IntenseWebs/freeipa
3
0
Fork 0
mirror of https://salsa.debian.org/freeipa-team/freeipa.git synced 2024-12-30 10:47:08 -06:00
Code Issues Packages Projects Releases Wiki Activity
d62b2d9be5
freeipa/ipalib/plugins/permission.py

445 lines
15 KiB
Python
Raw Normal View History

Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
# Authors:
# Rob Crittenden <rcritten@redhat.com>
#
# Copyright (C) 2010 Red Hat
# see file 'COPYING' for use and warranty information
#
Change FreeIPA license to GPLv3+ The changes include: * Change license blobs in source files to mention GPLv3+ not GPLv2 only * Add GPLv3+ license text * Package COPYING not LICENSE as the license blobs (even the old ones) mention COPYING specifically, it is also more common, I think https://fedorahosted.org/freeipa/ticket/239
2010-12-09 06:59:11 -06:00
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
Change FreeIPA license to GPLv3+ The changes include: * Change license blobs in source files to mention GPLv3+ not GPLv2 only * Add GPLv3+ license text * Package COPYING not LICENSE as the license blobs (even the old ones) mention COPYING specifically, it is also more common, I think https://fedorahosted.org/freeipa/ticket/239
2010-12-09 06:59:11 -06:00
# along with this program. If not, see <http://www.gnu.org/licenses/>.
ticket 1669 - improve i18n docstring extraction This patch reverts the use of pygettext for i18n string extraction. It was originally introduced because the help documentation for commands are in the class docstring and module docstring. Docstrings are a Python construct whereby any string which immediately follows a class declaration, function/method declaration or appears first in a module is taken to be the documentation for that object. Python automatically assigns that string to the __doc__ variable associated with the object. Explicitly assigning to the __doc__ variable is equivalent and permitted. We mark strings in the source for i18n translation by embedding them in _() or ngettext(). Specialized extraction tools (e.g. xgettext) scan the source code looking for strings with those markers and extracts the string for inclusion in a translation catalog. It was mistakingly assumed one could not mark for translation Python docstrings. Since some docstrings are vital for our command help system some method had to be devised to extract docstrings for the translation catalog. pygettext has the ability to locate and extract docstrings and it was introduced to acquire the documentation for our commands located in module and class docstrings. However pygettext was too large a hammer for this task, it lacked any fined grained ability to extract only the docstrings we were interested in. In practice it extracted EVERY docstring in each file it was presented with. This caused a large number strings to be extracted for translation which had no reason to be translated, the string might have been internal code documentation never meant to be seen by users. Often the superfluous docstrings were long, complex and likely difficult to translate. This placed an unnecessary burden on our volunteer translators. Instead what is needed is some method to extract only those strings intended for translation. We already have such a mechanism and it is already widely used, namely wrapping strings intended for translation in calls to _() or _negettext(), i.e. marking a string for i18n translation. Thus the solution to the docstring translation problem is to mark the docstrings exactly as we have been doing, it only requires that instead of a bare Python docstring we instead assign the marked string to the __doc__ variable. Using the hypothetical class foo as an example. class foo(Command): ''' The foo command takes out the garbage. ''' Would become: class foo(Command): __doc__ = _('The foo command takes out the garbage.') But which docstrings need to be marked for translation? The makeapi tool knows how to iterate over every command in our public API. It was extended to validate every command's documentation and report if any documentation is missing or not marked for translation. That information was then used to identify each docstring in the code which needed to be transformed. In summary what this patch does is: * Remove the use of pygettext (modification to install/po/Makefile.in) * Replace every docstring with an explicit assignment to __doc__ where the rhs of the assignment is an i18n marking function. * Single line docstrings appearing in multi-line string literals (e.g. ''' or """) were replaced with single line string literals because the multi-line literals were introducing unnecessary whitespace and newlines in the string extracted for translation. For example: ''' The foo command takes out the garbage. ''' Would appear in the translation catalog as: "\n The foo command takes out the garbage.\n " The superfluous whitespace and newlines are confusing to translators and requires us to strip leading and trailing whitespace from the translation at run time. * Import statements were moved from below the docstring to above it. This was necessary because the i18n markers are imported functions and must be available before the the doc is parsed. Technically only the import of the i18n markers had to appear before the doc but stylistically it's better to keep all the imports together. * It was observed during the docstring editing process that the command documentation was inconsistent with respect to the use of periods to terminate a sentence. Some doc had a trailing period, others didn't. Consistency was enforced by adding a period to end of every docstring if one was missing.
2011-08-24 21:48:30 -05:00
import copy
Limit permission and selfservice names to alphanumerics, -, _, space The DN and ACI code doesn't always escape special characters properly. Rather than trying to fix it, this patch takes the easy way out and enforces that the names are safe. https://fedorahosted.org/freeipa/ticket/2585
2012-04-06 03:56:46 -05:00
ticket 1669 - improve i18n docstring extraction This patch reverts the use of pygettext for i18n string extraction. It was originally introduced because the help documentation for commands are in the class docstring and module docstring. Docstrings are a Python construct whereby any string which immediately follows a class declaration, function/method declaration or appears first in a module is taken to be the documentation for that object. Python automatically assigns that string to the __doc__ variable associated with the object. Explicitly assigning to the __doc__ variable is equivalent and permitted. We mark strings in the source for i18n translation by embedding them in _() or ngettext(). Specialized extraction tools (e.g. xgettext) scan the source code looking for strings with those markers and extracts the string for inclusion in a translation catalog. It was mistakingly assumed one could not mark for translation Python docstrings. Since some docstrings are vital for our command help system some method had to be devised to extract docstrings for the translation catalog. pygettext has the ability to locate and extract docstrings and it was introduced to acquire the documentation for our commands located in module and class docstrings. However pygettext was too large a hammer for this task, it lacked any fined grained ability to extract only the docstrings we were interested in. In practice it extracted EVERY docstring in each file it was presented with. This caused a large number strings to be extracted for translation which had no reason to be translated, the string might have been internal code documentation never meant to be seen by users. Often the superfluous docstrings were long, complex and likely difficult to translate. This placed an unnecessary burden on our volunteer translators. Instead what is needed is some method to extract only those strings intended for translation. We already have such a mechanism and it is already widely used, namely wrapping strings intended for translation in calls to _() or _negettext(), i.e. marking a string for i18n translation. Thus the solution to the docstring translation problem is to mark the docstrings exactly as we have been doing, it only requires that instead of a bare Python docstring we instead assign the marked string to the __doc__ variable. Using the hypothetical class foo as an example. class foo(Command): ''' The foo command takes out the garbage. ''' Would become: class foo(Command): __doc__ = _('The foo command takes out the garbage.') But which docstrings need to be marked for translation? The makeapi tool knows how to iterate over every command in our public API. It was extended to validate every command's documentation and report if any documentation is missing or not marked for translation. That information was then used to identify each docstring in the code which needed to be transformed. In summary what this patch does is: * Remove the use of pygettext (modification to install/po/Makefile.in) * Replace every docstring with an explicit assignment to __doc__ where the rhs of the assignment is an i18n marking function. * Single line docstrings appearing in multi-line string literals (e.g. ''' or """) were replaced with single line string literals because the multi-line literals were introducing unnecessary whitespace and newlines in the string extracted for translation. For example: ''' The foo command takes out the garbage. ''' Would appear in the translation catalog as: "\n The foo command takes out the garbage.\n " The superfluous whitespace and newlines are confusing to translators and requires us to strip leading and trailing whitespace from the translation at run time. * Import statements were moved from below the docstring to above it. This was necessary because the i18n markers are imported functions and must be available before the the doc is parsed. Technically only the import of the i18n markers had to appear before the doc but stylistically it's better to keep all the imports together. * It was observed during the docstring editing process that the command documentation was inconsistent with respect to the use of periods to terminate a sentence. Some doc had a trailing period, others didn't. Consistency was enforced by adding a period to end of every docstring if one was missing.
2011-08-24 21:48:30 -05:00
from ipalib.plugins.baseldap import *
from ipalib import api, _, ngettext
from ipalib import Flag, Str, StrEnum
from ipalib.request import context
from ipalib import errors
__doc__ = _("""
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
Permissions
Make permission and privilege help text clearer ticket 996
2011-02-22 15:24:41 -06:00
A permission enables fine-grained delegation of rights. A permission is
a human-readable form of a 389-ds Access Control Rule, or instruction (ACI).
A permission grants the right to perform a specific task such as adding a
user, modifying a group, etc.
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
Make permission and privilege help text clearer ticket 996
2011-02-22 15:24:41 -06:00
A permission may not contain other permissions.
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
Fix style and grammatical issues in built-in command help. There is a rather large API.txt change but it is only due to changes in the doc string in parameters. ticket 729
2011-03-04 10:08:54 -06:00
* A permission grants access to read, write, add or delete.
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
* A privilege combines similar permissions (for example all the permissions
needed to add a user).
* A role grants a set of privileges to users, groups, hosts or hostgroups.
