Question

I found this nice little script that I’ve made into a bash script.

#!/bin/bash
svn log -v -r{$1}:HEAD | awk '/^r[0-9]+ / {user=$3} /./{if (user=="'$2'") {print}}' | grep -E "^   M|^   G|^   A|^   D|^   C|^   U" | awk '{print $2}' | sort | uniq

$1 = 2012-10-01 //for example
$2 = George //for example

This script seems to work but its only printing out commits from trunk.
When I svn log from the terminal I see all commits regardless of branch.

So far I’ve narrowed the problem down to awk.
We use spaces in our branch folder names. Don’t sure if that could be a problem.

Is /./ in awk area the issue?

So far, I’m getting the url from the svn info and putting it in a variable.

theURL=$(svn info | grep -E "^(URL: )http(s)?:")
theURL=${theURL:5}

Perhaps I can tell awk to only print out a line where the line matches part of my url scheme?

Thanks,

Answer 1

svn log without path parameter uses default “.” path (which must be and can be only Working Copy, in your case – linked to trunk of repo)

svn log URL at the same time can work with any part of tree for remote repository, unrelated to current setting of (some) WC.

Just add and use 3-rd parameter