Separate the phases of creating the list of outputs (and output directories) from the actual creation of these outputs. This lays the groundwork for just printing the inputs and outputs in a subsequent patch, enabling CMake to track dependencies properly. In the now-separate output generation phase, use shutil.copy2() to preserve metadata like access and modification times when possible, rather than re-reading every file and seeing if the contents are different. This has the same semantics but is clearer and potentially more efficient, as the initial copies have more accurate relative timestamps. Signed-off-by: Marti Bolivar <marti@foundries.io> |
||
|---|---|---|
| .. | ||
| extract_content.py | ||
| fix_tex.py | ||
| genrest.py | ||