------------------------- -------------
 445  developer/solarisstudio-124/backend                12.4-
 446  developer/solarisstudio-124/c++                    12.4-
 447  developer/solarisstudio-124/cc                     12.4-
 448  developer/solarisstudio-124/library/c++-libs       12.4-
 449  developer/solarisstudio-124/library/math-libs      12.4-
 450  developer/solarisstudio-124/library/studio-gccrt   12.4-
 451  developer/solarisstudio-124/studio-common          12.4-
 452  developer/solarisstudio-124/studio-ja              12.4-
 453  developer/solarisstudio-124/studio-legal           12.4-
 454  developer/solarisstudio-124/studio-zhCN            12.4-
 456 Compiling with Solaris Studio can sometimes be finicky. This is the exact
 457 version used by Oracle, which worked correctly at the time of writing:
 458 ```
 459 $ cc -V
 460 cc: Sun C 5.13 SunOS_i386 2014/10/20
 461 $ CC -V
 462 CC: Sun C++ 5.13 SunOS_i386 151846-10 2015/10/30
 463 ```
 465 ### Microsoft Visual Studio
 467 The minimum accepted version of Visual Studio is 2010. Older versions will not
 468 be accepted by `configure`. The maximum accepted version of Visual Studio is
 469 2013.
 471 If you have multiple versions of Visual Studio installed, `configure` will by
 472 default pick the latest. You can request a specific version to be used by
 473 setting `--with-toolchain-version`, e.g. `--with-toolchain-version=2010`.
 475 If you get `LINK: fatal error LNK1123: failure during conversion to COFF: file
 476 invalid` when building using Visual Studio 2010, you have encountered
 477 [KB2757355](http://support.microsoft.com/kb/2757355), a bug triggered by a
 478 specific installation order. However, the solution suggested by the KB article
 479 does not always resolve the problem. See [this stackoverflow discussion](
 480 https://stackoverflow.com/questions/10888391) for other suggestions.
 482 ### IBM XL C/C++
 484 The regular builds by SAP is using version 12.1, described as `IBM XL C/C++ for
 485 AIX, V12.1 (5765-J02, 5725-C72) Version: 12.01.0000.0017`.
 487 See the [OpenJDK PowerPC Port Status Page](
 488 http://cr.openjdk.java.net/~simonis/ppc-aix-port) for details.
 490 ## Boot JDK Requirements
 492 Paradoxically, building OpenJDK requires a pre-existing JDK. This is called the
 493 "boot JDK". The boot JDK does not have to be OpenJDK, though. If you are
 494 porting OpenJDK to a new platform, chances are that there already exists
 495 another JDK for that platform that is usable as boot JDK.
 497 The rule of thumb is that the boot JDK for building JDK major version *N*
 498 should be an JDK of major version *N-1*, so for building JDK 9 a JDK 8 would be
 499 suitable as boot JDK. However, OpenJDK should be able to "build itself", so an
 500 up-to-date build of the current OpenJDK source is an acceptable alternative. If
 501 you are following the *N-1* rule, make sure you got the latest update version,
 502 since JDK 8 GA might not be able to build JDK 9 on all platforms.
 504 If the Boot JDK is not automatically detected, or the wrong JDK is picked, use
 505 `--with-boot-jdk` to point to the JDK to use.
 507 ### JDK 8 on Linux
 509 On apt-based distros (like Debian and Ubuntu), `sudo apt-get install
 510 openjdk-8-jdk` is typically enough to install OpenJDK 8. On rpm-based distros
 511 (like Fedora and Red Hat), try `sudo yum install java-1.8.0-openjdk-devel`.
 513 ### JDK 8 on Windows
 515 No pre-compiled binaries of OpenJDK 8 are readily available for Windows at the
 516 time of writing. An alternative is to download the [Oracle JDK](
 517 http://www.oracle.com/technetwork/java/javase/downloads). Another is the [Adopt
 518 OpenJDK Project](https://adoptopenjdk.net/), which publishes experimental
 519 prebuilt binaries for Windows.
 521 ### JDK 8 on macOS
 523 No pre-compiled binaries of OpenJDK 8 are readily available for macOS at the
 524 time of writing. An alternative is to download the [Oracle JDK](
 525 http://www.oracle.com/technetwork/java/javase/downloads), or to install it
 526 using `brew cask install java`. Another option is the [Adopt OpenJDK Project](
 527 https://adoptopenjdk.net/), which publishes experimental prebuilt binaries for
 528 macOS.
 530 ### JDK 8 on AIX
 532 No pre-compiled binaries of OpenJDK 8 are readily available for AIX at the
 533 time of writing. A starting point for working with OpenJDK on AIX is
 534 the [PowerPC/AIX Port Project](http://openjdk.java.net/projects/ppc-aix-port/).
 536 ## External Library Requirements
 538 Different platforms require different external libraries. In general, libraries
 539 are not optional - that is, they are either required or not used.
 541 If a required library is not detected by `configure`, you need to provide the
 542 path to it. There are two forms of the `configure` arguments to point to an
 543 external library: `--with-<LIB>=<path>` or `--with-<LIB>-include=<path to
 544 include> --with-<LIB>-lib=<path to lib>`. The first variant is more concise,
 545 but require the include files an library files to reside in a default hierarchy
 546 under this directory. In most cases, it works fine.
 548 As a fallback, the second version allows you to point to the include directory
 549 and the lib directory separately.
 551 ### FreeType
 553 FreeType2 from [The FreeType Project](http://www.freetype.org/) is required on
 554 all platforms. At least version 2.3 is required.
 556   * To install on an apt-based Linux, try running `sudo apt-get install
 557     libcups2-dev`.
 558   * To install on an rpm-based Linux, try running `sudo yum install
 559     cups-devel`.
 560   * To install on Solaris, try running `pkg install system/library/freetype-2`.
 561   * To install on macOS, try running `brew install freetype`.
 562   * To install on Windows, see [below](#building-freetype-on-windows).
 564 Use `--with-freetype=<path>` if `configure` does not properly locate your
 565 FreeType files.
 567 #### Building FreeType on Windows
 569 On Windows, there is no readily available compiled version of FreeType. OpenJDK
 570 can help you compile FreeType from source. Download the FreeType sources and
 571 unpack them into an arbitrary directory:
 573 ```
 574 wget http://download.savannah.gnu.org/releases/freetype/freetype-2.5.3.tar.gz
 575 tar -xzf freetype-2.5.3.tar.gz
 576 ```
 578 Then run `configure` with `--with-freetype-src=<freetype_src>`. This will
 579 automatically build the freetype library into `<freetype_src>/lib64` for 64-bit
 580 builds or into `<freetype_src>/lib32` for 32-bit builds. Afterwards you can
 581 always use `--with-freetype-include=<freetype_src>/include` and
 582 `--with-freetype-lib=<freetype_src>/lib[32|64]` for other builds.
 584 Alternatively you can unpack the sources like this to use the default
 585 directory:
 587 ```
 588 tar --one-top-level=$HOME/freetype --strip-components=1 -xzf freetype-2.5.3.tar.gz
 589 ```
 591 ### CUPS
 593 CUPS, [Common UNIX Printing System](http://www.cups.org) header files are
 594 required on all platforms, except Windows. Often these files are provided by
 595 your operating system.
 597   * To install on an apt-based Linux, try running `sudo apt-get install
 598     libcups2-dev`.
 599   * To install on an rpm-based Linux, try running `sudo yum install
 600     cups-devel`.
 601   * To install on Solaris, try running `pkg install print/cups`.
 603 Use `--with-cups=<path>` if `configure` does not properly locate your CUPS
 604 files.
 606 ### X11
 608 Certain [X11](http://www.x.org/) libraries and include files are required on
 609 Linux and Solaris.
 611   * To install on an apt-based Linux, try running `sudo apt-get install
 612     libx11-dev libxext-dev libxrender-dev libxtst-dev libxt-dev`.
