Change Details

CMake syntax has been redesigned to pursue the following goals: * the former generator has used an output of `cmake --help` (and friends) commands, which is quite informal (not a well formed) and error prune * the current syntax was really out of date The new approach is to use a trivial Python script, which uses a prepared YAML data file and Jinja template to render the resulting syntax file. Key features of the new implementation are: * every command highlight only related named parameters -- e.g. `add_library` do not have `TARGET` named option, so it doesn't highlighted in this context * new highlight classes has introduced to distinct property names, aliased (imported) targets, special non-named arguments... * reuse RST syntax for comments * input data now in VCS and trackable there is still some work to do, but I've just want to know is there any interest in this approach...

CMake syntax has been redesigned to pursue the following goals: * the former generator has used an output of `cmake --help` (and friends) commands, which is quite informal (not a well-formed) and error prune * the current syntax was really out of date The new approach is to use a trivial Python script, which uses a prepared YAML data file and Jinja template to render the resulting syntax file. Key features of the new implementation are: * every command highlight only related named parameters -- e.g. `add_library` do not have `TARGET` named option, so it doesn't highlight in this context * new highlight classes have introduced two distinct property names, aliased (imported) targets, special non-named arguments... * reuse RST syntax for comments * input data now in VCS and trackable there is still some work to do, but I've just wanted to know is there any interest in this approach...

CMake syntax has been redesigned to pursue the following goals: * the former generator has used an output of `cmake --help` (and friends) commands, which is quite informal (not a well -formed) and error prune * the current syntax was really out of date The new approach is to use a trivial Python script, which uses a prepared YAML data file and Jinja template to render the resulting syntax file. Key features of the new implementation are: * every command highlight only related named parameters -- e.g. `add_library` do not have `TARGET` named option, so it doesn't highlighted in this context * new highlight classes hashave introduced two distinct property names, aliased (imported) targets, special non-named arguments... * reuse RST syntax for comments * input data now in VCS and trackable there is still some work to do, but I've just wanted to know is there any interest in this approach...