Merge branch 'master' into wildrepos

lots of conflicts, esp in gl-auth-command, due to refactoring the
"special commands" stuff on master

Conflicts:
	doc/3-faq-tips-etc.mkd
	src/gitolite.pm
	src/gl-auth-command
	src/gl-compile-conf

README.mkd

@@ -99,6 +99,11 @@ or in Unix, perl, shell, etc.)...  well I can't afford 1000 USD rewards like
 djb, so you'll have to settle for 1000 INR (Indian Rupees) as a "token" prize
 :-)
 
+Update 2010-01-31: this security promise does not apply if you enable any of
+the external command helpers (like rsync).  It's probably quite secure, but I
+just haven't thought about it enough to be able to make such promises, like I
+can for the rest of "master".
+
 ----
 
 ### contact and license

conf/example.conf

@@ -147,10 +147,18 @@ repo    git
 
 # DENY/EXCLUDE RULES
 
-# ***IMPORTANT NOTE: if you use deny rules, the order of the rules also makes
+# ***IMPORTANT NOTES ABOUT "DENY" RULES***:
-# a difference, where earlier it did not.  Please review your ruleset
+
-# carefully or test it.  In particular, do not use `@all` in a deny rule -- it
+# - deny rules do NOT affect read access.  They only apply to `W` and `+`.
-# won't work as you might expect***.
+#
+# - when using deny rules, the order of your rules starts to matter, where
+#   earlier it did not.  The first matching rule applies, where "matching" is
+#   defined as either permitting the operation you're attempting (`W` or `+`),
+#   which results in success, or a "deny" (`-`), which results in failure.
+#   (As before, a fallthrough also results in failure).
+#
+# - do not use `@all` when your config has any deny rules; it won't work as
+#   you probably expect it to!
 
 # in the example above, you cannot easily say "anyone can write any tag,
 # except version tags can only be written by junio".  The following might look
@@ -166,10 +174,6 @@ repo    git
         -       refs/tags/v[0-9]        =   linus pasky @others
         RW      refs/tags/              =   junio linus pasky @others
 
-# Briefly, the rule is: the first matching refex that has the operation you're
-# looking for (`W` or `+`), or a minus (`-`), results in success, or failure,
-# respectively.  A fallthrough also results in failure
-
 # FILE/DIR NAME BASED RESTRICTIONS
 # --------------------------------
 
@@ -264,3 +268,21 @@ repo gitolite
 #   security reasons.
 # - you can also use an absolute path if you like, although in the interests
 #   of cloning the admin-repo sanely you should avoid doing this!
+
+# EXTERNAL COMMAND HELPERS -- RSYNC
+# ---------------------------------
+
+# If $RSYNC_BASE is non-empty, the following config entries come into play
+# (otherwise they are ignored):
+
+# a "fake" git repository to collect rsync rules.  Gitolite does not
+# auto-create any repo whose name starts with EXTCMD/
+repo    EXTCMD/rsync
+# grant permissions to files/dirs within the $RSYNC_BASE tree.  A leading
+# NAME/ is required as a prefix; the actual path starts after that.  Matching
+# follows the same rules as elsewhere in gitolite.
+        RW  NAME/                       =   sitaram
+        RW  NAME/foo/                   =   user1
+        R   NAME/bar/                   =   user2
+# just to remind you that these are perl regexes, not shell globs
+        RW  NAME/baz/.*/*.c             =   user3

conf/example.gitolite.rc

@@ -106,6 +106,26 @@ $GIT_PATH="";
 # syntax: space separated list of gitolite usernames in *one* string variable.
 # $SHELL_USERS = "alice bob";
 
+# --------------------------------------
+
+# if you want to enable the "htpasswd" command, give this the absolute path to
+# whatever file apache (etc) expect to find the passwords in.
+
+$HTPASSWD_FILE = "";
+
+# Look in doc/3 ("easier to link gitweb authorisation with gitolite" section)
+# for more details on using this feature.
+
+# --------------------------------------
+# EXTERNAL COMMAND HELPER -- RSYNC
+#
+# base path of all the files that are accessible via rsync.  Must be an
+# absolute path.  Leave it undefined or set to the empty string to disable the
+# rsync helper.
+$RSYNC_BASE = "";
+# $RSYNC_BASE = "/home/git/up-down";
+# $RSYNC_BASE = "/tmp/up-down";
+
 # --------------------------------------
 # per perl rules, this should be the last line in such a file:
 1;

doc/3-faq-tips-etc.mkd

@@ -27,6 +27,7 @@ In this document:
       * "personal" branches
       * custom hooks and custom git config
       * repos named with wildcards
+      * access control for external commands
   * design choices
       * keeping the parser and the access control separate
 