 613   * To install on an rpm-based Linux, try running `sudo yum install
 614     libXtst-devel libXt-devel libXrender-devel libXi-devel`.
 615   * To install on Solaris, try running `pkg install x11/header/x11-protocols
 616     x11/library/libice x11/library/libpthread-stubs x11/library/libsm
 617     x11/library/libx11 x11/library/libxau x11/library/libxcb
 618     x11/library/libxdmcp x11/library/libxevie x11/library/libxext
 619     x11/library/libxrender x11/library/libxscrnsaver x11/library/libxtst
 620     x11/library/toolkit/libxt`.
 622 Use `--with-x=<path>` if `configure` does not properly locate your X11 files.
 624 ### ALSA
 626 ALSA, [Advanced Linux Sound Architecture](https://www.alsa-project.org/) is
 627 required on Linux. At least version 0.9.1 of ALSA is required.
 629   * To install on an apt-based Linux, try running `sudo apt-get install
 630     libasound2-dev`.
 631   * To install on an rpm-based Linux, try running `sudo yum install
 632     alsa-lib-devel`.
 634 Use `--with-alsa=<path>` if `configure` does not properly locate your ALSA
 635 files.
 637 ### libffi
 639 libffi, the [Portable Foreign Function Interface Library](
 640 http://sourceware.org/libffi) is required when building the Zero version of
 641 Hotspot.
 643   * To install on an apt-based Linux, try running `sudo apt-get install
 644     libffi-dev`.
 645   * To install on an rpm-based Linux, try running `sudo yum install
 646     libffi-devel`.
 648 Use `--with-libffi=<path>` if `configure` does not properly locate your libffi
 649 files.
 651 ### libelf
 653 libelf from the [elfutils project](http://sourceware.org/elfutils) is required
 654 when building the AOT feature of Hotspot.
 656   * To install on an apt-based Linux, try running `sudo apt-get install
 657     libelf-dev`.
 658   * To install on an rpm-based Linux, try running `sudo yum install
 659     elfutils-libelf-devel`.
 661 Use `--with-libelf=<path>` if `configure` does not properly locate your libelf
 662 files.
 664 ## Other Tooling Requirements
 666 ### GNU Make
 668 OpenJDK requires [GNU Make](http://www.gnu.org/software/make). No other flavors
 669 of make are supported.
 671 At least version 3.81 of GNU Make must be used. For distributions supporting
 672 GNU Make 4.0 or above, we strongly recommend it. GNU Make 4.0 contains useful
 673 functionality to handle parallel building (supported by `--with-output-sync`)
 674 and speed and stability improvements.
 676 Note that `configure` locates and verifies a properly functioning version of
 677 `make` and stores the path to this `make` binary in the configuration. If you
 678 start a build using `make` on the command line, you will be using the version
 679 of make found first in your `PATH`, and not necessarily the one stored in the
 680 configuration. This initial make will be used as "bootstrap make", and in a
 681 second stage, the make located by `configure` will be called. Normally, this
 682 will present no issues, but if you have a very old `make`, or a non-GNU Make
 683 `make` in your path, this might cause issues.
 685 If you want to override the default make found by `configure`, use the `MAKE`
 686 configure variable, e.g. `configure MAKE=/opt/gnu/make`.
 688 On Solaris, it is common to call the GNU version of make by using `gmake`.
 690 ### GNU Bash
 692 OpenJDK requires [GNU Bash](http://www.gnu.org/software/bash). No other shells
 693 are supported.
 695 At least version 3.2 of GNU Bash must be used.
 697 ### Autoconf
 699 If you want to modify the build system itself, you need to install [Autoconf](
 700 http://www.gnu.org/software/autoconf).
 702 However, if you only need to build OpenJDK or if you only edit the actual
 703 OpenJDK source files, there is no dependency on autoconf, since the source
 704 distribution includes a pre-generated `configure` shell script.
 706 See the section on [Autoconf Details](#autoconf-details) for details on how
 707 OpenJDK uses autoconf. This is especially important if you plan to contribute
 708 changes to OpenJDK that modifies the build system.
 710 ## Running Configure
 712 To build OpenJDK, you need a "configuration", which consists of a directory
 713 where to store the build output, coupled with information about the platform,
 714 the specific build machine, and choices that affect how OpenJDK is built.
 716 The configuration is created by the `configure` script. The basic invocation of
 717 the `configure` script looks like this:
 719 ```
 720 bash configure [options]
 721 ```
 723 This will create an output directory containing the configuration and setup an
 724 area for the build result. This directory typically looks like
 725 `build/linux-x64-normal-server-release`, but the actual name depends on your
 726 specific configuration. (It can also be set directly, see [Using Multiple
 727 Configurations](#using-multiple-configurations)). This directory is referred to
 728 as `$BUILD` in this documentation.
 730 `configure` will try to figure out what system you are running on and where all
 731 necessary build components are. If you have all prerequisites for building
 732 installed, it should find everything. If it fails to detect any component
 733 automatically, it will exit and inform you about the problem.
 735 Some command line examples:
 737   * Create a 32-bit build for Windows with FreeType2 in `C:\freetype-i586`:
 738     ```
 739     bash configure --with-freetype=/cygdrive/c/freetype-i586 --with-target-bits=32
 740     ```
 742   * Create a debug build with the `server` JVM and DTrace enabled:
 743     ```
 744     bash configure --enable-debug --with-jvm-variants=server --enable-dtrace
 745     ```
 747 ### Common Configure Arguments
 749 Here follows some of the most common and important `configure` argument.
 751 To get up-to-date information on *all* available `configure` argument, please
 752 run:
 753 ```
 754 bash configure --help
 755 ```
 757 (Note that this help text also include general autoconf options, like
 758 `--dvidir`, that is not relevant to OpenJDK. To list only OpenJDK specific
 759 features, use `bash configure --help=short` instead.)
 761 #### Configure Arguments for Tailoring the Build
 763   * `--enable-debug` - Set the debug level to `fastdebug` (this is a shorthand
 764     for `--with-debug-level=fastdebug`)
 765   * `--with-debug-level=<level>` - Set the debug level, which can be `release`,
 766     `fastdebug`, `slowdebug` or `optimized`. Default is `release`. `optimized`
 767     is variant of `release` with additional Hotspot debug code.
 768   * `--with-native-debug-symbols=<method>` - Specify if and how native debug
 769     symbols should be built. Available methods are `none`, `internal`,
 770     `external`, `zipped`. Default behavior depends on platform. See [Native
 771     Debug Symbols](#native-debug-symbols) for more details.
 772   * `--with-version-string=<string>` - Specify the version string this build
 773     will be identified with.
 774   * `--with-version-<part>=<value>` - A group of options, where `<part>` can be
 775     any of `pre`, `opt`, `build`, `major`, `minor`, `security` or `patch`. Use
 776     these options to modify just the corresponding part of the version string
 777     from the default, or the value provided by `--with-version-string`.
 778   * `--with-jvm-variants=<variant>[,<variant>...]` - Build the specified variant
 779     (or variants) of Hotspot. Valid variants are: `server`, `client`,
 780     `minimal`, `core`, `zero`, `zeroshark`, `custom`. Note that not all
 781     variants are possible to combine in a single build.
 782   * `--with-jvm-features=<feature>[,<feature>...]` - Use the specified JVM
 783     features when building Hotspot. The list of features will be enabled on top
 784     of the default list. For the `custom` JVM variant, this default list is
 785     empty. A complete list of available JVM features can be found using `bash
 786     configure --help`.
 787   * `--with-target-bits=<bits>` - Create a target binary suitable for running
 788     on a `<bits>` platform. Use this to create 32-bit output on a 64-bit build
 789     platform, instead of doing a full cross-compile. (This is known as a
 790     *reduced* build.)
