Skip to content

Conversation

@pritamdamania87
Copy link
Contributor

@pritamdamania87 pritamdamania87 commented Nov 1, 2019

Stack from ghstack:

  1. Enhanced autograd unit tests to test the
    torch.distributed.autograd.backward() API more thoroughly on Python UDFs.
  2. Enhanced python_error to override what such that it returns an
    appropriate error string if we call what() on this error. This ensures we can
    propagate exceptions over the wire during RPCs (since we get the error string
    by calling what() on the exception)

Differential Revision: D18273041

1) Enhanced autograd unit tests to test the
torch.distributed.autograd.backward() API more thoroughly on Python UDFs.
2) Enhanced `python_error` to override `what` such that it returns an
appropriate error string if we call `what()` on this error. This ensures we can
propagate exceptions over the wire during RPCs (since we get the error string
by calling what() on the exception)
ghstack-source-id: 92972494

Differential Revision: [D18273041](https://our.internmc.facebook.com/intern/diff/D18273041/)

[ghstack-poisoned]
pritamdamania87 pushed a commit that referenced this pull request Nov 1, 2019
1) Enhanced autograd unit tests to test the
torch.distributed.autograd.backward() API more thoroughly on Python UDFs.
2) Enhanced `python_error` to override `what` such that it returns an
appropriate error string if we call `what()` on this error. This ensures we can
propagate exceptions over the wire during RPCs (since we get the error string
by calling what() on the exception)
ghstack-source-id: 92972494

Differential Revision: [D18273041](https://our.internmc.facebook.com/intern/diff/D18273041/)

ghstack-source-id: 93087189
Pull Request resolved: #29041
@pritamdamania87
Copy link
Contributor Author

@ezyang I'm not able to reproduce the doxygen issue locally. This is what I did and check-doxygen.sh completed successfully:

cd docs/
make html
cd cpp/source
./check-doxygen.sh

check-doxygen.sh output:

++ dirname ./check-doxygen.sh
+ pushd ./../../..
~/local/pytorch ~/local/pytorch/docs/cpp/source
+ cp aten/src/ATen/common_with_cwrap.py tools/shared/cwrap_common.py
+ cp torch/_utils_internal.py tools/shared
+ python aten/src/ATen/gen.py -s aten/src/ATen -d build/aten/src/ATen aten/src/ATen/Declarations.cwrap aten/src/THNN/generic/THNN.h aten/src/THCUNN/generic/THCUNN.h aten/src/ATen/nn.yaml aten/src/ATen/native/native_functions.yaml
+ python tools/setup_helpers/generate_code.py --declarations-path build/aten/src/ATen/Declarations.yaml --nn-path aten/src
Skipped writing torch/csrc/autograd/generated/python_functions.h
Skipped writing torch/csrc/autograd/generated/python_functions.cpp
Skipped writing torch/csrc/autograd/generated/python_variable_methods.cpp
Skipped writing torch/csrc/autograd/generated/python_variable_methods_dispatch.h
Skipped writing torch/csrc/autograd/generated/python_torch_functions.cpp
Skipped writing torch/csrc/autograd/generated/python_torch_functions_dispatch.h
Skipped writing torch/csrc/autograd/generated/python_nn_functions.cpp
Skipped writing torch/csrc/autograd/generated/python_nn_functions.h
Skipped writing torch/csrc/autograd/generated/python_nn_functions_dispatch.h
Skipped writing torch/csrc/autograd/generated/VariableType.h
Skipped writing torch/csrc/autograd/generated/VariableType_0.cpp
Skipped writing torch/csrc/autograd/generated/VariableType_1.cpp
Skipped writing torch/csrc/autograd/generated/VariableType_2.cpp
Skipped writing torch/csrc/autograd/generated/VariableType_3.cpp
Skipped writing torch/csrc/autograd/generated/VariableType_4.cpp
Skipped writing torch/csrc/autograd/generated/VariableTypeEverything.cpp
Skipped writing torch/csrc/autograd/generated/Functions.h
Skipped writing torch/csrc/autograd/generated/Functions.cpp
Skipped writing torch/csrc/autograd/generated/variable_factories.h
Skipped writing torch/csrc/jit/generated/register_aten_ops_0.cpp
Skipped writing torch/csrc/jit/generated/register_aten_ops_1.cpp
Skipped writing torch/csrc/jit/generated/register_aten_ops_2.cpp
+ popd
~/local/pytorch/docs/cpp/source
+ doxygen
+ cp original-doxygen-log.txt doxygen-log.txt
+ ignore_warning 'warning: no uniquely matching class member found for'
+ set +e
+ grep -v 'warning: no uniquely matching class member found for' doxygen-log.txt
+ set -e
+ mv temp.txt doxygen-log.txt
+ ignore_warning 'warning: explicit link request to '\''Item'\'' could not be resolved'
+ set +e
+ grep -v 'warning: explicit link request to '\''Item'\'' could not be resolved' doxygen-log.txt
+ set -e
+ mv temp.txt doxygen-log.txt
+ ignore_warning 'warning: Included by graph for '\''types.h'\'' not generated, too many nodes'
+ set +e
+ grep -v 'warning: Included by graph for '\''types.h'\'' not generated, too many nodes' doxygen-log.txt
+ set -e
+ mv temp.txt doxygen-log.txt
++ grep warning: doxygen-log.txt
++ wc -l
+ warnings=0
+ echo 'Treating all remaining warnings as errors'
Treating all remaining warnings as errors
+ [[ 0 -ne 0 ]]
+ rm -f doxygen-log.txt original-doxygen-log.txt