@@ -331,9 +332,25 @@ This requires that:
   * the HTTP auth should use the same username (like "sitaram") as used in the
     gitolite config (for the corresponding user)
 
-Once that is done, it's easy.  Gitweb allows you to specify a subroutine to
+Normally a superuser sets up passwords for users using the "htpasswd" command,
-decide on access.  We use that feature and tie it to gitolite.  Sample code
+but this is an administrative chore.
-(untested, munged from something I saw [here][leho]) is given below.
+
+Robin Smidsrød had the *great* idea that, since each user already has pubkey
+access to `git@server`, this gives us a very neat way of using gitolite to let
+the users *manage their own HTTP passwords*.  Here's how:
+
+  * setup apache so that the htaccess file it looks for is owned by the "git"
+    user
+  * in the `~/.gitolite.rc` file, look for the variable `$HTPASSWD_FILE` and
+    point it to this file
+  * tell your users to type in `ssh git@server htpasswd` to set or change
+    their HTTP passwords
+
+Here's the rest of how it hangs together.
+
+Gitweb allows you to specify a subroutine to decide on access.  We use that
+feature and tie it to gitolite.  Sample code (untested by me, but others do
+use it, munged from something I saw [here][leho]) is given below.
 
 Note the **utter simplicity** of the actual check (just 1 line!).  This is an
 unexpected piece of luck coming from the decision to keep the config parse
@@ -349,7 +366,7 @@ already done and we just use it!
     $projectroot = '/home/git/repositories/';
     my $gl_conf_compiled = '/home/git/.gitolite/conf/gitolite.conf-compiled.pm';
 
-    # I assume this gives us the HTTP auth username
+    # I am told this gives us the HTTP auth username
     my $username = $cgi->remote_user;
 
     # ----------
@@ -359,10 +376,11 @@ already done and we just use it!
     die "parse $gl_conf_compiled failed: " . ($! or $@) unless do $gl_conf_compiled;
 
     # this is gitweb's mechanism; it calls whatever sub is pointed at by this
-    # variable to decide access yes/no
+    # variable to decide access yes/no.  Gitweb calls it with one argument
+    # containing the full path of the repo being accessed
     $export_auth_hook = sub {
         my $reponame = shift;
-        # gitweb passes us the full repo path; so we strip the beginning...
+        # take the full path provided, strip the beginning...
         $reponame =~ s/\Q$projectroot\E\/?//;
         # ...and the end, to get the repo name as it is specified in gitolite conf
         $reponame =~ s/\.git$//;
@@ -499,12 +517,9 @@ that code path to better use :-)
 
 #### "exclude" (or "deny") rules
 
-***IMPORTANT CAVEAT: if you use deny rules, the order of the rules also makes
+Here is an illustrative explanation of "deny" rules.  However, please be sure
-a difference, where earlier it did not.  Please review your ruleset carefully
+to read the "DENY/EXCLUDE RULES" section in `conf/example.conf` for important
-or test it.  In particular, do not use `@all` in a deny rule -- it won't work
+notes/caveats before using "deny" rules.
-as you might expect***.  Also, deny rules are only processed in the second
-level checks (see "two levels of access rights checking" above), which means
-they only apply to write operations.
 
 Take a look at the following snippet, which *seems* to say that "bruce" can
 write versioned tags (anything containing `refs/tags/v[0-9]`), but the other
@@ -614,6 +629,20 @@ per-repo "gitconfig" settings.  Please see `doc/2-admin.mkd` and
 **This feature only exists in the "wildrepos" branch!**  Please see
 `doc/4-wildcard-repositories.mkd` for all the details.
 
+#### access control for external commands
+
+Gitolite now has a mechanism for allowing access control for arbitrary
+external commands, as long as they are invoked via ssh and present a
+server-side command that contains enough information to make an access control
+decision.  The first (and only, so far) such command implemented is rsync.
+
+Note that this is incompatible with giving people shell access as described in
+`doc/6-ssh-troubleshooting.mkd` -- people who have shell access are not
+subject to this mechanism (it wouldn't make sense to try and control someone
+who has shell access anyway).
+
+Please see the config files (both of them) for examples and usage.
+
 ### design choices
 
 #### keeping the parser and the access control separate

src/gitolite.pm

@@ -292,4 +292,118 @@ sub expand_wild
     }
 }
 
+# ----------------------------------------------------------------------------
+#       S P E C I A L   C O M M A N D S
+# ----------------------------------------------------------------------------
+
+sub special_cmd
+{
+    my ($GL_ADMINDIR, $GL_CONF_COMPILED, $RSYNC_BASE, $HTPASSWD_FILE) = @_;
+
+    my $cmd = $ENV{SSH_ORIGINAL_COMMAND};
+    my $user = $ENV{GL_USER};
+
+    # check each special command we know about and call it if enabled
+    if ($cmd eq 'info') {
+        &report_basic($GL_ADMINDIR, $GL_CONF_COMPILED, $user);
+        print "you also have shell access\n\r" if $shell_allowed;
+    } elsif ($HTPASSWD_FILE and $cmd eq 'htpasswd') {
+        &ext_cmd_htpasswd($HTPASSWD_FILE);
+    } elsif ($RSYNC_BASE and $cmd =~ /^rsync /) {
+        &ext_cmd_rsync($GL_CONF_COMPILED, $RSYNC_BASE, $cmd);
+    } else {
+        # if the user is allowed a shell, just run the command
+        exec $ENV{SHELL}, "-c", $cmd if $shell_allowed;
+
+        die "bad command: $cmd\n";
+    }
+}
+
+# ----------------------------------------------------------------------------
+#       generic check access routine
+# ----------------------------------------------------------------------------
+
+sub check_access
+{
+    my ($GL_CONF_COMPILED, $repo, $path, $perm) = @_;
+    my $ref = "NAME/$path";
+
+    &parse_acl($GL_CONF_COMPILED);
+
+    # until I do some major refactoring (which will bloat the update hook a
+    # bit, sadly), this code duplicates stuff in the current update hook.
+
+    my @allowed_refs;
+    # we want specific perms to override @all, so they come first
+    push @allowed_refs, @ { $repos{$repo}{$ENV{GL_USER}} || [] };
+    push @allowed_refs, @ { $repos{$repo}{'@all'} || [] };
+
+    for my $ar (@allowed_refs) {
+        my $refex = (keys %$ar)[0];
+        next unless $ref =~ /^$refex/;
+        die "$perm $ref $ENV{GL_USER} DENIED by $refex\n" if $ar->{$refex} eq '-';
+        return if ($ar->{$refex} =~ /\Q$perm/);
+    }
+    die "$perm $ref $ENV{GL_REPO} $ENV{GL_USER} DENIED by fallthru\n";
+}
+
+# ----------------------------------------------------------------------------
+#       external command helper: rsync
+# ----------------------------------------------------------------------------
+
+sub ext_cmd_rsync
+{
+    my ($GL_CONF_COMPILED, $RSYNC_BASE, $cmd) = @_;
+
+    # test the command patterns; reject if they don't fit.  Rsync sends
+    # commands that looks like one of these to the server (the first one is
+    # for a read, the second for a write)
+    #   rsync --server --sender -some.flags . some/path
+    #   rsync --server -some.flags . some/path
+
+    die "bad rsync command: $cmd"
+        unless $cmd =~ /^rsync --server( --sender)? -[\w.]+(?: --(?:delete|partial))* \. (\S+)$/;
+    my $perm = "W";
+    $perm = "R" if $1;
+    my $path = $2;
+    die "I dont like absolute paths in $cmd\n" if $path =~ /^\//;
+    die "I dont like '..' paths in $cmd\n" if $path =~ /\.\./;
+
+    # ok now check if we're permitted to execute a $perm action on $path
+    # (taken as a refex) using rsync.
+
+    &check_access($GL_CONF_COMPILED, 'EXTCMD/rsync', $path, $perm);
+        # that should "die" if there's a problem
+
+    wrap_chdir($RSYNC_BASE);
+    &log_it("$ENV{GL_TS}\t$ENV{SSH_ORIGINAL_COMMAND}\t$ENV{USER}\n");
+    exec $ENV{SHELL}, "-c", $ENV{SSH_ORIGINAL_COMMAND};
+}
+
+# ----------------------------------------------------------------------------
+#       external command helper: htpasswd
+# ----------------------------------------------------------------------------
+
+sub ext_cmd_htpasswd
+{
+    my $HTPASSWD_FILE = shift;
+
+    die "$HTPASSWD_FILE doesn't exist or is not writable\n" unless -w $HTPASSWD_FILE;
+    $|++;
+    print <<EOFhtp;
+Please type in your new htpasswd at the prompt.  You only have to type it once.
+
+NOTE THAT THE PASSWORD WILL BE ECHOED, so please make sure no one is
+shoulder-surfing, and make sure you clear your screen as well as scrollback
+history after you're done (or close your terminal instance).
+
+EOFhtp
+    print "new htpasswd:";
+
+    my $password = <>;
+    $password =~ s/[\n\r]*$//;
+    my $rc = system("htpasswd", "-b", $HTPASSWD_FILE, $ENV{GL_USER}, $password);
+    die "htpasswd command seems to have failed with $rc return code...\n" if $rc;
+}
+
 1;