A permission is made up of a number of different parts:
1. The name of the permission.
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
2. The target of the permission.
Fix style and grammatical issues in built-in command help. There is a rather large API.txt change but it is only due to changes in the doc string in parameters. ticket 729
2011-03-04 10:08:54 -06:00
3. The rights granted by the permission.
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
Fix style and grammatical issues in built-in command help. There is a rather large API.txt change but it is only due to changes in the doc string in parameters. ticket 729
2011-03-04 10:08:54 -06:00
Rights define what operations are allowed, and may be one or more
of the following:
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
1. write - write one or more attributes
2. read - read one or more attributes
3. add - add a new entry to the tree
4. delete - delete an existing entry
5. all - all permissions are granted
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
Read permission is granted for most attributes by default so the read
permission is not expected to be used very often.
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
Note the distinction between attributes and entries. The permissions are
independent, so being able to add a user does not mean that the user will
Fix style and grammatical issues in built-in command help. There is a rather large API.txt change but it is only due to changes in the doc string in parameters. ticket 729
2011-03-04 10:08:54 -06:00
be editable.
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
There are a number of allowed targets:
1. type: a type of object (user, group, etc).
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
2. memberof: a member of a group or hostgroup
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
3. filter: an LDAP filter
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
4. subtree: an LDAP filter specifying part of the LDAP DIT. This is a
Fix style and grammatical issues in built-in command help. There is a rather large API.txt change but it is only due to changes in the doc string in parameters. ticket 729
2011-03-04 10:08:54 -06:00
super-set of the "type" target.
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
5. targetgroup: grant access to modify a specific group (such as granting
the rights to manage group membership)
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
EXAMPLES:
Add a permission that grants the creation of users:
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
ipa permission-add --type=user --permissions=add "Add Users"
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
Add a permission that grants the ability to manage group membership:
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
ipa permission-add --attrs=member --permissions=write --type=group "Manage Group Members"
ticket 1669 - improve i18n docstring extraction This patch reverts the use of pygettext for i18n string extraction. It was originally introduced because the help documentation for commands are in the class docstring and module docstring. Docstrings are a Python construct whereby any string which immediately follows a class declaration, function/method declaration or appears first in a module is taken to be the documentation for that object. Python automatically assigns that string to the __doc__ variable associated with the object. Explicitly assigning to the __doc__ variable is equivalent and permitted. We mark strings in the source for i18n translation by embedding them in _() or ngettext(). Specialized extraction tools (e.g. xgettext) scan the source code looking for strings with those markers and extracts the string for inclusion in a translation catalog. It was mistakingly assumed one could not mark for translation Python docstrings. Since some docstrings are vital for our command help system some method had to be devised to extract docstrings for the translation catalog. pygettext has the ability to locate and extract docstrings and it was introduced to acquire the documentation for our commands located in module and class docstrings. However pygettext was too large a hammer for this task, it lacked any fined grained ability to extract only the docstrings we were interested in. In practice it extracted EVERY docstring in each file it was presented with. This caused a large number strings to be extracted for translation which had no reason to be translated, the string might have been internal code documentation never meant to be seen by users. Often the superfluous docstrings were long, complex and likely difficult to translate. This placed an unnecessary burden on our volunteer translators. Instead what is needed is some method to extract only those strings intended for translation. We already have such a mechanism and it is already widely used, namely wrapping strings intended for translation in calls to _() or _negettext(), i.e. marking a string for i18n translation. Thus the solution to the docstring translation problem is to mark the docstrings exactly as we have been doing, it only requires that instead of a bare Python docstring we instead assign the marked string to the __doc__ variable. Using the hypothetical class foo as an example. class foo(Command): ''' The foo command takes out the garbage. ''' Would become: class foo(Command): __doc__ = _('The foo command takes out the garbage.') But which docstrings need to be marked for translation? The makeapi tool knows how to iterate over every command in our public API. It was extended to validate every command's documentation and report if any documentation is missing or not marked for translation. That information was then used to identify each docstring in the code which needed to be transformed. In summary what this patch does is: * Remove the use of pygettext (modification to install/po/Makefile.in) * Replace every docstring with an explicit assignment to __doc__ where the rhs of the assignment is an i18n marking function. * Single line docstrings appearing in multi-line string literals (e.g. ''' or """) were replaced with single line string literals because the multi-line literals were introducing unnecessary whitespace and newlines in the string extracted for translation. For example: ''' The foo command takes out the garbage. ''' Would appear in the translation catalog as: "\n The foo command takes out the garbage.\n " The superfluous whitespace and newlines are confusing to translators and requires us to strip leading and trailing whitespace from the translation at run time. * Import statements were moved from below the docstring to above it. This was necessary because the i18n markers are imported functions and must be available before the the doc is parsed. Technically only the import of the i18n markers had to appear before the doc but stylistically it's better to keep all the imports together. * It was observed during the docstring editing process that the command documentation was inconsistent with respect to the use of periods to terminate a sentence. Some doc had a trailing period, others didn't. Consistency was enforced by adding a period to end of every docstring if one was missing.
