tensorflow: Errors while building Tensorflow 1.12

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Void Linux glibc x86_64
• TensorFlow installed from (source or binary): Source
• TensorFlow version: r1.12
• Python version: 3.6
• Installed using virtualenv? pip? conda?: pip, no virtualenv
• Bazel version (if compiling from source): 0.22
• GCC/Compiler version (if compiling from source): 8.2.0
• CUDA/cuDNN version: No
• GPU model and memory: None

Describe the problem

Hello guys, I’m trying to build TF in a non Debian based distro. But what I’m getting is this error, as referrenced by this windows build from source, especially related to eigen.

Provide the exact sequence of commands / steps that you executed before running into the problem

bazel build --incompatible_remove_native_http_archive=false --config=opt --copt=-mavx --copt=-mavx2 --config=monolithic //tensorflow/tools/pip_package:build_pip_package
INFO: Invocation ID: 851c2110-08b2-4ed8-a7b0-4c995b8388e9
DEBUG: /home/neel/.cache/bazel/_bazel_neel/e655abc18ce719a580533b95739e1032/external/build_bazel_rules_apple/apple/repositories.bzl:35:5: 
WARNING: `build_bazel_rules_apple` depends on `bazel_skylib` loaded from https://github.com/bazelbuild/bazel-skylib.git (tag 0.6.0), but we have detected it already loaded into your workspace from None (tag None). You may run into compatibility issues. To silence this warning, pass `ignore_version_differences = True` to `apple_rules_dependencies()`.

INFO: Build options --compiler and --cpu have changed, discarding analysis cache.
ERROR: /home/neel/Programming_Workspace/Not-My-Repos/tensorflow/third_party/eigen3/BUILD:34:1: no such package '@eigen_archive//': Traceback (most recent call last):
	File "/home/neel/Programming_Workspace/Not-My-Repos/tensorflow/third_party/repo.bzl", line 106
		_apply_patch(ctx, ctx.attr.patch_file)
	File "/home/neel/Programming_Workspace/Not-My-Repos/tensorflow/third_party/repo.bzl", line 68, in _apply_patch
		fail("patch command is not found, ple...")
patch command is not found, please install it and referenced by '//third_party/eigen3:eigen3'
ERROR: Analysis of target '//tensorflow/tools/pip_package:build_pip_package' failed; build aborted: no such package '@eigen_archive//': Traceback (most recent call last):
	File "/home/neel/Programming_Workspace/Not-My-Repos/tensorflow/third_party/repo.bzl", line 106
		_apply_patch(ctx, ctx.attr.patch_file)
	File "/home/neel/Programming_Workspace/Not-My-Repos/tensorflow/third_party/repo.bzl", line 68, in _apply_patch
		fail("patch command is not found, ple...")
patch command is not found, please install it
INFO: Elapsed time: 3.142s
INFO: 0 processes.
FAILED: Build did NOT complete successfully (42 packages loaded, 470 targets c\
onfigured)
    currently loading: tensorflow/core/kernels
    Fetching @eigen_archive; fetching
    Fetching @double_conversion; fetching

Any other info / logs

I just cloned the latest repo from git and checked out TF r1.12 last commit. After that I passed only CPU options to “configure” and also had to install pip3. I also installed eigen3 package from my distro’s repos, but no help. I don’t know what to do, eigen does seem to be in the TF/tools/ folder but bazel doesn’t import it. Please help !!

Here’s the folder structure of “third_parties” folder:

ls ./third_party/
android                     libxsmm.BUILD
arm_neon_2_x86_sse.BUILD    linenoise.BUILD
astor.BUILD                 llvm
aws                         lmdb.BUILD
backports_weakref.BUILD     mkl
boringssl                   mkl_dnn
BUILD                       mpi
clang_toolchain             mpi_collectives
codegen.BUILD               nanopb.BUILD
com_google_absl.BUILD       nasm
common.bzl                  nccl
cub.BUILD                   ngraph
curl.BUILD                  ortools
cython.BUILD                pasta
double_conversion.BUILD     pcre.BUILD
eigen3                      png.BUILD
eigen.BUILD                 png_fix_rpi.patch
enum34.BUILD                pprof.BUILD
examples                    protobuf
farmhash.BUILD              py
fft2d                       python_runtime
flatbuffers                 repo.bzl
gast.BUILD                  six.BUILD
gif.BUILD                   snappy.BUILD
git                         sqlite.BUILD
googleapis.BUILD            swig.BUILD
gpus                        sycl
grpc                        systemlibs
hadoop                      tensorrt
highwayhash                 termcolor.BUILD
hwloc                       tflite_mobilenet.BUILD
icu                         tflite_mobilenet_float.BUILD
jpeg                        tflite_mobilenet_quant.BUILD
jsoncpp.BUILD               tflite_ovic_testdata.BUILD
kafka                       tflite_smartreply.BUILD
keras_applications_archive  toolchains
kissfft                     zlib.BUILD

Heres my git branches:

git branch
  master
* r1.12

git show --oneline r1.12

a6d8ffae09 (HEAD -> r1.12, tag: v1.12.0) Fix a bug in tpu.py and xla.py that while creating an identity node for control input edges under rewrite context, the parent control flow context is lost. (#23446)
diff --git a/tensorflow/contrib/compiler/xla.py b/tensorflow/contrib/compiler/xla.py
index 873b03580d..83d9d8c54a 100644
--- a/tensorflow/contrib/compiler/xla.py
+++ b/tensorflow/contrib/compiler/xla.py
@@ -179,14 +179,11 @@ class XLACompileContext(control_flow_ops.XLAControlFlowContext):
     if external_control_inputs:
       # Use an identity to pull control inputs as data inputs. Note that we
       # ignore ops which don't have outputs. TODO(phawkins): fix that.
-      with ops.control_dependencies(None):
-        self.Enter()
-        external_control_inputs = [
-            array_ops.identity(x.outputs[0]).op
-            for x in external_control_inputs
-            if x.outputs
-        ]
-        self.Exit()
+      external_control_inputs = [
+          array_ops.identity(x.outputs[0]).op