src/gl-auth-command

@@ -24,7 +24,7 @@ use warnings;
 # ----------------------------------------------------------------------------
 
 # these are set by the "rc" file
-our ($GL_LOGT, $GL_CONF_COMPILED, $REPO_BASE, $GIT_PATH, $REPO_UMASK, $GL_ADMINDIR);
+our ($GL_LOGT, $GL_CONF_COMPILED, $REPO_BASE, $GIT_PATH, $REPO_UMASK, $GL_ADMINDIR, $RSYNC_BASE, $HTPASSWD_FILE);
 # and these are set by gitolite.pm
 our ($R_COMMANDS, $W_COMMANDS, $REPONAME_PATT, $REPOPATT_PATT);
 our %repos;
@@ -48,6 +48,8 @@ $ENV{PATH} .= ":$GIT_PATH" if $GIT_PATH;
 # set the umask before creating any files
 umask($REPO_UMASK);
 
+my $repo_base_abs = ( $REPO_BASE =~ m(^/) ? $REPO_BASE : "$ENV{HOME}/$REPO_BASE" );
+
 # ----------------------------------------------------------------------------
 #       start...
 # ----------------------------------------------------------------------------
@@ -63,25 +65,40 @@ if ($ARGV[0] eq '-s') {
 # first, fix the biggest gripe I have with gitosis, a 1-line change
 my $user=$ENV{GL_USER}=shift;       # there; now that's available everywhere!
 
+# ----------------------------------------------------------------------------
+#       logging, timestamp env vars
+# ----------------------------------------------------------------------------
+
+# timestamp
+my ($s, $min, $h, $d, $m, $y) = (localtime)[0..5];
+$y += 1900; $m++;               # usual adjustments
+for ($s, $min, $h, $d, $m) {
+    $_ = "0$_" if $_ < 10;
+}
+$ENV{GL_TS} = "$y-$m-$d.$h:$min:$s";
+
+# substitute template parameters and set the logfile name
+$GL_LOGT =~ s/%y/$y/g;
+$GL_LOGT =~ s/%m/$m/g;
+$GL_LOGT =~ s/%d/$d/g;
+$ENV{GL_LOG} = $GL_LOGT;
+
 # ----------------------------------------------------------------------------
 #       sanity checks on SSH_ORIGINAL_COMMAND
 # ----------------------------------------------------------------------------
 
-# print basic access info if SSH_ORIGINAL_COMMAND does not exist
+# no SSH_ORIGINAL_COMMAND given...
 unless ($ENV{SSH_ORIGINAL_COMMAND}) {
-    # unless the user is allowed to use a shell
+    # if the user is allowed to use a shell, give him one
     if ($shell_allowed) {
         my $shell = $ENV{SHELL};
         $shell =~ s/.*\//-/;    # change "/bin/bash" to "-bash"
         exec { $ENV{SHELL} } $shell;
     }
-    &report_basic($GL_ADMINDIR, $GL_CONF_COMPILED, $user);
+    # otherwise, pretend he typed in "info" and carry on...
-    exit 1;
+    $ENV{SSH_ORIGINAL_COMMAND} = 'info';
 }
 
