Negative transcript entries can be shadowed by updates

[hfath]

When an update changes owner or access information of negative files and directories, they will show up in the positive transcript generated.

This should not happen, is a mis-feature or more likely a bug, and should be replaced by a warning at best, maybe just result in a commented-out transcript line. Creating a new transcript should heed negative transcripts. Until then, extra care is needed to spot these overrides in the new transcript, and remove them.

[sth0]

But isn't a negative transcript supposed to enforce the owner/group and permissions for the file but not the contents or any sub files if it is a directory. What you want is an "exclude" with a wildcard which just drops the file and any sub files. (technically two excludes for a directory)

x dir_to_exclude
x dir_to_exclude/**

[hfath]

But isn't a negative transcript supposed to enforce the owner/group and permissions for the file but not the contents or any sub files if it is a directory.

Yes, and the expectation (mine, at least) is that this information does not get superseded by a freshly created (positive) transcript. Because 99.9% of the time that is not what I want.

[sth0]

Yes, those do end up in the lcreate transcripts, because those are changes noted by radmind. This is why I started using exclude transcripts for each OS version.