 792 #### Configure Arguments for Native Compilation
 794   * `--with-devkit=<path>` - Use this devkit for compilers, tools and resources
 795   * `--with-sysroot=<path>` - Use this directory as sysroot
 796   * `--with-extra-path=<path>[;<path>]` - Prepend these directories to the
 797     default path when searching for all kinds of binaries
 798   * `--with-toolchain-path=<path>[;<path>]` - Prepend these directories when
 799     searching for toolchain binaries (compilers etc)
 800   * `--with-extra-cflags=<flags>` - Append these flags when compiling JDK C
 801     files
 802   * `--with-extra-cxxflags=<flags>` - Append these flags when compiling JDK C++
 803     files
 804   * `--with-extra-ldflags=<flags>` - Append these flags when linking JDK
 805     libraries
 807 #### Configure Arguments for External Dependencies
 809   * `--with-boot-jdk=<path>` - Set the path to the [Boot JDK](
 810     #boot-jdk-requirements)
 811   * `--with-freetype=<path>` - Set the path to [FreeType](#freetype)
 812   * `--with-cups=<path>` - Set the path to [CUPS](#cups)
 813   * `--with-x=<path>` - Set the path to [X11](#x11)
 814   * `--with-alsa=<path>` - Set the path to [ALSA](#alsa)
 815   * `--with-libffi=<path>` - Set the path to [libffi](#libffi)
 816   * `--with-libelf=<path>` - Set the path to [libelf](#libelf)
 817   * `--with-jtreg=<path>` - Set the path to JTReg. See [Running Tests](
 818     #running-tests)
 820 Certain third-party libraries used by OpenJDK (libjpeg, giflib, libpng, lcms
 821 and zlib) are included in the OpenJDK repository. The default behavior of the
 822 OpenJDK build is to use this version of these libraries, but they might be
 823 replaced by an external version. To do so, specify `system` as the `<source>`
 824 option in these arguments. (The default is `bundled`).
 826   * `--with-libjpeg=<source>` - Use the specified source for libjpeg
 827   * `--with-giflib=<source>` - Use the specified source for giflib
 828   * `--with-libpng=<source>` - Use the specified source for libpng
 829   * `--with-lcms=<source>` - Use the specified source for lcms
 830   * `--with-zlib=<source>` - Use the specified source for zlib
 832 On Linux, it is possible to select either static or dynamic linking of the C++
 833 runtime. The default is static linking, with dynamic linking as fallback if the
 834 static library is not found.
 836   * `--with-stdc++lib=<method>` - Use the specified method (`static`, `dynamic`
 837     or `default`) for linking the C++ runtime.
 839 ### Configure Control Variables
 841 It is possible to control certain aspects of `configure` by overriding the
 842 value of `configure` variables, either on the command line or in the
 843 environment.
 845 Normally, this is **not recommended**. If used improperly, it can lead to a
 846 broken configuration. Unless you're well versed in the build system, this is
 847 hard to use properly. Therefore, `configure` will print a warning if this is
 848 detected.
 850 However, there are a few `configure` variables, known as *control variables*
 851 that are supposed to be overriden on the command line. These are variables that
 852 describe the location of tools needed by the build, like `MAKE` or `GREP`. If
 853 any such variable is specified, `configure` will use that value instead of
 854 trying to autodetect the tool. For instance, `bash configure
 855 MAKE=/opt/gnumake4.0/bin/make`.
 857 If a configure argument exists, use that instead, e.g. use `--with-jtreg`
 858 instead of setting `JTREGEXE`.
 860 Also note that, despite what autoconf claims, setting `CFLAGS` will not
 861 accomplish anything. Instead use `--with-extra-cflags` (and similar for
 862 `cxxflags` and `ldflags`).
 864 ## Running Make
 866 When you have a proper configuration, all you need to do to build OpenJDK is to
 867 run `make`. (But see the warning at [GNU Make](#gnu-make) about running the
 868 correct version of make.)
 870 When running `make` without any arguments, the default target is used, which is
 871 the same as running `make default` or `make jdk`. This will build a minimal (or
 872 roughly minimal) set of compiled output (known as an "exploded image") needed
 873 for a developer to actually execute the newly built JDK. The idea is that in an
 874 incremental development fashion, when doing a normal make, you should only
 875 spend time recompiling what's changed (making it purely incremental) and only
 876 do the work that's needed to actually run and test your code.
 878 The output of the exploded image resides in `$BUILD/jdk`. You can test the
 879 newly built JDK like this: `$BUILD/jdk/bin/java -version`.
 881 ### Common Make Targets
 883 Apart from the default target, here are some common make targets:
 885   * `hotspot` - Build all of hotspot (but only hotspot)
 886   * `hotspot-<variant>` - Build just the specified jvm variant
 887   * `images` or `product-images` - Build the JRE and JDK images
 888   * `docs` or `docs-image` - Build the documentation image
 889   * `test-image` - Build the test image
 890   * `all` or `all-images` - Build all images (product, docs and test)
 891   * `bootcycle-images` - Build images twice, second time with newly built JDK
 892     (good for testing)
 893   * `clean` - Remove all files generated by make, but not those generated by
 894     configure
 895   * `dist-clean` - Remove all files, including configuration
 897 Run `make help` to get an up-to-date list of important make targets and make
 898 control variables.
 900 It is possible to build just a single module, a single phase, or a single phase
 901 of a single module, by creating make targets according to these followin
 902 patterns. A phase can be either of `gensrc`, `gendata`, `copy`, `java`,
 903 `launchers`, `libs` or `rmic`. See [Using Fine-Grained Make Targets](
 904 #using-fine-grained-make-targets) for more details about this functionality.
 906   * `<phase>` - Build the specified phase and everything it depends on
 907   * `<module>` - Build the specified module and everything it depends on
 908   * `<module>-<phase>` - Compile the specified phase for the specified module
 909     and everything it depends on
 911 Similarly, it is possible to clean just a part of the build by creating make
 912 targets according to these patterns:
 914   * `clean-<outputdir>` - Remove the subdir in the output dir with the name
 915   * `clean-<phase>` - Remove all build results related to a certain build
 916     phase
 917   * `clean-<module>` - Remove all build results related to a certain module
 918   * `clean-<module>-<phase>` - Remove all build results related to a certain
 919     module and phase
 921 ### Make Control Variables
 923 It is possible to control `make` behavior by overriding the value of `make`
 924 variables, either on the command line or in the environment.
 926 Normally, this is **not recommended**. If used improperly, it can lead to a
 927 broken build. Unless you're well versed in the build system, this is hard to
 928 use properly. Therefore, `make` will print a warning if this is detected.
 930 However, there are a few `make` variables, known as *control variables* that
 931 are supposed to be overriden on the command line. These make up the "make time"
 932 configuration, as opposed to the "configure time" configuration.
 934 #### General Make Control Variables
 936   * `JOBS` - Specify the number of jobs to build with. See [Build
 937     Performance](#build-performance).
 938   * `LOG` - Specify the logging level and functionality. See [Checking the
 939     Build Log File](#checking-the-build-log-file)
 940   * `CONF` and `CONF_NAME` - Selecting the configuration(s) to use. See [Using
 941     Multiple Configurations](#using-multiple-configurations)
 943 #### Test Make Control Variables
 945 These make control variables only make sense when running tests. Please see
 946 [Testing OpenJDK](testing.html) for details.
 948   * `TEST`
 949   * `TEST_JOBS`
 950   * `JTREG`
 951   * `GTEST`
 953 #### Advanced Make Control Variables
 955 These advanced make control variables can be potentially unsafe. See [Hints and
 956 Suggestions for Advanced Users](#hints-and-suggestions-for-advanced-users) and
 957 [Understanding the Build System](#understanding-the-build-system) for details.
 959   * `SPEC`
 960   * `CONF_CHECK`
 961   * `COMPARE_BUILD`
 962   * `JDK_FILTER`
 964 ## Running Tests
 966 Most of the OpenJDK tests are using the [JTReg](http://openjdk.java.net/jtreg)
 967 test framework. Make sure that your configuration knows where to find your
 968 installation of JTReg. If this is not picked up automatically, use the
 969 `--with-jtreg=<path to jtreg home>` option to point to the JTReg framework.
