Both function annotations and variable annotations are turned on using the
"python:annotations" feature. Example:
%feature("python:annotations", "c");
struct V {
float val;
};
The generated code contains a variable annotation containing the C float type:
class V(object):
val: "float" = property(_example.V_val_get, _example.V_val_set)
...
Python 3.5 and earlier do not support variable annotations, so variable
annotations can be turned off with a "python:annotations:novar" feature flag.
Example turning on function annotations but not variable annotations globally:
%feature("python:annotations", "c");
%feature("python:annotations:novar");
or via the command line:
-features python:annotations=c,python:annotations:novar
Closes#1951
Testing is skipped where there is no support for it, that is,
using -builtin or -fastproxy. How to add this support in needs
determining, it's not clear how to do so.
Python function annotations containing C/C++ types are no longer
generated when using the -py3 option. Function annotations support
has been moved to a feature to provide finer grained control.
It can be turned on globally by adding:
%feature("python:annotations", "c");
or by using the command line argument:
-features python:annotations=c
The implementation is designed to be expandable to support different
annotations implementations. Future implementations could implement
something like the following for generating pure Python types:
%feature("python:annotations", "python");
or typing module types to conform to PEP-484:
%feature("python:annotations", "typing");
Closes#1561
Issue #735
- Improved documentation for using declarations.
- Issue new warning WARN_LANG_USING_NAME_DIFFERENT when there
is a conflict in the target language name to be used when
introducing a method via a using declaration. Previously
the method was silently ignored. Issue #1840. Issue #655.
Even in the case of just creating a `DerivedClass` this test says:
```
swig/python detected a memory leak of type 'DerivedClass *', no destructor found.
```
even though the destructor is defined in the base class.
Fix access to C++ static member functions using Python class
staticmethod syntax, such as Klass.memberfunction instead of
Klass_memberfunction, when using -fastproxy and -builtin in
combination with %callback.
The docstring containing the callback pointers were not being patched
during module initialisation.
Accept keyword arguments accessing C++ static member functions when
using -builtin and kwargs feature and Python class staticmethod syntax.
The missing keyword argument support was only when using the
class staticmethod syntax, not when using the flat static method
syntax.
Use Python class staticmethod syntax to access C++ static member functions,
such as Klass.memberfunction, instead of Klass_memberfunction.
Examples and test-suite changes in preparation for issue #2137.
- Document change in CHANGES file
- Minor tweaks and whitespace fixes in stype.c
- Enhance testcase
- Synchronise Java and Python runt test in testcase
Closes#1977Fixes#1603
Squashed commit of the following:
commit 61f794184e
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 20 15:20:25 2021 +0100
added python version for second part of testcase
commit bb80e236b8
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 20 15:10:14 2021 +0100
reworked comment in fixed code
commit 5a94bcc481
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 20 15:07:34 2021 +0100
removed new unittests; I switched to the official test cases
commit 36603f3c8d
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 20 15:05:33 2021 +0100
removed new unittests; I switched to the official test cases
commit 1f20ea00d2
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 20 14:39:19 2021 +0100
adapted testcase
commit ab492794c0
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 20 14:29:10 2021 +0100
fixed test
commit 9b5dd0c8f9
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 20 13:04:06 2021 +0100
C99 compatible comments
commit 1a89425ac8
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 20 11:41:48 2021 +0100
added extra checks to prevent accidental partial matches
commit 20e76f511a
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Thu Mar 18 21:42:56 2021 +0100
minor
commit 0e383bbb76
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Thu Mar 18 21:41:45 2021 +0100
cleanup
commit b644767121
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Thu Mar 18 21:39:43 2021 +0100
cleanup
commit 2574468c0f
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Thu Mar 18 21:38:35 2021 +0100
cleanup
commit 9dc7f1ed30
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Thu Mar 18 21:35:28 2021 +0100
removed ide files
commit a442a9df46
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Thu Mar 18 21:32:36 2021 +0100
removed printf
commit da4c6e91e3
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Thu Mar 18 21:29:02 2021 +0100
possible fix found
commit 6fad8d40e3
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Thu Mar 18 21:01:11 2021 +0100
found difference between good/bad case... next: analyze why replacement did not work here...