-my $cmd = $ENV{SSH_ORIGINAL_COMMAND};
-my $repo_base_abs = ( $REPO_BASE =~ m(^/) ? $REPO_BASE : "$ENV{HOME}/$REPO_BASE" );
-
 # ----------------------------------------------------------------------------
 #       get and set perms for actual repo created by wildcard-autoviv
 # ----------------------------------------------------------------------------
@@ -92,7 +109,8 @@ my $CUSTOM_COMMANDS=qr/^\s*(expand|getperms|setperms)\s/;
 # back; they all blithely take advantage of the fact that processing custom
 # commands is sort of a dead end for normal (git) processing
 
-if ($cmd =~ $CUSTOM_COMMANDS) {
+if ($ENV{SSH_ORIGINAL_COMMAND} =~ $CUSTOM_COMMANDS) {
+    my $cmd = $ENV{SSH_ORIGINAL_COMMAND};
     my ($verb, $repo) = ($cmd =~ /^\s*(\S+)\s+\/?(.*?)(?:.git)?$/);
     if ($repo =~ $REPONAME_PATT and $verb =~ /getperms|setperms/) {
         # with an actual reponame, you can "getperms" or "setperms"
@@ -108,35 +126,36 @@ if ($cmd =~ $CUSTOM_COMMANDS) {
     exit 0;
 }
 
-# people allowed to get a shell can get basic access info by asking nicely
-if ($cmd eq 'info') {
-    &report_basic($GL_ADMINDIR, $GL_CONF_COMPILED, $user);
-    print "you also have shell access\n\r" if $shell_allowed;
-    exit 0;
-}
-
 # ----------------------------------------------------------------------------
-#       normal (git) processing
+#       non-git commands
 # ----------------------------------------------------------------------------
 
-# split into command and arguments; the pattern allows old style as well as
+# if the command does NOT fit the pattern of a normal git command, send it off
-# new style: "git-subcommand arg" or "git subcommand arg", just like gitosis
+# somewhere else...
-# does, although I'm not sure how necessary that is
+
-#
+# side notes on detecting a normal git command: the pattern we check allows
-# keep in mind this is how git sends across the command:
+# old style as well as new style ("git-subcommand arg" or "git subcommand
+# arg"), just like gitosis does, although I'm not sure how necessary that is.
+# Currently, this is how git sends across the command (including the single
+# quotes):
 #       git-receive-pack 'reponame.git'
-# including the single quotes
 
-my ($verb, $repo) = ($cmd =~ /^\s*(git\s+\S+|\S+)\s+'\/?(.*?)(?:\.git)?'/);
+my ($verb, $repo) = ($ENV{SSH_ORIGINAL_COMMAND} =~ /^\s*(git\s+\S+|\S+)\s+'\/?(.*?)(?:\.git)?'/);
 unless ( $verb and ( $verb =~ $R_COMMANDS or $verb =~ $W_COMMANDS ) and $repo and $repo =~ $REPONAME_PATT ) {
-    # if the user is allowed a shell, just run the command
+    # ok, it's not a normal git command; call the special command helper
-    exec $ENV{SHELL}, "-c", $ENV{SSH_ORIGINAL_COMMAND} if $shell_allowed;
+    &special_cmd ($GL_ADMINDIR, $GL_CONF_COMPILED, $RSYNC_BASE, $HTPASSWD_FILE);
-    # otherwise, whine
+    exit;
-    die "bad command: $cmd\n";
 }
 die "$repo ends with a slash; I don't like that\n" if $repo =~ /\/$/;
 die "$repo has two consecutive periods; I don't like that\n" if $repo =~ /\.\./;
 
+# reponame
+$ENV{GL_REPO}=$repo;
+
+# ----------------------------------------------------------------------------
+#       the real git commands (git-receive-pack, etc...)
+# ----------------------------------------------------------------------------
+
 # ----------------------------------------------------------------------------
 #       first level permissions check
 # ----------------------------------------------------------------------------
@@ -163,32 +182,11 @@ die "$perm access for $repo DENIED to $user\n"
     unless $repos{$repo}{$perm}{$user}
         or $repos{$repo}{$perm}{'@all'};
 
