0

I am using tree and diff to compare a local directory with a USB backup version of the same directory, and to identify noteworthy / discardable differences.

$ cd home/dir/ $ tree --charset=ascii -af . > ~/tree__local $ cd {usb}/home/dir/ $ tree --charset=ascii -af . > ~/tree__usb $ diff ~/tree__local ~/tree__usb > ~/diff-analysis.txt

But I've noticed that tree has the habit of leaving out parent branches for last items … Ugh, this is hard to describe with words:

$ tree --charset=ascii -af . . |-- ./bar | |-- ./bar/babushka | |-- ./bar/galeda | `-- ./bar/helga `-- ./foo |-- ./foo/alpha |-- ./foo/beta `-- ./foo/gamma 8 directories, 0 files # add a new final item. $ touch zebra # same command. $ tree --charset=ascii -af . . |-- ./bar | |-- ./bar/babushka | |-- ./bar/galeda | `-- ./bar/helga |-- ./foo | |-- ./foo/alpha | |-- ./foo/beta | `-- ./foo/gamma `-- ./zebra 8 directories, 1 file

Notice how the files in foo/ have gained an additional pipe branch. The diff command marks these as differences even though the files themselves are actually not different at all.

This is very suboptimal for my purposes. Do you guys happen to know how to circumvent or prevent that? Can you force homogenous pipe branches all the way down? The tree manual makes no mention of formatting the "last" or "final" item.

Improve this question
1
  • You might want to compare find . in each, or use rsync as noted in one of the answers. You're trying to hammer in a screw. Commented Sep 26, 2021 at 23:36

3 Answers 3

Reset to default
4

The output of tree is meant for display purposes and is less useful for parsing.

It would be better to compare the two directory hierarchies directly. You may do this with, e.g., rsync like so:

rsync --dry-run --archive --delete --itemize-changes home/dir/ "$usb_path"/home/dir/

(Including --dry-run, or the short variant of this option, -n, is very important.)

This command would tell you precisely the difference between the local directory and the directory on the USB device. Nothing would be changed on disk as we are running with the --dry-run option. The --itemize-changes option makes rsync output a terse status string for each file or directory that it would have copied or deleted, encoding the reason for the synchronization.

Add --omit-dir-times to the end of the list of options to avoid listing directories whose timestamps are different. Add --ignore-existing to make it ignore files that exist in both locations, even though their sizes or meta-data differs. With these two extra option added to the command above, you would expect to see lines of output prefixed with

  • *deleting, for things deleted from the local directory but not from the USB device,
  • cd+++++++++, for directories added to the local directory, or
  • >f+++++++++, for files added to the local directory.
Improve this answer
3
  • This sounds like a smart approach … except, I get a lot of false positives: cd+++++++++ entries show up for directories with identical file listings inside them in LOCAL and USB; and there's way too many of them. — Can the directory/file permissions be a factor here? I've got 644 for files, 755 for directories on LOCAL vs. 777 for all on USB (because I still didn't get around to try this user mapping file solution for this other issue that I have), askubuntu.com/a/74851/846446 Commented Oct 1, 2021 at 14:12
  • @WoodrowShigeru You could try using --no-perms after --archive. Commented Oct 1, 2021 at 15:49
  • Unfortunately, I see no difference. I've also tried -rlgoD instead of --archive w/o any effect. When comparing the dirs with ls, I sometimes see another time on the dir or a subdir, but not any of the files inside – or I see another link count. detailed.wordpress.com/2017/10/28/… — I've had to read up on it – stackoverflow.com/questions/34407685/link-count-and-ls-command –, not sure how a link count could affect rsync like that. Commented Oct 1, 2021 at 16:35
1

I guess you can get rid of the complete indentation and branch formatting in order to work better with diff:

-i
Makes tree not print the indentation lines, useful when used in conjunction with the -f option. Also removes as much whitespace as possible when used with the -J or -x options.

$ tree --charset=ascii -afi . . ./bar ./bar/babushka ./bar/galeda ./bar/helga ./foo ./foo/alpha ./foo/beta ./foo/gamma ./zebra 8 directories, 1 file

Too bad: the indentation helps me visually understand these things faster. But oh well …

Improve this answer
0

In each line turn everything before the first . into spaces:

tree --charset=ascii -af . | awk 'BEGIN{FS=OFS="."} {gsub(/./, " ", $1)} 1'
Improve this answer

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.