8b08a28c89
* fixes issue #13981: unsafe_writes block appeared too late in the atomic_move workflow. This led to errno.EBUSY to not be managed in the context of issue #!#981 * Reduce changes to fix #13981 * Abstract the unsafe_writes fallback into a helper method. Explicitly try/except os.rename part of the code and call this helper method. If the code fails in shutil.copy2 or shutil.move this should not be related to issue #13981 since they write to b_tmp_dest_name. (as suggested by @abadger) * Check if unsafe_writes in the caller, not in _unsafe_writes. That way the function call reads as "Do an unsafe write" and not as "I think we should do an unsafe_write. |
||
|---|---|---|
| .. | ||
| ansible | ||