 970 Note that this option should point to the JTReg home, i.e. the top directory,
 971 containing `lib/jtreg.jar` etc.
 973 To execute the most basic tests (tier 1), use:
 974 ```
 975 make run-test-tier1
 976 ```
 978 For more details on how to run tests, please see the [Testing
 979 OpenJDK](testing.html) document.
 981 ## Cross-compiling
 983 Cross-compiling means using one platform (the *build* platform) to generate
 984 output that can ran on another platform (the *target* platform).
 986 The typical reason for cross-compiling is that the build is performed on a more
 987 powerful desktop computer, but the resulting binaries will be able to run on a
 988 different, typically low-performing system. Most of the complications that
 989 arise when building for embedded is due to this separation of *build* and
 990 *target* systems.
 992 This requires a more complex setup and build procedure. This section assumes
 993 you are familiar with cross-compiling in general, and will only deal with the
 994 particularities of cross-compiling OpenJDK. If you are new to cross-compiling,
 995 please see the [external links at Wikipedia](
 996 https://en.wikipedia.org/wiki/Cross_compiler#External_links) for a good start
 997 on reading materials.
 999 Cross-compiling OpenJDK requires you to be able to build both for the build
1000 platform and for the target platform. The reason for the former is that we need
1001 to build and execute tools during the build process, both native tools and Java
1002 tools.
1004 If all you want to do is to compile a 32-bit version, for the same OS, on a
1005 64-bit machine, consider using `--with-target-bits=32` instead of doing a
1006 full-blown cross-compilation. (While this surely is possible, it's a lot more
1007 work and will take much longer to build.)
1009 ### Boot JDK and Build JDK
1011 When cross-compiling, make sure you use a boot JDK that runs on the *build*
1012 system, and not on the *target* system.
1014 To be able to build, we need a "Build JDK", which is a JDK built from the
1015 current sources (that is, the same as the end result of the entire build
1016 process), but able to run on the *build* system, and not the *target* system.
1017 (In contrast, the Boot JDK should be from an older release, e.g. JDK 8 when
1018 building JDK 9.)
1020 The build process will create a minimal Build JDK for you, as part of building.
1021 To speed up the build, you can use `--with-build-jdk` to `configure` to point
1022 to a pre-built Build JDK. Please note that the build result is unpredictable,
1023 and can possibly break in subtle ways, if the Build JDK does not **exactly**
1024 match the current sources.
1026 ### Specifying the Target Platform
1028 You *must* specify the target platform when cross-compiling. Doing so will also
1029 automatically turn the build into a cross-compiling mode. The simplest way to
1030 do this is to use the `--openjdk-target` argument, e.g.
1031 `--openjdk-target=arm-linux-gnueabihf`. or `--openjdk-target=aarch64-oe-linux`.
1032 This will automatically set the `--build`, `--host` and `--target` options for
1033 autoconf, which can otherwise be confusing. (In autoconf terminology, the
1034 "target" is known as "host", and "target" is used for building a Canadian
1035 cross-compiler.)
1037 ### Toolchain Considerations
1039 You will need two copies of your toolchain, one which generates output that can
1040 run on the target system (the normal, or *target*, toolchain), and one that
1041 generates output that can run on the build system (the *build* toolchain). Note
1042 that cross-compiling is only supported for gcc at the time being. The gcc
1043 standard is to prefix cross-compiling toolchains with the target denominator.
1044 If you follow this standard, `configure` is likely to pick up the toolchain
1045 correctly.
1047 The *build* toolchain will be autodetected just the same way the normal
1048 *build*/*target* toolchain will be autodetected when not cross-compiling. If
1049 this is not what you want, or if the autodetection fails, you can specify a
1050 devkit containing the *build* toolchain using `--with-build-devkit` to
1051 `configure`, or by giving `BUILD_CC` and `BUILD_CXX` arguments.
1053 It is often helpful to locate the cross-compilation tools, headers and
1054 libraries in a separate directory, outside the normal path, and point out that
1055 directory to `configure`. Do this by setting the sysroot (`--with-sysroot`) and
1056 appending the directory when searching for cross-compilations tools
1057 (`--with-toolchain-path`). As a compact form, you can also use `--with-devkit`
1058 to point to a single directory, if it is correctly setup. (See `basics.m4` for
1059 details.)
1061 If you are unsure what toolchain and versions to use, these have been proved
1062 working at the time of writing:
1064   * [aarch64](
1065 https://releases.linaro.org/archive/13.11/components/toolchain/binaries/gcc-linaro-aarch64-linux-gnu-4.8-2013.11_linux.tar.xz)
1066   * [arm 32-bit hardware floating  point](
1067 https://launchpad.net/linaro-toolchain-unsupported/trunk/2012.09/+download/gcc-linaro-arm-linux-gnueabihf-raspbian-2012.09-20120921_linux.tar.bz2)
1069 ### Native Libraries
1071 You will need copies of external native libraries for the *target* system,
1072 present on the *build* machine while building.
1074 Take care not to replace the *build* system's version of these libraries by
1075 mistake, since that can render the *build* machine unusable.
1077 Make sure that the libraries you point to (ALSA, X11, etc) are for the
1078 *target*, not the *build*, platform.
1080 #### ALSA
1082 You will need alsa libraries suitable for your *target* system. For most cases,
1083 using Debian's pre-built libraries work fine.
1085 Note that alsa is needed even if you only want to build a headless JDK.
1087   * Go to [Debian Package Search](https://www.debian.org/distrib/packages) and
1088     search for the `libasound2` and `libasound2-dev` packages for your *target*
1089     system. Download them to /tmp.
1091   * Install the libraries into the cross-compilation toolchain. For instance:
1092 ```
1093 cd /tools/gcc-linaro-arm-linux-gnueabihf-raspbian-2012.09-20120921_linux/arm-linux-gnueabihf/libc
1094 dpkg-deb -x /tmp/libasound2_1.0.25-4_armhf.deb .
1095 dpkg-deb -x /tmp/libasound2-dev_1.0.25-4_armhf.deb .
1096 ```
1098   * If alsa is not properly detected by `configure`, you can point it out by
1099     `--with-alsa`.
1101 #### X11
1103 You will need X11 libraries suitable for your *target* system. For most cases,
1104 using Debian's pre-built libraries work fine.
1106 Note that X11 is needed even if you only want to build a headless JDK.
1108   * Go to [Debian Package Search](https://www.debian.org/distrib/packages),
1109     search for the following packages for your *target* system, and download them
1110     to /tmp/target-x11:
1111       * libxi
1112       * libxi-dev
1113       * x11proto-core-dev
1114       * x11proto-input-dev
1115       * x11proto-kb-dev
1116       * x11proto-render-dev
1117       * x11proto-xext-dev
1118       * libice-dev
1119       * libxrender
1120       * libxrender-dev
1121       * libsm-dev
1122       * libxt-dev
1123       * libx11
1124       * libx11-dev
1125       * libxtst
1126       * libxtst-dev
1127       * libxext
1128       * libxext-dev
1130   * Install the libraries into the cross-compilation toolchain. For instance:
1131     ```
1132     cd /tools/gcc-linaro-arm-linux-gnueabihf-raspbian-2012.09-20120921_linux/arm-linux-gnueabihf/libc/usr
1133     mkdir X11R6
1134     cd X11R6
1135     for deb in /tmp/target-x11/*.deb ; do dpkg-deb -x $deb . ; done
1136     mv usr/* .
1137     cd lib
1138     cp arm-linux-gnueabihf/* .
1139     ```
1141     You can ignore the following messages. These libraries are not needed to
1142     successfully complete a full JDK build.
