prune: Skip calculating potential freed space in the dry run #5813
Conversation
| n_objects_pruned, formatted_freed_size); | ||
| if (!opt_prune_dry_run) | ||
| { | ||
| formatted_freed_size = g_format_size_full (objsize_total, 0); |
There was a problem hiding this comment.
The declaration of the formatted_freed_size variable should be moved to this if block now.
But, what's the reasoning for not printing this on dry runs? Does it always return zero or a bogus value on dry runs?
The commit message could have some of that reasoning.
There was a problem hiding this comment.
The data will be null/zero because it will no longer traverse all objects.
| if (!dry_run) | ||
| { | ||
| g_info ("Pruning unreachable objects"); | ||
| g_timer_start (timer); | ||
|
|
||
| if (!prune_unreachable_loose_objects (repo, &data, cancellable, error)) | ||
| return FALSE; | ||
| if (!prune_unreachable_loose_objects (repo, &data, cancellable, error)) | ||
| return FALSE; | ||
|
|
||
| g_timer_stop (timer); | ||
| g_info ("Elapsed time: %.1f sec", g_timer_elapsed (timer, NULL)); | ||
| g_timer_stop (timer); | ||
| g_info ("Elapsed time: %.1f sec", g_timer_elapsed (timer, NULL)); | ||
| } |
There was a problem hiding this comment.
Well this is not just "calculating", right? This is actually pruning stuff on a dry run, seems like an actual problem here.
There was a problem hiding this comment.
It is not actually pruning. Somewhere down the stackcode there's data.dont_prune defined that causes this function to stat/statx each object to calculate potential space-saving.
No description provided.