Python bindings in 3.0rc1

Clang Frontend
#1

Hi,

I wasn't aware of the 3.0 branching when I posted some patches for the
python bindings to cfe-commits. I just want to make some
clarifications on the patches in the light of this. (I hope I got the
right code owners).

* [PATCH 1/4] [PYTHON] Add xdata field to Cursor structures.

  This is a bugfix. Without this the python bindings are non functional
  and basically crashes immediately.

* [PATCH 2/4] [PYTHON] Fix error message for unknown type kind.

  This is a non-important bugfix. It is a copy-paste error giving a
  confusing error message (for something that never should happen).

* [PATCH 3/4] [PYTHON] Add TypeKind.CONSTANTARRAY

  This is a bugfix. Without this fix one get an exception when trying
  to get the kind of a constantarray type object in python.
  
* [PATCH 4/4] [PYTHON] Add support for constant arrays.

  This adds support for getting the element type and array size on
  constant array type objects. So this is more of a new feature.

anders

#2

Hi,

I wasn't aware of the 3.0 branching when I posted some patches for the
python bindings to cfe-commits. I just want to make some
clarifications on the patches in the light of this. (I hope I got the
right code owners).

* [PATCH 1/4] [PYTHON] Add xdata field to Cursor structures.

This is a bugfix. Without this the python bindings are non functional
and basically crashes immediately.

Okay. Bill, please merge Clang r142474 to the release branch.

* [PATCH 2/4] [PYTHON] Fix error message for unknown type kind.

This is a non-important bugfix. It is a copy-paste error giving a
confusing error message (for something that never should happen).

Not important for the release.

* [PATCH 3/4] [PYTHON] Add TypeKind.CONSTANTARRAY

This is a bugfix. Without this fix one get an exception when trying
to get the kind of a constantarray type object in python.

Okay. Bill, please merge Clang r142476 to the release branch.

* [PATCH 4/4] [PYTHON] Add support for constant arrays.

This adds support for getting the element type and array size on
constant array type objects. So this is more of a new feature.

Right. Not important for the release (but appreciated nonetheless :slight_smile: )

  - Doug

#3

Okay. Done.

[Irk:3.0-release] ./merge.sh -r 142474 -p cfe
# Updating tree
At revision 142721.
# Merging r142474 into cfe
--- Merging r142474 into '.':
U bindings/python/clang/cindex.py
# Committing changes
Sending .
Sending bindings/python/clang/cindex.py
Sending test/SemaCXX/warn-unreachable.cpp
Transmitting file data .
Committed revision 142722.
[Irk:3.0-release] ./merge.sh -r 142476 -p cfe
# Updating tree
At revision 142722.
# Merging r142476 into cfe
--- Merging r142476 into '.':
U bindings/python/tests/cindex/test_type.py
U bindings/python/clang/cindex.py
# Committing changes
Sending .
Sending bindings/python/clang/cindex.py
Sending bindings/python/tests/cindex/test_type.py
Sending test/SemaCXX/warn-unreachable.cpp
Transmitting file data ..
Committed revision 142723.

-bw

#4

Neither of the patches touched warn-unreachable.cpp yet they show up
here? Those merges can also be seen in the log for that file.

http://llvm.org/viewvc/llvm-project/cfe/branches/release_30/test/SemaCXX/warn-unreachable.cpp?view=log

But least they are empty diffs.

anders

Okay. Done.

[Irk:3.0-release] ./merge.sh -r 142474 -p cfe

...

Sending test/SemaCXX/warn-unreachable.cpp

???

Committed revision 142722.

[Irk:3.0-release] ./merge.sh -r 142476 -p cfe

...

Sending test/SemaCXX/warn-unreachable.cpp

???

#5

It's an artifact of the 'svn merge' command. It's harmless, but annoying. :frowning:

-bw