1143     ```
1144     cp: cannot stat `arm-linux-gnueabihf/libICE.so': No such file or directory
1145     cp: cannot stat `arm-linux-gnueabihf/libSM.so': No such file or directory
1146     cp: cannot stat `arm-linux-gnueabihf/libXt.so': No such file or directory
1147     ```
1149   * If the X11 libraries are not properly detected by `configure`, you can
1150     point them out by `--with-x`.
1152 ### Building for ARM/aarch64
1154 A common cross-compilation target is the ARM CPU. When building for ARM, it is
1155 useful to set the ABI profile. A number of pre-defined ABI profiles are
1156 available using `--with-abi-profile`: arm-vfp-sflt, arm-vfp-hflt, arm-sflt,
1157 armv5-vfp-sflt, armv6-vfp-hflt. Note that soft-float ABIs are no longer
1158 properly supported on OpenJDK.
1160 OpenJDK contains two different ports for the aarch64 platform, one is the
1161 original aarch64 port from the [AArch64 Port Project](
1162 http://openjdk.java.net/projects/aarch64-port) and one is a 64-bit version of
1163 the Oracle contributed ARM port. When targeting aarch64, by the default the
1164 original aarch64 port is used. To select the Oracle ARM 64 port, use
1165 `--with-cpu-port=arm64`. Also set the corresponding value (`aarch64` or
1166 `arm64`) to --with-abi-profile, to ensure a consistent build.
1168 ### Verifying the Build
1170 The build will end up in a directory named like
1171 `build/linux-arm-normal-server-release`.
1173 Inside this build output directory, the `images/jdk` and `images/jre` will
1174 contain the newly built JDK and JRE, respectively, for your *target* system.
1176 Copy these folders to your *target* system. Then you can run e.g.
1177 `images/jdk/bin/java -version`.
1179 ## Build Performance
1181 Building OpenJDK requires a lot of horsepower. Some of the build tools can be
1182 adjusted to utilize more or less of resources such as parallel threads and
1183 memory. The `configure` script analyzes your system and selects reasonable
1184 values for such options based on your hardware. If you encounter resource
1185 problems, such as out of memory conditions, you can modify the detected values
1186 with:
1188   * `--with-num-cores` -- number of cores in the build system, e.g.
1189     `--with-num-cores=8`.
1191   * `--with-memory-size` -- memory (in MB) available in the build system, e.g.
1192     `--with-memory-size=1024`
1194 You can also specify directly the number of build jobs to use with
1195 `--with-jobs=N` to `configure`, or `JOBS=N` to `make`. Do not use the `-j` flag
1196 to `make`. In most cases it will be ignored by the makefiles, but it can cause
1197 problems for some make targets.
1199 It might also be necessary to specify the JVM arguments passed to the Boot JDK,
1200 using e.g. `--with-boot-jdk-jvmargs="-Xmx8G"`. Doing so will override the
1201 default JVM arguments passed to the Boot JDK.
1203 At the end of a successful execution of `configure`, you will get a performance
1204 summary, indicating how well the build will perform. Here you will also get
1205 performance hints. If you want to build fast, pay attention to those!
1207 If you want to tweak build performance, run with `make LOG=info` to get a build
1208 time summary at the end of the build process.
1210 ### Disk Speed
1212 If you are using network shares, e.g. via NFS, for your source code, make sure
1213 the build directory is situated on local disk (e.g. by `ln -s
1214 /localdisk/jdk-build $JDK-SHARE/build`). The performance penalty is extremely
1215 high for building on a network share; close to unusable.
1217 Also, make sure that your build tools (including Boot JDK and toolchain) is
1218 located on a local disk and not a network share.
1220 As has been stressed elsewhere, do use SSD for source code and build directory,
1221 as well as (if possible) the build tools.
1223 ### Virus Checking
1225 The use of virus checking software, especially on Windows, can *significantly*
1226 slow down building of OpenJDK. If possible, turn off such software, or exclude
1227 the directory containing the OpenJDK source code from on-the-fly checking.
1229 ### Ccache
1231 The OpenJDK build supports building with ccache when using gcc or clang. Using
1232 ccache can radically speed up compilation of native code if you often rebuild
1233 the same sources. Your milage may vary however, so we recommend evaluating it
1234 for yourself. To enable it, make sure it's on the path and configure with
1235 `--enable-ccache`.
1237 ### Precompiled Headers
1239 By default, the Hotspot build uses preccompiled headers (PCH) on the toolchains
1240 were it is properly supported (clang, gcc, and Visual Studio). Normally, this
1241 speeds up the build process, but in some circumstances, it can actually slow
1242 things down.
1244 You can experiment by disabling precompiled headers using
1245 `--disable-precompiled-headers`.
1247 ### Icecc / icecream
1249 [icecc/icecream](http://github.com/icecc/icecream) is a simple way to setup a
1250 distributed compiler network. If you have multiple machines available for
1251 building OpenJDK, you can drastically cut individual build times by utilizing
1252 it.
1254 To use, setup an icecc network, and install icecc on the build machine. Then
1255 run `configure` using `--enable-icecc`.
1257 ### Using sjavac
1259 To speed up Java compilation, especially incremental compilations, you can try
1260 the experimental sjavac compiler by using `--enable-sjavac`.
1262 ### Building the Right Target
1264 Selecting the proper target to build can have dramatic impact on build time.
1265 For normal usage, `jdk` or the default target is just fine. You only need to
1266 build `images` for shipping, or if your tests require it.
1268 See also [Using Fine-Grained Make Targets](#using-fine-grained-make-targets) on
1269 how to build an even smaller subset of the product.
1271 ## Troubleshooting
1273 If your build fails, it can sometimes be difficult to pinpoint the problem or
1274 find a proper solution.
1276 ### Locating the Source of the Error
1278 When a build fails, it can be hard to pinpoint the actual cause of the error.
1279 In a typical build process, different parts of the product build in parallel,
1280 with the output interlaced.
1282 #### Build Failure Summary
1284 To help you, the build system will print a failure summary at the end. It looks
1285 like this:
1287 ```
1288 ERROR: Build failed for target 'hotspot' in configuration 'linux-x64' (exit code 2)
1290 === Output from failing command(s) repeated here ===
1291 * For target hotspot_variant-server_libjvm_objs_psMemoryPool.o:
1292 /localhome/hg/jdk9-sandbox/hotspot/src/share/vm/services/psMemoryPool.cpp:1:1: error: 'failhere' does not name a type
1293    ... (rest of output omitted)
1295 * All command lines available in /localhome/hg/jdk9-sandbox/build/linux-x64/make-support/failure-logs.
1296 === End of repeated output ===
1298 === Make failed targets repeated here ===
1299 lib/CompileJvm.gmk:207: recipe for target '/localhome/hg/jdk9-sandbox/build/linux-x64/hotspot/variant-server/libjvm/objs/psMemoryPool.o' failed
1300 make/Main.gmk:263: recipe for target 'hotspot-server-libs' failed
1301 === End of repeated output ===
1303 Hint: Try searching the build log for the name of the first failed target.
1304 Hint: If caused by a warning, try configure --disable-warnings-as-errors.
1305 ```
1307 Let's break it down! First, the selected configuration, and the top-level
1308 target you entered on the command line that caused the failure is printed.
1310 Then, between the `Output from failing command(s) repeated here` and `End of
1311 repeated output` the first lines of output (stdout and stderr) from the actual
1312 failing command is repeated. In most cases, this is the error message that
1313 caused the build to fail. If multiple commands were failing (this can happen in
1314 a parallel build), output from all failed commands will be printed here.
1316 The path to the `failure-logs` directory is printed. In this file you will find
1317 a `<target>.log` file that contains the output from this command in its
1318 entirety, and also a `<target>.cmd`, which contain the complete command line
1319 used for running this command. You can re-run the failing command by executing
1320 `. <path to failure-logs>/<target>.cmd` in your shell.
