运行命令“bazel build tensorflow/examples/image_retraining:retrain”时构建失败

Question

我正在尝试使用 python 2.7 在没有 GPU 的 Ubuntu 14.04 LTS 上构建 tensorflow。当我在终端上运行以下命令时，紧随其后的是 tutorial:

bazel build tensorflow/examples/image_retraining:retrain

它说使用日志构建失败：

ERROR: /home/yuan/tensorflow_source/tensorflow/tensorflow/python/BUILD:1826:1: Linking of rule '//tensorflow/python:_pywrap_tensorflow.so' failed: gcc failed: error executing command 
  (cd /home/yuan/.cache/bazel/_bazel_yuan/e5b8b6538ba16bbae5b1e0f5c26b7a12/execroot/tensorflow && \
  exec env - \
  /usr/bin/gcc -shared -o bazel-out/local-fastbuild/bin/tensorflow/python/_pywrap_tensorflow.so -Wl,--version-script tensorflow/tf_version_script.lds -pthread -Wl,-no-as-needed -B/usr/bin -B/usr/bin -pass-exit-codes '-Wl,--build-id=md5' '-Wl,--hash-style=gnu' -Wl,-S -Wl,@bazel-out/local-fastbuild/bin/tensorflow/python/_pywrap_tensorflow.so-2.params): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 1.
bazel-out/local-fastbuild/bin/tensorflow/core/libversion_lib.a(version_info.pic.o): In function `tf_git_version()':
version_info.cc:(.text+0x0): multiple definition of `tf_git_version()'
bazel-out/local-fastbuild/bin/tensorflow/core/libframework_internal.lo(version_info.pic.o):version_info.cc:(.text+0x0): first defined here
bazel-out/local-fastbuild/bin/tensorflow/core/libversion_lib.a(version_info.pic.o): In function `tf_compiler_version()':
version_info.cc:(.text+0xd): multiple definition of `tf_compiler_version()'
bazel-out/local-fastbuild/bin/tensorflow/core/libframework_internal.lo(version_info.pic.o):version_info.cc:(.text+0xd): first defined here
collect2: error: ld returned 1 exit status
Target //tensorflow/examples/image_retraining:retrain failed to build
INFO: Elapsed time: 52.885s, Critical Path: 34.63s

这是我的构建环境：

操作系统：Ubuntu 14.04 LTS 64 位，RAM 12Gib
gcc 版本：4.8.4
蟒蛇版本：2.7.6
bazel 版本：0.3.2
tensorflow源的git版本：v0.11.0rc0-1541-g3737ac3

有没有人想办法解决这个错误？非常感谢！

Answer 1

@David 感谢您的帮助，看来我的 ./configure 结果与预期不符。

我终于通过以下步骤构建成功

1. 将 tensorflow 源同步到 HEAD

2. 再次运行./configure，如有错误请参考this。

3. 再次运行构建命令

   bazel 构建 tensorflow/examples/image_retraining:retrain

而且我的构建结果不会有错误。