commit 72a7693340
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Wed Mar 17 22:43:08 2021 +0100
ideintified location where the template-template type is inserted in the final type (with the missing specialization).
commit 1b53312c07
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Wed Mar 17 22:28:19 2021 +0100
more notes (to be reverted)
commit a9a0b58938
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Wed Mar 17 22:19:16 2021 +0100
some experiments (to be reverted)
commit 0e7a24bbd5
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Wed Mar 17 21:46:20 2021 +0100
added some notes
commit 2f77911a12
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Tue Mar 16 22:01:01 2021 +0100
comment changed
commit 2cb7213b06
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Tue Mar 16 21:55:47 2021 +0100
renamed example template parameters to easily distinguish them.
commit ff457d7397
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Tue Mar 16 21:21:10 2021 +0100
added reset parser functionality (required for unittests, to reset parser)
commit 617bbde3b4
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Tue Mar 16 20:47:41 2021 +0100
adjusted test for simple templates
commit beb7e7f77c
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Mon Mar 15 07:41:17 2021 +0100
added note how it should be...
commit 7b3328431c
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 17:07:24 2021 +0100
found a place which is maybe problematic for the template_template problem (#1603)
commit 46c2443d15
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 16:11:56 2021 +0100
unitests: reset parser for each testcase
commit b3a0f1516f
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 15:30:47 2021 +0100
first experiment with templates
commit 32a11c6c77
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 14:31:17 2021 +0100
wip
commit 37b805ba6e
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 14:21:05 2021 +0100
integrated unittests in ctest
commit 79f7bee168
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 14:18:26 2021 +0100
wip: new test created; problem: not in c++ mode
commit 345d503d55
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 14:08:09 2021 +0100
cleanup test code
commit 0a26adec10
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 14:05:24 2021 +0100
some more tests (first steps with c++ code)
commit 6f628e0fa9
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sun Mar 14 12:56:41 2021 +0100
experiments
commit c4a13bf3e1
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 13 18:18:01 2021 +0100
first experiment
commit 7d26586105
Author: goto40 <pierre.bayerl@googlemail.com>
Date: Sat Mar 13 17:23:13 2021 +0100
setup catch2 framework (w/o tests)
template_typedef_cplx2 files are generated by the template_typedef_import.multicpptest
but can also be cleaned by the template_typedef_cplx2.cpptest target.
* commit '8245277ad3acd9308ce28c40508b999e9496b27e':
Remove test for unsupported complex or _Complex by itself
More C99 complex fixes, plus Python tests
Restore _Complex as standalone type
Small corrections for handling C99 _Complex
Properly handle C99 complex types even in C++ mode
Conflicts:
Examples/test-suite/python/complextest_runme.py
* remove-dependency-on-2to3:
Remove need for Python 2to3
Modify examples to be both Python 2 and 3 compatible
Remove python3 specific runme3.py test files
Convert python tests using 2to3
Convert python test scripts to be Python 2 and 3 compatible
Convert swigobject python test to be python 2 and 3 compatible
Convert two tests to work with both Python 2 and 3
Improve director_exception Python test
Remove further print statements from Python tests
Improve Python testing catching exceptions
Improve contract Python testcase testing
Remove print statements from Python tests
All Python examples and tests have been written to be both Python 2 and Python 3
compatible, removing the need for 2to3 to run the examples or test-suite.
The 2to3 executable is not always available and even when available does not
always work, e.g. with pyenv. An alternative would be to use the lib2to3 Python
module instead, but this isn't available in some older versions of Python 3.
I had this problem on Ubuntu Bionic on Travis:
checking Examples/python/callback
pyenv: 2to3-3.8: command not found
The `2to3-3.8' command exists in these Python versions:
3.8
3.8.1
Reference issues:
https://github.com/pypa/virtualenv/issues/1399https://travis-ci.community/t/2to3-command-not-found-in-venv-in-bionic/4495
This error is given for any use of variable called "l" (and also "I" and
"O", but we don't seem to have any of those) and it doesn't seem to be
worth changing this variable name in the tests code, as it's really not
that meaningful there anyhow, so just disable the warning to let the CI
builds, which now use pycodestyle 2.6, instead of 2.5 which didn't have
this error and which is still the latest available in Debian Sid, pass.
