Small documentation fixes, add better verbosity when saving connection cache data.

Make sure the hasSubordinates attribute is defined before checking its value.

fix 'ls -R' output, minor style cleanup

Repair broken path behavior, remove unneeded #path_to_dn 'relative' flag.

[PATCH 19/19] remove rdn_to_dn() after its last users are gone From 892013debac0aef9937ecfbf2c8aab72c88e07cc Mon Sep 17 00:00:00 2001 Signed-off-by: Peter Marschall <peter@adpm.de> --- shelldap | 21 --------------------- 1 files changed, 0 insertions(+), 21 deletions(-)

run_cat: convert to using path_to_dn() run_edit: convert to using path_to_dn() run_copy & run_move: convert to using path_to_dn() run_grep: convert to using path_to_dn() run_passwd: convert to using path_to_dn()

add method path_to_dn() to convert a given "path" to a DN path_to_dn() replaces the occurrences of '~', '.' and '..' in a path given and returns a DN. However, it does not check whether the DN is valid. Especially: - on return it is not guaranteed that the DN exists - on return the first part does not need to be a valid RDN

mkdir: support more objectclasses Depending on the naming attribute given, support the objectclasses 'country' and 'organization' in addition to the default 'organizationalUnit'.

make_filter: cope with filters that are already parenthesized Treat filter elements correctly that may be more complex filters themselves; e.g. '(&(sn=Doe)(givenname=John))'

run_list: new argument syntax: [<options>] [<filter>] [<attributes>] From 232fbd24ff43c9c0d0691cf0e1b51a82ef099489 Mon Sep 17 00:00:00 2001 Make run_list work with a properly defined argument syntax: - start with (optional) options: -R -l - continue with filter ['(objectclass=*)' as fallback if none given] - end with attributes (also optional) Add method is_valid_filter() to check whether a strig is a legal LDAP filter.

fix attribute lists for LDAP queries LDAP does not know of an attribute named 'dn'. To get only the DN in a search, the attriibute list to use is '1.1'. On all other cases, the DN of the entries found is automaticlly part of the result set too.

remove now unused parent_dn() method

cd: flexible treatment of repeated '..', even as prefix Treat '..' as any shell does: - while the path given starts with '..', strip away the first element of the current base DN - use ',' as separator - if anything remains in thep ath given, prepend it to the stripped down baseDN - use the result as the new base DN

small style cleanup

base(): make more secure, allow '' as DN Only accept DNs as arguments to base that are legal DNs. Convert DN given into canonical form.

slight cleanup: make more clear, it's an array

use sane way to get a default basedn: RootDSe's namingContexts

Add documentation for the additional short flags.

accept short option names for some options Accept short name equivalents like in ldap... commands for 'server', 'basedn' and 'binddn'.

simplify over-complex call of N:L:E->get_value() @{ Net::LDAP::Entry->get_value(..., asref => 1) } is equivalent to Net::LDAP::Entry->get_value(...)

use symbolic LDAP error codes instead of numbers

Exit with a nicer error message if IO::Socket::SSL isn't installed, but the user is requesting SSL/TLS. (this is normally required by Net::LDAP.)

Allow '-' on RDN name when copying

Add a quick Makefile to automate future release tarballs.

Bump to version 0.4.

Added tag 0.4 for changeset 664bbe3dcd44

Follow regular man page conventions. Patch from Salvatore Bonaccorso <salvatore.bonaccorso@gmail.com>.

Minor cleanup.

Improve performance for cd/ls for containers with a large number of entries. Patch from Yann Cezard <yann.cezard@univ-pau.fr>.

Tagging for release 0.3.