1322 Another way to trace the failure is to follow the chain of make targets, from
1323 top-level targets to individual file targets. Between `Make failed targets
1324 repeated here` and `End of repeated output` the output from make showing this
1325 chain is repeated. The first failed recipe will typically contain the full path
1326 to the file in question that failed to compile. Following lines will show a
1327 trace of make targets why we ended up trying to compile that file.
1329 Finally, some hints are given on how to locate the error in the complete log.
1330 In this example, we would try searching the log file for "`psMemoryPool.o`".
1331 Another way to quickly locate make errors in the log is to search for "`]
1332 Error`" or "`***`".
1334 Note that the build failure summary will only help you if the issue was a
1335 compilation failure or similar. If the problem is more esoteric, or is due to
1336 errors in the build machinery, you will likely get empty output logs, and `No
1337 indication of failed target found` instead of the make target chain.
1339 #### Checking the Build Log File
1341 The output (stdout and stderr) from the latest build is always stored in
1342 `$BUILD/build.log`. The previous build log is stored as `build.log.old`. This
1343 means that it is not necessary to redirect the build output yourself if you
1344 want to process it.
1346 You can increase the verbosity of the log file, by the `LOG` control variable
1347 to `make`. If you want to see the command lines used in compilations, use
1348 `LOG=cmdlines`. To increase the general verbosity, use `LOG=info`, `LOG=debug`
1349 or `LOG=trace`. Both of these can be combined with `cmdlines`, e.g.
1350 `LOG=info,cmdlines`. The `debug` log level will show most shell commands
1351 executed by make, and `trace` will show all. Beware that both these log levels
1352 will produce a massive build log!
1354 ### Fixing Unexpected Build Failures
1356 Most of the time, the build will fail due to incorrect changes in the source
1357 code.
1359 Sometimes the build can fail with no apparent changes that have caused the
1360 failure. If this is the first time you are building OpenJDK on this particular
1361 computer, and the build fails, the problem is likely with your build
1362 environment. But even if you have previously built OpenJDK with success, and it
1363 now fails, your build environment might have changed (perhaps due to OS
1364 upgrades or similar). But most likely, such failures are due to problems with
1365 the incremental rebuild.
1367 #### Problems with the Build Environment
1369 Make sure your configuration is correct. Re-run `configure`, and look for any
1370 warnings. Warnings that appear in the middle of the `configure` output is also
1371 repeated at the end, after the summary. The entire log is stored in
1372 `$BUILD/configure.log`.
1374 Verify that the summary at the end looks correct. Are you indeed using the Boot
1375 JDK and native toolchain that you expect?
1377 By default, OpenJDK has a strict approach where warnings from the compiler is
1378 considered errors which fail the build. For very new or very old compiler
1379 versions, this can trigger new classes of warnings, which thus fails the build.
1380 Run `configure` with `--disable-warnings-as-errors` to turn of this behavior.
1381 (The warnings will still show, but not make the build fail.)
1383 #### Problems with Incremental Rebuilds
1385 Incremental rebuilds mean that when you modify part of the product, only the
1386 affected parts get rebuilt. While this works great in most cases, and
1387 significantly speed up the development process, from time to time complex
1388 interdependencies will result in an incorrect build result. This is the most
1389 common cause for unexpected build problems, together with inconsistencies
1390 between the different Mercurial repositories in the forest.
1392 Here are a suggested list of things to try if you are having unexpected build
1393 problems. Each step requires more time than the one before, so try them in
1394 order. Most issues will be solved at step 1 or 2.
1396  1. Make sure your forest is up-to-date
1398     Run `bash get_source.sh` to make sure you have the latest version of all
1399     repositories.
1401  2. Clean build results
1403     The simplest way to fix incremental rebuild issues is to run `make clean`.
1404     This will remove all build results, but not the configuration or any build
1405     system support artifacts. In most cases, this will solve build errors
1406     resulting from incremental build mismatches.
1408  3. Completely clean the build directory.
1410     If this does not work, the next step is to run `make dist-clean`, or
1411     removing the build output directory (`$BUILD`). This will clean all
1412     generated output, including your configuration. You will need to re-run
1413     `configure` after this step. A good idea is to run `make
1414     print-configuration` before running `make dist-clean`, as this will print
1415     your current `configure` command line. Here's a way to do this:
1417     ```
1418     make print-configuration > current-configuration
1419     make dist-clean
1420     bash configure $(cat current-configuration)
1421     make
1422     ```
1424  4. Re-clone the Mercurial forest
1426     Sometimes the Mercurial repositories themselves gets in a state that causes
1427     the product to be un-buildable. In such a case, the simplest solution is
1428     often the "sledgehammer approach": delete the entire forest, and re-clone
1429     it. If you have local changes, save them first to a different location
1430     using `hg export`.
1432 ### Specific Build Issues
1434 #### Clock Skew
1436 If you get an error message like this:
1437 ```
1438 File 'xxx' has modification time in the future.
1439 Clock skew detected. Your build may be incomplete.
1440 ```
1441 then the clock on your build machine is out of sync with the timestamps on the
1442 source files. Other errors, apparently unrelated but in fact caused by the
1443 clock skew, can occur along with the clock skew warnings. These secondary
1444 errors may tend to obscure the fact that the true root cause of the problem is
1445 an out-of-sync clock.
1447 If you see these warnings, reset the clock on the build machine, run `make
1448 clean` and restart the build.
1450 #### Out of Memory Errors
1452 On Solaris, you might get an error message like this:
1453 ```
1454 Trouble writing out table to disk
1455 ```
1456 To solve this, increase the amount of swap space on your build machine.
1458 On Windows, you might get error messages like this:
1459 ```
1460 fatal error - couldn't allocate heap
1461 cannot create ... Permission denied
1462 spawn failed
1463 ```
1464 This can be a sign of a Cygwin problem. See the information about solving
1465 problems in the [Cygwin](#cygwin) section. Rebooting the computer might help
1466 temporarily.
1468 ### Getting Help
1470 If none of the suggestions in this document helps you, or if you find what you
1471 believe is a bug in the build system, please contact the Build Group by sending
1472 a mail to [build-dev@openjdk.java.net](mailto:build-dev@openjdk.java.net).
1473 Please include the relevant parts of the configure and/or build log.
1475 If you need general help or advice about developing for OpenJDK, you can also
1476 contact the Adoption Group. See the section on [Contributing to OpenJDK](
1477 #contributing-to-openjdk) for more information.
1479 ## Hints and Suggestions for Advanced Users
1481 ### Setting Up a Forest for Pushing Changes (defpath)
1483 To help you prepare a proper push path for a Mercurial repository, there exists
1484 a useful tool known as [defpath](
1485 http://openjdk.java.net/projects/code-tools/defpath). It will help you setup a
1486 proper push path for pushing changes to OpenJDK.
1488 Install the extension by cloning
1489 `http://hg.openjdk.java.net/code-tools/defpath` and updating your `.hgrc` file.
1490 Here's one way to do this:
1492 ```
1493 cd ~
1494 mkdir hg-ext
1495 cd hg-ext
1496 hg clone http://hg.openjdk.java.net/code-tools/defpath
1497 cat << EOT >> ~/.hgrc
1498 [extensions]
1499 defpath=~/hg-ext/defpath/defpath.py
1500 EOT
1501 ```
1503 You can now setup a proper push path using:
1504 ```
1505 hg defpath -d -u <your OpenJDK username>
1506 ```
1508 If you also have the `trees` extension installed in Mercurial, you will
1509 automatically get a `tdefpath` command, which is even more useful. By running
1510 `hg tdefpath -du <username>` in the top repository of your forest, all repos
1511 will get setup automatically. This is the recommended usage.
1513 ### Bash Completion
1515 The `configure` and `make` commands tries to play nice with bash command-line
1516 completion (using `<tab>` or `<tab><tab>`). To use this functionality, make
1517 sure you enable completion in your `~/.bashrc` (see instructions for bash in
1518 your operating system).