The "endlink" command is processed in processWordCommands, which by
default skips space occuring after the command, which is intended for
removing leading space from a command argument. For "end" commands,
we don't want to do this. Note that certain end commands such as
"endcode" aren't processed by processWordCommands (believe
addCommandUnique ends up handling them).
Update usage of \link in doxygen_translate_all_tags.i to test handling
of space after \endlink.
Tweaking some of the usage in doxygen_misc_constructs.i to remove what
seems to be an extra space from the input (otherwise we would need to
add an extra space to the expected output).
In addition to the changes in the previous commit, also avoid syntax
errors in the generated Python docstrings by splitting them into several
parts if there are 3 quotes in a row in the input, as it's impossible to
have them inside triple-quoted strings, generally speaking (i.e. if
there are occurrences of both """ and ''' inside the string).
Single-line Doxygen comments ending with a double quote resulted in
syntactically-invalid Python docstrings in the output, so use triple
single quotes as delimiters in this case to avoid it.
This test code tests the upcast code:
swig_assert_equal_simple(-1, base_num2([Derived(7)]))
Although there is no explicit memory leak fix test, it does at least run the code.
Handling of None needs fixing in Python (it is working in Ruby)
Note that the Ruby implementation has a partial template specialization
for shared_ptr, whereas the Python implementation is in the generic
code!
Issue #1512
Don't attempt to use the class's __dict__ for setting 'this' when
a user has extended a class with:
__slots__ = ['this'].
Was segfaulting. Now we fall back to a simple PyObject_SetAttr if the
usual approach to setting 'this' in __dict__ does not work.
Closes#1673Closes#1674
Due to confusion in build_combined_docstring(), we could call
DohDelete() on the "feature:docstring" string, which resulted in a crash
when trying to use it later.
Fix this and simplify the code at the same time by ensuring that we
always use a copy of "feature:docstring" if it's not empty or don't use
it at all if it's empty -- like this we don't have to check for its
length each time before using it.
Closes#1648.
* Issue-1632:
Minor workaround in doxygen_basic_translate_style3 test
Add new test doxygen_basic_translate_style3.i
Fix for newline handling in doxygen "///" style comments
* doxy/commands:
Update documentation for doxygen tags
Fix doxygen translation of \p command for python
Fix doxygen handling of \em tag for python
Minor formatting updates to doxygen docs
Reformat tag lists in doxygen documentation
Add doxygen_code_blocks_runme.java
Special handling for python doctest code blocks
Add new doxygen test doxygen_code_blocks
Handle doxygen code command with language option
Improve doxygen parser handling of \code content
Flag optional arguments in doxygen pydoc output
Add parameter direction to doxygen pydoc output
Support doxygen \param[] commands
Instead of silenty ignoring them, now a "TypeError: f() takes no keyword arguments"
exception is thrown if keyword arguments are used. Hence constructors and normal
methods/functions behave in the same way.
Closes issue #1595
With the "///" doxygen comment style, comments within a code block do
not get handled correctly. Modifying the test to remove this case and
adding a note about it for future reference.
The fix is when using kwargs feature or -keyword.
The fix is in the argument error checking when wrapping zero
argument constructors only. Supplied keyword args were silently
ignored.
Issue #1595
Processing doxygen @param comments for a parameter whose name did not
appear in the function declaration would cause a segfault due to a
null pointer dereference.
Adding test cases for both variadic function (no specified arguments)
and @param comment that references an argument that is not named in
the function prototype. Both of these cases previously segfaulted.
A doctest code block begins with ">>>" and is not indented. Identify
these in doxygen comments and treat them accordingly. Also add check
to padCodeAndVerbatimBlocks for these because Sphinx requires an empty
line before.
Add test case to doxygen_code_blocks.i.
This does somewhat more detailed testing of the code block parsing,
and also exercises the language identification of python doctest
features. For now, it is only tested by python (javadoc translation
may not correctly handle some of the characters that are used here).
If the parameter has a default value, add the string ", optional" to
the parameter type description in the translated python comments.
