Question

Using svn diff with the --summarize flag returns something like the below. How would we pass this to sed or grep to do the following:

1. Remove any lines that start with “D” (deleted files)
2. Remove the prefix of the “M”, “A” or “MM” (or any other cases) along with the tab afterward.
3. Remove the URL path only leaving the filenames/folder.
4. Store in a file

Example:

D   https://localhost/example/test1.php
D   https://localhost/example/test2.php
M   https://localhost/example/test3.php
M   https://localhost/example/test4.php
A   https://localhost/example/test5.php
M   https://localhost/example/test6.php
A   https://localhost/example/test7.php
M   https://localhost/example/test8.php
M   https://localhost/example/test9.php
M   https://localhost/example/test10.php
A   https://localhost/example/test11.php
M   https://localhost/example/test12.php
M   https://localhost/example/test13.php
MM  https://localhost/example/test.php
M   https://localhost/test0.php

Would then become:

/example/test3.php
/example/test4.php
/example/test5.php
/example/test6.php
/example/test7.php
/example/test8.php
/example/test9.php
/example/test10.php
/example/test11.php
/example/test12.php
/example/test13.php
/example/test.php
/test0.php

Answer 1

Like this with sed:

$ svn diff --summarize | sed -e '/^D/d' -e 's/.*host//'
/example/test3.php
/example/test4.php
/example/test5.php
/example/test6.php
/example/test7.php
/example/test8.php
/example/test9.php
/example/test10.php
/example/test11.php
/example/test12.php
/example/test13.php
/example/test.php
/test0.php

# Redirect output to file
$ svn diff --summarize | sed -e '/^D/d' -e 's/.*host//' > file

You need to pipe | the output of svn to sed. The first part '/^D/d' deletes all the lines starting with D and the second s/.*host// substitutes everything up to host with nothing, to store in to a file use redirect > file.

Similar logic with grep:

$ svn diff --summarize | grep '^[^D]' file | grep -Po '(?<=host).*' > file

The first grep filters out the lines that start with D and the second uses positive lookahead with -Po to display only the part of the line after host.