@ezyang
Copy link
Contributor

ezyang commented Nov 1, 2019

This error may be sensitive to doxygen or dot version, and you may not have lined things up exactly the same way.

I was reading through the CI script and it looks like we filter out similar looking warnings. So it might be acceptable to just filter this one out too.

@ezyang
Copy link
Contributor

ezyang commented Nov 1, 2019

1) Enhanced autograd unit tests to test the
torch.distributed.autograd.backward() API more thoroughly on Python UDFs.
2) Enhanced `python_error` to override `what` such that it returns an
appropriate error string if we call `what()` on this error. This ensures we can
propagate exceptions over the wire during RPCs (since we get the error string
by calling what() on the exception)

Differential Revision: [D18273041](https://our.internmc.facebook.com/intern/diff/D18273041/)

[ghstack-poisoned]
pritamdamania87 pushed a commit that referenced this pull request Nov 1, 2019
Pull Request resolved: #29041

1) Enhanced autograd unit tests to test the
torch.distributed.autograd.backward() API more thoroughly on Python UDFs.
2) Enhanced `python_error` to override `what` such that it returns an
appropriate error string if we call `what()` on this error. This ensures we can
propagate exceptions over the wire during RPCs (since we get the error string
by calling what() on the exception)
ghstack-source-id: 93098679
ghstack-source-id: 93098679

Differential Revision: [D18273041](https://our.internmc.facebook.com/intern/diff/D18273041/)
@pritamdamania87
Copy link
Contributor Author

you could also do this https://github.com/pytorch/pytorch/pull/26613/files#diff-90c72f4e4e61cafd042a47b211ed4e05

Followed what Alban did there and quick_checks are now passing. Could you approve this PR? Thanks!

Copy link
Contributor

@mrshenli mrshenli left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is a retry of #28824, and the only difference is in docs/cpp/source/Doxyfile.

Lint failures on python_engine.cpp is irrelevant.

@facebook-github-bot
Copy link
Contributor

This pull request has been merged in e8e7d93.

@facebook-github-bot facebook-github-bot deleted the gh/pritamdamania87/20/head branch November 5, 2019 15:18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants