Use package.xml version in cmake (using python)

[scpeters]

Alternative to #232 for addressing the duplicate package versions specified in cmake and package.xml. Whereas #232 adds a github workflow to verify that the versions are identical, this copies a python script and cmake helper from gazebosim/gz-cmake#456 to get the version from a package.xml file and use it in cmake. This makes python a hard requirement for configuring the urdfdom package.

[clalancette]

I'm not opposed to this, but something occurred to me. The data in the package.xml is highly structured XML. Could we not do something as simple as doing a regex? Something like the following:

file(READ package.xml PACKAGE_XML_DATA)
string(REGEX MATCH "<version>([0-9]+)\.([0-9]+)\.([0-9]+)</version>" _ ${PACKAGE_XML_DATA})
set(URDF_MAJOR_VERSION ${CMAKE_MATCH_1})
set(URDF_MINOR_VERSION ${CMAKE_MATCH_2})
set(URDF_PATCH_VERSION ${CMAKE_MATCH_3})

[traversaro]

I'm not opposed to this, but something occurred to me. The data in the package.xml is highly structured XML. Could we not do something as simple as doing a regex? Something like the following:

file(READ package.xml PACKAGE_XML_DATA)
string(REGEX MATCH "<version>([0-9]+)\.([0-9]+)\.([0-9]+)</version>" _ ${PACKAGE_XML_DATA})
set(URDF_MAJOR_VERSION ${CMAKE_MATCH_1})
set(URDF_MINOR_VERSION ${CMAKE_MATCH_2})
set(URDF_PATCH_VERSION ${CMAKE_MATCH_3})

dartsim does something similar: https://github.com/dartsim/dart/blob/59f39923015604b2984b82fc8c964dad9cd598f4/CMakeLists.txt#L72 .

[scpeters]

I'm not opposed to this, but something occurred to me. The data in the package.xml is highly structured XML. Could we not do something as simple as doing a regex? Something like the following:

file(READ package.xml PACKAGE_XML_DATA)
string(REGEX MATCH "<version>([0-9]+)\.([0-9]+)\.([0-9]+)</version>" _ ${PACKAGE_XML_DATA})
set(URDF_MAJOR_VERSION ${CMAKE_MATCH_1})
set(URDF_MINOR_VERSION ${CMAKE_MATCH_2})
set(URDF_PATCH_VERSION ${CMAKE_MATCH_3})

sure, this is reasonable enough and is more concise. I'll open a separate PR with this approach

[scpeters]

I'm not opposed to this, but something occurred to me. The data in the package.xml is highly structured XML. Could we not do something as simple as doing a regex? Something like the following:

file(READ package.xml PACKAGE_XML_DATA)
string(REGEX MATCH "<version>([0-9]+)\.([0-9]+)\.([0-9]+)</version>" _ ${PACKAGE_XML_DATA})
set(URDF_MAJOR_VERSION ${CMAKE_MATCH_1})
set(URDF_MINOR_VERSION ${CMAKE_MATCH_2})
set(URDF_PATCH_VERSION ${CMAKE_MATCH_3})

sure, this is reasonable enough and is more concise. I'll open a separate PR with this approach

see #236

[scpeters]

the approach from #236 was used instead