Three examples with default values were already present in the test
cases, so their expected python output has been updated accordingly.
For doxygen comments that specify parameter direction (i.e.,
\param[in], \param[out], and \param[in,out]), the direction is
appended to the type definition in the generated Python documentation.
Updated expected python output for doxygen test case.
Recognize \param[in], \param[out], and \param[in,out]. Currently they
are all treated the same as \param, but the information is now
available for inclusion in the translated comments. This is done
using new utility functions getBaseCommand and getEndOfWordCommand,
which will also generalize to treatment of code command options,
e.g. \code{.py}. Preliminary treatment of the extended version of
\code is already in place in these functions.
Added examples of all three new \param commands to the
doxygen_translate_all_tags test and updated the python and java
expected output.
This testcase was only run in Ruby and Python and implements an
obscure feature where a C++ class inherits from a std::wstring.
The li_std_wstring test is left in place to be modified in next commit
for more regular wstring testing across all languages.
* annotation_quoting:
Examples: Unify string quoting in the Python sources
Test-suite: Unify string quoting in the Python sources
Tools: Unify string quoting in the Python sources
Generator: Unify string quoting in generated Python sources
Generator: Remove trailing whitespaces in the Python module
Fix how end of paragraph is identified when parsing \code and
\verbatim blocks that appear within a paragraph and contain an empty
line. Previously this would generate a warning for unexpected end of
doxygen comments, and it could generate a segfault due to
dereferencing an invalid iterator value.
The doxygen_basic_translate.i and doxygen_basic_translate_style2.i
tests have been updated to serve as regression tests for this
behavior. Prior to this fix, inclusion of the empty code block line
in this context produced a segfault.
Sphinx requires an empty line before code and math blocks, whereas
doxygen does not. This update ensures that a blank line is included
before generated code and math blocks in the pydoc output. This is
done by post-processing the docstring line by line to check whether
any newlines need to be added. This way, if the original doxygen
source already includes an empty line before a block, an additional
unnecessary empty line is not added.
Updating the expected test output for doxygen_basic_translate, which
now adds the necessary empty line before the code block. Adding
further test cases to doxygen_translate_all_tags to explicitly verify
that a newline is added in the pydoc output before both code and math
blocks that appear within a paragraph.
Additionally, empty lines previously appearing at the beginning of the
generated docstrings are now removed. This does not alter the
behavior of the tests.
Remove a newline character that was added to the translated comments
prior to the code block. This way the structure of the pydoc output
more closely resembles that of the original doxygen comments.
Updating tests accordingly.
The _runme.py code for the recently added test case was using
references to an old module name for the test case, which was later
changed but not updated in the runme file.
Observed that with this second comment style, there is no line break
after the function overload headings in the translated
comments. Updating the test results accordingly.
This style looks like:
/** Line 1
* Line 2
*/
This is needed to verify fixes to some of the indentation in the
translated comments.
The test is copied from doxygen_basic_translate.i. One adjustment was
made to change the comment style on the last function that left out
the intermediate "*" characters. This does not produce correct output
when combined with this style of starting the text on the first
comment line.
Test results are copied directly from doxygen_basic_translate. A
minor difference in the Python results will be updated in a subsequent
commit.
Eliminate extra leading and trailing newlines around the \f[ and \f{
block math commands for doxygen comment translation for python.
Update tests accordingly.
If \code is used as a block command (probably the main use case), an
extra newline was included in the translated Python comments. This is
now removed and doxygen python test case output updated.
If \code was used inline, it produced an extra indent versus block
usage. This extra indent was also stored in the test output. This
update resolves this by simply removing a space that was being added
unnecessarily in handleTagVerbatim. Updating test case output
accordingly.
The low-level API for setting static member variables stopped working
when the fastunpack option was turned on by default. The PyMethodDef
setup requires METH_O, not METH_VARARGS with fastunpack.
* ZackerySpytz-director-classes-final-methods:
Warning tweaks for destructors that are final in director classes
Documentation for directors and virtual final methods
Fixes for final destructors in director classes
Warning fix for final destructor in directors
Remove a useless warning filter
Fix the handling of director classes with final methods
Fixes wrapping overloaded functions/constructors where a vararg
function is declared after a non-vararg function.