1520 Make completion will work out of the box, and will complete valid make targets.
1521 For instance, typing `make jdk-i<tab>` will complete to `make jdk-image`.
1523 The `configure` script can get completion for options, but for this to work you
1524 need to help `bash` on the way. The standard way of running the script, `bash
1525 configure`, will not be understood by bash completion. You need `configure` to
1526 be the command to run. One way to achieve this is to add a simple helper script
1527 to your path:
1529 ```
1530 cat << EOT > /tmp/configure
1531 #!/bin/bash
1532 if [ \$(pwd) = \$(cd \$(dirname \$0); pwd) ] ; then
1533   echo >&2 "Abort: Trying to call configure helper recursively"
1534   exit 1
1535 fi
1537 bash \$PWD/configure "\$@"
1538 EOT
1539 chmod +x /tmp/configure
1540 sudo mv /tmp/configure /usr/local/bin
1541 ```
1543 Now `configure --en<tab>-dt<tab>` will result in `configure --enable-dtrace`.
1545 ### Using Multiple Configurations
1547 You can have multiple configurations for a single source forest. When you
1548 create a new configuration, run `configure --with-conf-name=<name>` to create a
1549 configuration with the name `<name>`. Alternatively, you can create a directory
1550 under `build` and run `configure` from there, e.g. `mkdir build/<name> && cd
1551 build/<name> && bash ../../configure`.
1553 Then you can build that configuration using `make CONF_NAME=<name>` or `make
1554 CONF=<pattern>`, where `<pattern>` is a substring matching one or several
1555 configurations, e.g. `CONF=debug`. The special empty pattern (`CONF=`) will
1556 match *all* available configuration, so `make CONF= hotspot` will build the
1557 `hotspot` target for all configurations. Alternatively, you can execute `make`
1558 in the configuration directory, e.g. `cd build/<name> && make`.
1560 ### Handling Reconfigurations
1562 If you update the forest and part of the configure script has changed, the
1563 build system will force you to re-run `configure`.
1565 Most of the time, you will be fine by running `configure` again with the same
1566 arguments as the last time, which can easily be performed by `make
1567 reconfigure`. To simplify this, you can use the `CONF_CHECK` make control
1568 variable, either as `make CONF_CHECK=auto`, or by setting an environment
1569 variable. For instance, if you add `export CONF_CHECK=auto` to your `.bashrc`
1570 file, `make` will always run `reconfigure` automatically whenever the configure
1571 script has changed.
1573 You can also use `CONF_CHECK=ignore` to skip the check for a needed configure
1574 update. This might speed up the build, but comes at the risk of an incorrect
1575 build result. This is only recommended if you know what you're doing.
1577 From time to time, you will also need to modify the command line to `configure`
1578 due to changes. Use `make print-configure` to show the command line used for
1579 your current configuration.
1581 ### Using Fine-Grained Make Targets
1583 The default behavior for make is to create consistent and correct output, at
1584 the expense of build speed, if necessary.
1586 If you are prepared to take some risk of an incorrect build, and know enough of
1587 the system to understand how things build and interact, you can speed up the
1588 build process considerably by instructing make to only build a portion of the
1589 product.
1591 #### Building Individual Modules
1593 The safe way to use fine-grained make targets is to use the module specific
1594 make targets. All source code in JDK 9 is organized so it belongs to a module,
1595 e.g. `java.base` or `jdk.jdwp.agent`. You can build only a specific module, by
1596 giving it as make target: `make jdk.jdwp.agent`. If the specified module
1597 depends on other modules (e.g. `java.base`), those modules will be built first.
1599 You can also specify a set of modules, just as you can always specify a set of
1600 make targets: `make jdk.crypto.cryptoki jdk.crypto.ec jdk.crypto.mscapi
1601 jdk.crypto.ucrypto`
1603 #### Building Individual Module Phases
1605 The build process for each module is divided into separate phases. Not all
1606 modules need all phases. Which are needed depends on what kind of source code
1607 and other artifact the module consists of. The phases are:
1609   * `gensrc` (Generate source code to compile)
1610   * `gendata` (Generate non-source code artifacts)
1611   * `copy` (Copy resource artifacts)
1612   * `java` (Compile Java code)
1613   * `launchers` (Compile native executables)
1614   * `libs` (Compile native libraries)
1615   * `rmic` (Run the `rmic` tool)
1617 You can build only a single phase for a module by using the notation
1618 `$MODULE-$PHASE`. For instance, to build the `gensrc` phase for `java.base`,
1619 use `make java.base-gensrc`.
1621 Note that some phases may depend on others, e.g. `java` depends on `gensrc` (if
1622 present). Make will build all needed prerequisites before building the
1623 requested phase.
1625 #### Skipping the Dependency Check
1627 When using an iterative development style with frequent quick rebuilds, the
1628 dependency check made by make can take up a significant portion of the time
1629 spent on the rebuild. In such cases, it can be useful to bypass the dependency
1630 check in make.
1632 > **Note that if used incorrectly, this can lead to a broken build!**
1634 To achieve this, append `-only` to the build target. For instance, `make
1635 jdk.jdwp.agent-java-only` will *only* build the `java` phase of the
1636 `jdk.jdwp.agent` module. If the required dependencies are not present, the
1637 build can fail. On the other hand, the execution time measures in milliseconds.
1639 A useful pattern is to build the first time normally (e.g. `make
1640 jdk.jdwp.agent`) and then on subsequent builds, use the `-only` make target.
1642 #### Rebuilding Part of java.base (JDK\_FILTER)
1644 If you are modifying files in `java.base`, which is the by far largest module
1645 in OpenJDK, then you need to rebuild all those files whenever a single file has
1646 changed. (This inefficiency will hopefully be addressed in JDK 10.)
1648 As a hack, you can use the make control variable `JDK_FILTER` to specify a
1649 pattern that will be used to limit the set of files being recompiled. For
1650 instance, `make java.base JDK_FILTER=javax/crypto` (or, to combine methods,
1651 `make java.base-java-only JDK_FILTER=javax/crypto`) will limit the compilation
1652 to files in the `javax.crypto` package.
1654 ### Learn About Mercurial
1656 To become an efficient OpenJDK developer, it is recommended that you invest in
1657 learning Mercurial properly. Here are some links that can get you started:
1659   * [Mercurial for git users](http://www.mercurial-scm.org/wiki/GitConcepts)
1660   * [The official Mercurial tutorial](http://www.mercurial-scm.org/wiki/Tutorial)
1661   * [hg init](http://hginit.com/)
1662   * [Mercurial: The Definitive Guide](http://hgbook.red-bean.com/read/)
1664 ## Understanding the Build System
1666 This section will give you a more technical description on the details of the
1667 build system.
1669 ### Configurations
1671 The build system expects to find one or more configuration. These are
1672 technically defined by the `spec.gmk` in a subdirectory to the `build`
1673 subdirectory. The `spec.gmk` file is generated by `configure`, and contains in
1674 principle the configuration (directly or by files included by `spec.gmk`).
1676 You can, in fact, select a configuration to build by pointing to the `spec.gmk`
1677 file with the `SPEC` make control variable, e.g. `make SPEC=$BUILD/spec.gmk`.
1678 While this is not the recommended way to call `make` as a user, it is what is
1679 used under the hood by the build system.
1681 ### Build Output Structure
1683 The build output for a configuration will end up in `build/<configuration
1684 name>`, which we refer to as `$BUILD` in this document. The `$BUILD` directory
1685 contains the following important directories:
1687 ```
1688 buildtools/
1689 configure-support/
1690 hotspot/
1691 images/
1692 jdk/
1693 make-support/
1694 support/
1695 test-results/
1696 test-support/
1697 ```
1699 This is what they are used for:
1701   * `images`: This is the directory were the output of the `*-image` make
1702     targets end up. For instance, `make jdk-image` ends up in `images/jdk`.