-# ----------------------------------------------------------------------------
-#       logging, timestamp.  also setup env vars for later
-# ----------------------------------------------------------------------------
-
-# reponame
-$ENV{GL_REPO}=$repo;
-
-# timestamp
-my ($s, $min, $h, $d, $m, $y) = (localtime)[0..5];
-$y += 1900; $m++;               # usual adjustments
-for ($s, $min, $h, $d, $m) {
-    $_ = "0$_" if $_ < 10;
-}
-$ENV{GL_TS} = "$y-$m-$d.$h:$min:$s";
-
-# substitute template parameters and set the logfile name
-$GL_LOGT =~ s/%y/$y/g;
-$GL_LOGT =~ s/%m/$m/g;
-$GL_LOGT =~ s/%d/$d/g;
-$ENV{GL_LOG} = $GL_LOGT;
-
-&log_it("$ENV{GL_TS}\t$ENV{SSH_ORIGINAL_COMMAND}\t$user\n");
-
 # ----------------------------------------------------------------------------
 #       over to git now
 # ----------------------------------------------------------------------------
 
+&log_it("$ENV{GL_TS}\t$ENV{SSH_ORIGINAL_COMMAND}\t$user\n");
+
 $repo = "'$REPO_BASE/$repo.git'";
 exec("git", "shell", "-c", "$verb $repo");

src/gl-compile-conf

@@ -312,7 +312,6 @@ sub parse_conf_file
             die "$WARN $fragment attempting to set description for $repo\n" if
                 $fragment ne 'master' and $fragment ne $repo and ($groups{"\@$fragment"}{$repo} || '') ne 'master';
             $desc{"$repo.git"} = $desc;
-            $owner =~ s/ /+/g if $owner;    # gitweb/INSTALL wants more, but meh...!
             $owner{"$repo.git"} = $owner || '';
         }
         else
@@ -389,8 +388,9 @@ wrap_chdir("$repo_base_abs");
 
 for my $repo (sort keys %repos) {
     next unless $repo =~ $REPONAME_PATT;
-    print STDERR "creating $repo...\n";
+    next if $repo =~ m(^EXTCMD/);   # these are not real repos
     unless (-d "$repo.git") {
+        print STDERR "creating $repo...\n";
         new_repo($repo, "$GL_ADMINDIR/src/hooks");
         # new_repo would have chdir'd us away; come back
         wrap_chdir("$repo_base_abs");
@@ -451,16 +451,32 @@ for my $repo (sort keys %repos) {
         $projlist{"$repo.git"} = 1;
         # add the description file; no messages to user or error checking :)
         $desc{"$repo.git"} and open(DESC, ">", $desc_file) and print DESC $desc{"$repo.git"} . "\n" and close DESC;
+        if ($owner{"$repo.git"}) {
+            # set the repository owner
+            system("git", "--git-dir=$repo.git", "config", "gitweb.owner", $owner{"$repo.git"});
+        } else {
+            # remove the repository owner setting
+            system("git --git-dir=$repo.git config --unset-all gitweb.owner 2>/dev/null");
+        }
     } else {
         # delete the description file; no messages to user or error checking :)
         unlink $desc_file;
+        # remove the repository owner setting
+        system("git --git-dir=$repo.git config --unset-all gitweb.owner 2>/dev/null");
+    }
+
+    # unless there are other gitweb.* keys set, remove the section to keep the
+    # config file clean
+    my $keys = `git --git-dir=$repo.git config --get-regexp '^gitweb\\.' 2>/dev/null`;
+    if (length($keys) == 0) {
+        system("git --git-dir=$repo.git config --remove-section gitweb 2>/dev/null");
     }
 }
 
 # update the project list
 my $projlist_fh = wrap_open( ">", $PROJECTS_LIST);
 for my $proj (sort keys %projlist) {
-    print $projlist_fh "$proj" . ( $owner{$proj} ? " $owner{$proj}" : "" ) . "\n";
+    print $projlist_fh "$proj\n";
 }
 close $projlist_fh;