tensorflow: workaround tensorboard collision from upstream pip hack

Upstream lists tensorboard as a utility from tensorflow despite it
being provided by tensorboard to stop pip for deleting it.  This
causes python.withPackages collisions between the two.

1 files changed, 9 insertions, 0 deletions

diff --git a/pkgs/development/python-modules/tensorflow/bin.nix b/pkgs/development/python-modules/tensorflow/bin.nix
index 4e3621157f93..42e35df7497a 100644
--- a/pkgs/development/python-modules/tensorflow/bin.nix
+++ b/pkgs/development/python-modules/tensorflow/bin.nix
@@ -61,6 +61,15 @@ in buildPythonPackage rec {
   # bleach) Hence we disable dependency checking for now.
   installFlags = lib.optional isPy36 "--no-dependencies";
 
+
+  # Upstream has a pip hack that results in bin/tensorboard being in both tensorflow
+  # and the propageted input tensorflow-tensorboard which causes environment collisions.
+  #
+  # https://github.com/tensorflow/tensorflow/blob/v1.7.1/tensorflow/tools/pip_package/setup.py#L79
+  postInstall = ''
+    rm $out/bin/tensorboard
+  '';
+
   # Note that we need to run *after* the fixup phase because the
   # libraries are loaded at runtime. If we run in preFixup then
   # patchelf --shrink-rpath will remove the cuda libraries.