find svn revision by removed text

Question

Is there a way to find an SVN revision by searching for a text string that got removed in the file? I know the exact text to search for and which file to look in, but there are hundreds of revisions.

Answer 1

Building on khmarbaise's script, I came up with this:

#!/bin/bash
file="$1"
REVISIONS=`svn log $file -q --stop-on-copy |grep "^r" | cut -d"r" -f2 | cut -d" " -f1`
for rev in $REVISIONS; do
    prevRev=$(($rev-1))
    difftext=`svn diff --old=$file@$prevRev --new=$file@$rev | tr -s " " | grep -v " -\ \- " | grep -e "$2"`
    if [ -n "$difftext" ]; then
        echo "$rev: $difftext"
    fi
done

pass the file name and search string on the command line:

xyz.sh "filename" "text to search"

svn diff gives me both the rev where it's added and where it's deleted; I'll leave it here in case it's useful to anyone. There's an error message at the last revision that I don't know how to get rid of (I still got a lot of bash to learn :) ) but the rev numbers are correct.

Answer 2

just a little bash script which filters out the changed lines...If you change pom.xml into your file may with supplemental URL you have what you need...(If you are on Unix like system). Put the following into a script file (xyz.sh) and do a filter on the output.

#!/bin/bash
REVISIONS=`svn log pom.xml -q|grep "^r" | cut -d"r" -f2 | cut -d" " -f1`
for rev in $REVISIONS; do
    svn blame -r$rev:$rev pom.xml | tr -s " " | grep -v " -\ \- "
done


xyz.sh | grep "Text you are searching for"

The printout will be something like:

256 ......

The 256 is the revision in which the change had been made.

Answer 3

I like blameall.py.

It's output for a file contains all existing and deleted lines so that I can see the deleted lines in the surrounding context.

Answer 4

Made a simple PHP version of the accepted answer to be run on the command line:

<?php
$file = $argv[1];
$searchfor = $argv[2];

if (!$file) die('Please specify file name to search through as the first argument.');
if (!$searchfor) die('Please specify text to search for as the second argument.');

echo PHP_EOL .'Searching '. $file .' for: '. $searchfor . PHP_EOL . PHP_EOL;

$cmd = 'svn log '. $file .' -q';
$output = array();
exec($cmd, $output);

// Find all revisions
$all_revisions = array();
foreach ($output as $line) {
    if (preg_match("/^r(\\d+) /", $line, $match)) {
        $all_revisions[] = $match[1];
    }
}

echo 'Checking '. count($all_revisions) .' revisions...'. PHP_EOL . PHP_EOL;

// Get diff for each revision
foreach ($all_revisions as $keykey => $rev) {
    $prev_rev = $all_revisions[$keykey+1];

    $cmd = 'svn diff --old='. $file .'@'. $prev_rev .' --new='. $file .'@'. $rev;
    $output = array();
    exec($cmd, $output);

    // Check if string is present
    $str = implode(PHP_EOL, $output);
    if (strpos($str, $searchfor) !== false) {
        echo 'Found in revision '. $prev_rev .' but removed from '. $rev .'.'. PHP_EOL;
        break;
    } else {
        echo 'Not found in revision '. $rev .'.'. PHP_EOL;
    }
}
echo 'Done.'. PHP_EOL;

Run the script the script with the following command:

php scriptname.php "filetosearchthrough.php" "text to search for"

The script could be fine-tuned but does what you need. Keep in mind that it takes a long time to search through many revisions.

Answer 5

If you can't or you don't want to use scripts:

svn log --diff [path_to_file] > log.txt

(the [path_to_file] is optional, if you don't specify it, it will include all the files in the current folder and subfolders)

There you can find all added lines (they begin with +), all removed lines (they begin with -) and some context lines (they begin with a space). Modified lines will appear with the + and repeated with the -

You can process the resulting file with any decent text editor (e.g. Notepad++), grep, scripts or anything you like.

UPDATE:

In case the file was deleted and added again at some point of the history, you can do the following to search before that point:

• First find when the file was deleted:
  • svn log -v [path-to-parent-folder] > logParent.txt
  • search for "D [path-to-file]" in logParent.txt
  • for earch revision found (let's call it [deleted-revision]) subtract 1: [prev-deleted-revision] = [deleted-revision]-1
  • svn log --diff [path-to-file]@[prev-deleted-revision] > log.txt