This is a long standing bug in the Python layer exposed since fastunpack
was turned on by default.
Generated SwigDirector_* classes were attempting to override
methods marked as final.
In addition, give a warning if the destructor of a director class is
final.
Closes#564.
Tests for std::vector of pointers added which check
std::vector<T*>::const_reference and std::vector<T*>::reference
usage which gave compilation errors in Python and Perl which had
specialized these vectors incorrectly.
Fix regression introduced by 3f5c17824c
which resulted in using "$ignore" instead of the real parameter name in
Python if an %ignore/%rename($ignore) for the parameter name was used
(as could happen not necessarily intentionally when using wild card
ignores with regex matches) by explicitly checking if we're dealing with
a parameter node in apply_rename(), used by Swig_name_make(), and not
renaming it to "$ignore" in this case.
Extend the test suite to check for this case.
Closes#1460.
One of side effects of 15b369028f was that
the default values were only included in Python doc strings if we could
be sure that they could be interpreted as valid Python expressions, but
this change was actually undesirable as it may be useful to see C++
expression for the default value in the doc string even when it isn't
valid in Python.
Undo this part of the change and extend autodoc unit test to check that
this stays fixed.
Closes#1271.
* vadz-better-param-names:
Enable keyword arguments for keyword_rename unit test
Update error messages test suite
Add more tests for Python parameter renaming
Improve handling parameters clashing with language keywords
- Static method wrappers were using the 'fastproxy' approach by default.
This is inconsistent with instance method wrappers. The fastproxy approach
is now turned off by default to be consistent with instance methods.
Static method wrappers can now also be controlled using the -fastproxy and
-olddefs options.
Example:
struct Klass {
static int statmethod(int a = 2);
};
generates:
class Klass(object):
...
@staticmethod
def statmethod(a=2):
return _example.Klass_statmethod(a)
instead of:
class Klass(object):
...
statmethod = staticmethod(_example.Klass_statmethod)
- Modernise wrappers for static methods to use decorator syntax - @staticmethod.
- Add missing runtime test for static class methods and using the actual
class method.
Test added to check fixes for:
- Issue #948 and issue #1019 and issue #1273 - raw string delimiters
not being stripped off
- Issue #538 - Ruby support for "docstring" feature
The typecheck typemaps succeed for non pointers (SWIGTYPE, SWIGTYPE&,
SWIGTYPE&&) when the equivalent to C NULL is passed from the target
language. This commit implements a fix for Python to not accept a Python
None for non-pointer types.
Issue #1202
Code handling %template in the parser created a totally new top-level
module child of namespace type when handling templates inside a
namespace and copied the nodes from the previously parsed C++ template
declaration to it. However copies of this node kept their original
values of "abstracts" attribute, which contained pointers to the classes
in the original template declaration, i.e. outside of the subtree
created for the instantiated template. This, in turn, meant that during
the types resolution pass, the code in TypePass did not update the types
used in the methods of the classes appearing in the "abstracts" List,
even though it did update the types for the children of the instantiated
template subtree.
And this finally resulted in wrongly detecting overridden virtual
methods as abstract in Allocate::is_abstract_inherit() during the next
pass, as the signatures of the overridden method -- using resolved types
-- and of the method from the class pointed to by "abstract" -- using
the original types from C++ code -- didn't match.
Resolve this simply by not copying "abstracts" attributes when creating
the template subtree and doing another pass over this tree to recreate
them using the new nodes, just as it's already done for "defaultargs"
attribute, presumably for similar reasons. Note that doing another pass
over the tree is not as efficient as doing everything in a single pass,
but merging the new update_abstracts() with update_defaultargs() is not
completely obvious, so for now keep it simple and optimize it later if
necessary.
Also, add a test checking for the situation described above.
Closes#1353.
There were only needed to support Python < 2.2, and we now require at
least Python 2.6.
Conflicts:
.travis.yml
Examples/test-suite/python/autodoc_runme.py
Source/Modules/python.cxx
This is a cherry-pick and merge from patch in #1261
Olly Betts