Which commit has this blob?

Both of the following scripts take the blob’s SHA1 as the first argument, and after it, optionally, any arguments that git log will understand. E.g. --all to search in all branches instead of just the current one, or -g to search in the reflog, or whatever else you fancy.

Here it is as a shell script – short and sweet, but slow:

#!/bin/sh
obj_name="$1"
shift
git log "$@" --pretty=tformat:'%T %h %s' \
| while read tree commit subject ; do
    if git ls-tree -r $tree | grep -q "$obj_name" ; then
        echo $commit "$subject"
    fi
done

And an optimised version in Perl, still quite short but much faster:

#!/usr/bin/perl
use 5.008;
use strict;
use Memoize;

my $obj_name;

sub check_tree {
    my ( $tree ) = @_;
    my @subtree;

    {
        open my $ls_tree, '-|', git => 'ls-tree' => $tree
            or die "Couldn't open pipe to git-ls-tree: $!\n";

        while ( <$ls_tree> ) {
            /\A[0-7]{6} (\S+) (\S+)/
                or die "unexpected git-ls-tree output";
            return 1 if $2 eq $obj_name;
            push @subtree, $2 if $1 eq 'tree';
        }
    }

    check_tree( $_ ) && return 1 for @subtree;

    return;
}

memoize 'check_tree';

die "usage: git-find-blob <blob> [<git-log arguments ...>]\n"
    if not @ARGV;

my $obj_short = shift @ARGV;
$obj_name = do {
    local $ENV{'OBJ_NAME'} = $obj_short;
     `git rev-parse --verify \$OBJ_NAME`;
} or die "Couldn't parse $obj_short: $!\n";
chomp $obj_name;

open my $log, '-|', git => log => @ARGV, '--pretty=format:%T %h %s'
    or die "Couldn't open pipe to git-log: $!\n";

while ( <$log> ) {
    chomp;
    my ( $tree, $commit, $subject ) = split " ", $_, 3;
    print "$commit $subject\n" if check_tree( $tree );
}

Leave a Comment