Conan 1.29: Updates to multiple generators, new tools.rename function, experimental toolchain for gnu make, lockfile bug fixes
Conan 1.29 has been released. This release doesn’t contain any major new features, but instead brings more completeness and quality to existing features.
Updates to Generators
Some of the newer cpp_info
properties have now been added to both the QMake
and Qbs
generators, including cpp_info.system_libs
,
cpp_info.framework_paths
and cpp_info.frameworks
. Additionally, the json
generator learned how to generate the user_info
data from the build context.
The cmake_find_package
and cmake_find_package_multi
both had some bugs in
the previous release surrounding the new filenames
attribute of cpp_info
.
Also, there was a bug in the cmake_multi
generator in which it was producing
cmake
files which were failing under valid conditions. These have all been
fixed.
New tools.rename Function
On Windows, many users have reported an intermittant problem with python’s
os.rename
function when used in some conanfile.py
contexts. Sometimes the
function will fail simply with a permission error. Numerous workarounds have
been offerred, but in this release we’ve merged a new tools function called
tools.rename
which uses a more robust mechanism to avoid these errors.
Experimental Toolchain for Gnu Make
Continuing our work on Conan’s new toolchain strategy, this release contains a
toolchain class for the Gnu Make
build system. Like the existing CMake
toolchain which had been added previously, this new toolchain class
generates .mak
files for the Gnu Make
build system which contain all the
relevant build-related variables from Conan. To clarify, this does NOT include
variables related to dependencies: this is still the domain of generators.
Lockfile Bug Fixes
We have also continued to evolve and improve the experience around Lockfiles. In
this release, we fixed one bug where some package’s were erroneously marked as
modified, and another bug where Package_ID_Unknown
were not being updated
correctly. We’ve also added a more helpful message when users try to compute
the build_order
of a graph from a “base” lockfile (which isn’t possible). Now
the output should be much more helpful.
Besides the items listed above, there were some minor bug fixes you may wish to read about. If so, please refer to the changelog for the complete list.
We hope you enjoy this release, and look forward to your feedback.