You definitely want to use fast-import, but you probably want to do something a lot closer to fast-export for monotone (read: use its automate stdio interface and avoid expensive calls).
Here's a simple monotone to git converter I wrote. You'll need the Monotone::AutomateStdio perl module to use it (which I think I got it from monotone's net.venge.monotone.contrib.lib.automate-stdio branch). It is very fast; it can convert the OpenEmbedded repo in something like 5-10 minutes on my machine.
Note that for monotone export to go fast you absolutely /must/ avoid the get_manifest operation. In my converter I use the revision information directly. Getting the renames right with this is a little tricky; IIRC, the ordering that works is:
* Rename all renamed files, innermost files first, to temporary names.
* Delete all deleted files, innermost first.
* Rename all temporary names to permanent names, outermost first.
* Add all new/modified files.
Conveniently, all of the above can be done by directly emitting fast-import commands, so you don't have to keep track of trees directly. (With one exception, which I'll elaborate on in a different email.)
-bcd
#!/usr/bin/perl # Copyright (C) 2007-2008 Brian Downing # This program is licensed under version 2 of the GNU GPL.
use strict; use Monotone::AutomateStdio; use Date::Parse;
my $D = 0;
my $m = Monotone::AutomateStdio->new($ARGV[0]);
my $revlist = []; $m->graph($revlist); my $sorted = [];
for my $rev (@$revlist) {
push(@$sorted, $rev->{revision_id});
}my $leaves = []; $m->leaves($leaves);
$m->toposort($sorted, @$sorted);
my $marks = {}; my $mark = 1; my $c = 0;
sub quote_file {
$_ = shift;
return $_;
s/\\/\\\\/g;
s/\n/\\n/g;
s/"/\\"/g;
return qq("$_");
}sub lprint {
my $fh = shift;
print @_ if $D;
print $fh @_;
}sub lprintf {
my $fh = shift;
printf @_ if $D;
printf $fh @_;
}my $tmptag = "624d893e-ae1a-42d8-90a9-926a6ceffae8";
open my $fi, '|-', 'git-fast-import --export-marks=file';
for my $rev (@$sorted) {
my ($time, $author, $msg) = ("0", "__UNKNOWN__", "__UNKNOWN__");
my @certs;
my @branches;
$m->certs(\@certs, $rev);
for my $cert (@certs) {
my ($n, $v) = ($cert->{name}, $cert->{value});
$author = $v if ($n eq 'author');
$time = $v if ($n eq 'date');
$msg = $v if ($n eq 'changelog');
push(@branches, $v) if ($n eq 'branch');
}
my $email = $author;
$msg .= "\nmtn-revision: $rev\n";
for my $b (sort @branches) {
$msg .= "mtn-branch: $b\n";
}
$time = str2time($time, 'UTC');
my $mfest = [];
$m->get_revision($mfest, $rev);
my $orcount = 0;
my $add_files = {};
my $add_dirs = {};
my $delete_files = {};
my $from_tmpnames = {};
my $to_tmpnames = {};
my $curtmp = 0;
my @parents;
for my $e (@$mfest) {
if ($e->{type} eq 'old_revision') {
push(@parents, $e->{revision_id});
++$orcount;
}
next if $orcount > 1;
if ($e->{type} eq 'add_file' || $e->{type} eq 'patch') {
my $id = $e->{file_id} || $e->{to_file_id};
$add_files->{$e->{name}} = $id;
unless ($marks->{$id}) {
my $data;
$m->get_file(\$data, $id);
print "new file $id\n" if $D;
print $fi "blob\n";
my $len = length($data);
print $fi "mark :$mark\n";
$marks->{$id} = $mark++;
print $fi "data $len\n$data\n";
}
} elsif ($e->{type} eq 'add_dir') {
$add_dirs->{$e->{name}} = 1;
} elsif ($e->{type} eq 'delete') {
$delete_files->{$e->{name}} = 1;
} elsif ($e->{type} eq 'rename') {
$curtmp++;
$from_tmpnames->{$e->{from_name}} = "__tmp_${tmptag}_$curtmp";
$to_tmpnames->{$e->{to_name}} = "__tmp_${tmptag}_$curtmp";
}
}
printf("rev $rev (%d/%d, %.2f%)\n",
++$c, scalar(@$sorted), 100*$c/scalar(@$sorted));
print $fi "reset refs/import\n" unless @parents;
lprint $fi, "commit refs/import\n";
print $fi "mark :$mark\n";
$marks->{$rev} = $mark++;
if ($author =~ m(\s*(.*?\S)\s*<(.*)>\s*)) {
$author = $1;
$email = $2;
}
$author =~ s/[<>]/_/g;
$email =~ s/[<>]/_/g;
$author =~ s/@.*//;
print $fi "committer $author <$email> $time +0000\n";
my $len = length($msg);
print $fi "data $len\n$msg\n";
my $from = "from";
for my $p (@parents) {
lprint $fi, "$from :$marks->{$p}\n";
$from = "merge";
}
for my $f (sort { length($b) <=> length ($a) } keys %$from_tmpnames) {
lprintf($fi, "R %s %s\n",
quote_file($f), quote_file($from_tmpnames->{$f}));
}
for my $f (sort { length($b) <=> length ($a) } keys %$delete_files) {
lprintf($fi, "D %s\n", quote_file($f));
}
for my $f (sort { length($a) <=> length ($b) } keys %$to_tmpnames) {
lprintf($fi, "R %s %s\n",
quote_file($to_tmpnames->{$f}), quote_file($f));
}
for my $f (keys %$add_files) {
lprintf($fi, "M 0644 :%s %s\n",
$marks->{$add_files->{$f}}, quote_file($f));
}
for my $f (keys %$add_dirs) {
$f .= "/" if $f;
lprintf($fi, "M 0644 inline %s\n", quote_file("$f.gitignore"));
lprint($fi, "data 0\n\n");
}
print $fi "\n";
}my $branches = {};
for my $rev (@$leaves) {
my $branch;
my @certs;
$m->certs(\@certs, $rev);
for my $cert (@certs) {
my ($n, $v) = ($cert->{name}, $cert->{value});
$branch = $v if ($n eq 'branch');
}
my $r = $branches->{$branch};
$branches->{$branch}--;
if ($marks->{$rev}) {
print $fi "reset refs/heads/$branch$r\n";
print $fi "from :$marks->{$rev}\n\n";
}
}close $fi;
Felipe Contreras