1704   * `jdk`: This is the "exploded image". After `make jdk`, you will be able to
1705     launch the newly built JDK by running `$BUILD/jdk/bin/java`.
1707   * `test-results`: This directory contains the results from running tests.
1709   * `support`: This is an area for intermediate files needed during the build,
1710     e.g. generated source code, object files and class files. Some noteworthy
1711     directories in `support` is `gensrc`, which contains the generated source
1712     code, and the `modules_*` directories, which contains the files in a
1713     per-module hierarchy that will later be collapsed into the `jdk` directory
1714     of the exploded image.
1716   * `buildtools`: This is an area for tools compiled for the build platform
1717     that are used during the rest of the build.
1719   * `hotspot`: This is an area for intermediate files needed when building
1720     hotspot.
1722   * `configure-support`, `make-support` and `test-support`: These directories
1723     contain files that are needed by the build system for `configure`, `make`
1724     and for running tests.
1726 ### Fixpath
1728 Windows path typically look like `C:\User\foo`, while Unix paths look like
1729 `/home/foo`. Tools with roots from Unix often experience issues related to this
1730 mismatch when running on Windows.
1732 In the OpenJDK build, we always use Unix paths internally, and only just before
1733 calling a tool that does not understand Unix paths do we convert them to
1734 Windows paths.
1736 This conversion is done by the `fixpath` tool, which is a small wrapper that
1737 modifies unix-style paths to Windows-style paths in command lines. Fixpath is
1738 compiled automatically by `configure`.
1740 ### Native Debug Symbols
1742 Native libraries and executables can have debug symbol (and other debug
1743 information) associated with them. How this works is very much platform
1744 dependent, but a common problem is that debug symbol information takes a lot of
1745 disk space, but is rarely needed by the end user.
1747 The OpenJDK supports different methods on how to handle debug symbols. The
1748 method used is selected by `--with-native-debug-symbols`, and available methods
1749 are `none`, `internal`, `external`, `zipped`.
1751   * `none` means that no debug symbols will be generated during the build.
1753   * `internal` means that debug symbols will be generated during the build, and
1754     they will be stored in the generated binary.
1756   * `external` means that debug symbols will be generated during the build, and
1757     after the compilation, they will be moved into a separate `.debuginfo` file.
1758     (This was previously known as FDS, Full Debug Symbols).
1760   * `zipped` is like `external`, but the .debuginfo file will also be zipped
1761     into a `.diz` file.
1763 When building for distribution, `zipped` is a good solution. Binaries built
1764 with `internal` is suitable for use by developers, since they facilitate
1765 debugging, but should be stripped before distributed to end users.
1767 ### Autoconf Details
1769 The `configure` script is based on the autoconf framework, but in some details
1770 deviate from a normal autoconf `configure` script.
1772 The `configure` script in the top level directory of OpenJDK is just a thin
1773 wrapper that calls `common/autoconf/configure`. This in turn provides
1774 functionality that is not easily expressed in the normal Autoconf framework,
1775 and then calls into the core of the `configure` script, which is the
1776 `common/autoconf/generated-configure.sh` file.
1778 As the name implies, this file is generated by Autoconf. It is checked in after
1779 regeneration, to alleviate the common user to have to install Autoconf.
1781 The build system will detect if the Autoconf source files have changed, and
1782 will trigger a regeneration of `common/autoconf/generated-configure.sh` if
1783 needed. You can also manually request such an update by `bash
1784 common/autoconf/autogen.sh`.
1786 If you make changes to the build system that requires a re-generation, note the
1787 following:
1789   * You must use *exactly* version 2.69 of autoconf for your patch to be
1790     accepted. This is to avoid spurious changes in the generated file. Note
1791     that Ubuntu 16.04 ships a patched version of autoconf which claims to be
1792     2.69, but is not.
1794   * You do not need to include the generated file in reviews.
1796   * If the generated file needs updating, the Oracle JDK closed counter-part
1797     will also need to be updated. It is very much appreciated if you ask for an
1798     Oracle engineer to sponsor your push so this can be made in tandem.
1800 ### Developing the Build System Itself
1802 This section contains a few remarks about how to develop for the build system
1803 itself. It is not relevant if you are only making changes in the product source
1804 code.
1806 While technically using `make`, the make source files of the OpenJDK does not
1807 resemble most other Makefiles. Instead of listing specific targets and actions
1808 (perhaps using patterns), the basic modus operandi is to call a high-level
1809 function (or properly, macro) from the API in `make/common`. For instance, to
1810 compile all classes in the `jdk.internal.foo` package in the `jdk.foo` module,
1811 a call like this would be made:
1813 ```
1814 $(eval $(call SetupJavaCompilation, BUILD_FOO_CLASSES, \
1816     SRC := $(JDK_TOPDIR)/src/jkd.foo/share/classes, \
1817     INCLUDES := jdk/internal/foo, \
1818     BIN := $(SUPPORT_OUTPUTDIR)/foo_classes, \
1819 ))
1820 ```
1822 By encapsulating and expressing the high-level knowledge of *what* should be
1823 done, rather than *how* it should be done (as is normal in Makefiles), we can
1824 build a much more powerful and flexible build system.
1826 Correct dependency tracking is paramount. Sloppy dependency tracking will lead
1827 to improper parallelization, or worse, race conditions.
1829 To test for/debug race conditions, try running `make JOBS=1` and `make
1830 JOBS=100` and see if it makes any difference. (It shouldn't).
1832 To compare the output of two different builds and see if, and how, they differ,
1833 run `$BUILD1/compare.sh -o $BUILD2`, where `$BUILD1` and `$BUILD2` are the two
1834 builds you want to compare.
1836 To automatically build two consecutive versions and compare them, use
1837 `COMPARE_BUILD`. The value of `COMPARE_BUILD` is a set of variable=value
1838 assignments, like this:
1839 ```
1840 make COMPARE_BUILD=CONF=--enable-new-hotspot-feature:MAKE=hotspot
1841 ```
1842 See `make/InitSupport.gmk` for details on how to use `COMPARE_BUILD`.
1844 To analyze build performance, run with `LOG=trace` and check `$BUILD/build-trace-time.log`.
1845 Use `JOBS=1` to avoid parallelism.
1847 Please check that you adhere to the [Code Conventions for the Build System](
1848 http://openjdk.java.net/groups/build/doc/code-conventions.html) before
1849 submitting patches. Also see the section in [Autoconf Details](
1850 #autoconf-details) about the generated configure script.
1852 ## Contributing to OpenJDK
1854 So, now you've build your OpenJDK, and made your first patch, and want to
1855 contribute it back to the OpenJDK community.
1857 First of all: Thank you! We gladly welcome your contribution to the OpenJDK.
1858 However, please bear in mind that OpenJDK is a massive project, and we must ask
1859 you to follow our rules and guidelines to be able to accept your contribution.
1861 The official place to start is the ['How to contribute' page](
1862 http://openjdk.java.net/contribute/). There is also an official (but somewhat
1863 outdated and skimpy on details) [Developer's Guide](
1864 http://openjdk.java.net/guide/).
1866 If this seems overwhelming to you, the Adoption Group is there to help you! A
1867 good place to start is their ['New Contributor' page](
1868 https://wiki.openjdk.java.net/display/Adoption/New+Contributor), or start
1869 reading the comprehensive [Getting Started Kit](
1870 https://adoptopenjdk.gitbooks.io/adoptopenjdk-getting-started-kit/en/). The
1871 Adoption Group will also happily answer any questions you have about
1872 contributing. Contact them by [mail](
1873 http://mail.openjdk.java.net/mailman/listinfo/adoption-discuss) or [IRC](
1874 http://openjdk.java.net/irc/).
1876 ---
1877 # Override styles from the base CSS file that are not ideal for this document.
1878 header-includes:
1879  - '<style type="text/css">pre, code, tt { color: #1d6ae5; }</style>'
1880 ---