2011-08-24 21:48:30 -05:00
""")
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
ACI_PREFIX=u"permission"
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
output_params = (
Str('ipapermissiontype',
label=_('Permission Type'),
),
Fix raw format for ACI commands ACI plugins (permission, selfservice and delegation) were not prepared to serve ACIs in a raw format, i.e. raw "aci" attribute taken from LDAP. This patch fixes all these plugins and their commands to provide provide this format. Few ACI raw format unit tests were added for all these plugins. https://fedorahosted.org/freeipa/ticket/2010 https://fedorahosted.org/freeipa/ticket/2223 https://fedorahosted.org/freeipa/ticket/2228 https://fedorahosted.org/freeipa/ticket/2232
2012-02-02 14:28:15 -06:00
Str('aci',
label=_('ACI'),
),
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
)
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
class permission(LDAPObject):
"""
Permission object.
"""
container_dn = api.env.container_permission
Fixed object_name and object_name_plural internationalization The object_name, object_name_plural and messages that use these attributes have been converted to support translation. The label attribute in the Param class has been modified to accept unicode string. Ticket #1435
2011-07-12 11:01:25 -05:00
object_name = _('permission')
object_name_plural = _('permissions')
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
object_class = ['groupofnames', 'ipapermission']
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
default_attributes = ['cn', 'member', 'memberof',
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
'memberindirect', 'ipapermissiontype',
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
]
Fix raw format for ACI commands ACI plugins (permission, selfservice and delegation) were not prepared to serve ACIs in a raw format, i.e. raw "aci" attribute taken from LDAP. This patch fixes all these plugins and their commands to provide provide this format. Few ACI raw format unit tests were added for all these plugins. https://fedorahosted.org/freeipa/ticket/2010 https://fedorahosted.org/freeipa/ticket/2223 https://fedorahosted.org/freeipa/ticket/2228 https://fedorahosted.org/freeipa/ticket/2232
2012-02-02 14:28:15 -06:00
aci_attributes = ['aci', 'group', 'permissions', 'attrs', 'type',
Display the value of memberOf ACIs in permission plugin. There were two problems: 1. memberof wasn't in the list of things we looked for in the return value from aci_show() 2. The value wasn't being translated into a group name. Use the DN class to retrieve the group name from the memberof URI. Note that I changed the parsing for targetgroup as well. We now save a lookup and potentially returning a NotFound if an aci points to a group that no longer exists. https://fedorahosted.org/freeipa/ticket/2100
2011-12-06 17:15:41 -06:00
'filter', 'subtree', 'targetgroup', 'memberof',
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
]
attribute_members = {
'member': ['privilege'],
}
Only apply validation rules when adding and updating. There may be cases, for whatever reason, that an otherwise illegal entry gets created that doesn't match the criteria for a valid user/host/group name. If this happens (i.e. migration) there is no way to remove this using the IPA tools because we always applied the name pattern. So you can't, for example, delete a user with an illegal name. Primary keys are cloned with query=True in PKQuery which causes no rules to be applied on mod/show/find. This reverts a change from commit 3a5e26a0 which applies class rules when query=True (for enforcing no white space). Replace rdnattr with rdn_is_primary_key. This was meant to tell us when an RDN change was necessary to do a rename. There could be a disconnect where the rdnattr wasn't the primary key and in that case we don't need to do an RDN change, so use a boolean instead so that it is clear that RDN == primary key. Add a test to ensure that nowhitespace is actually enforced. https://fedorahosted.org/freeipa/ticket/2115 Related: https://fedorahosted.org/freeipa/ticket/2089 Whitespace tickets: https://fedorahosted.org/freeipa/ticket/1285 https://fedorahosted.org/freeipa/ticket/1286 https://fedorahosted.org/freeipa/ticket/1287
2012-02-29 12:31:20 -06:00
rdn_is_primary_key = True
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
label = _('Permissions')
Fixed label capitalization The CSS text-transform sometimes produces incorrect capitalization, so the code has been modified to use translated labels that already contain the correct capitalization. Ticket #1424
2011-07-13 21:10:47 -05:00
label_singular = _('Permission')
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
takes_params = (
Str('cn',
cli_name='name',
label=_('Permission name'),
primary_key=True,
Limit permission and selfservice names to alphanumerics, -, _, space The DN and ACI code doesn't always escape special characters properly. Rather than trying to fix it, this patch takes the easy way out and enforces that the names are safe. https://fedorahosted.org/freeipa/ticket/2585
2012-04-06 03:56:46 -05:00
pattern='^[-_ a-zA-Z0-9]+$',
pattern_errmsg="May only contain letters, numbers, -, _, and space",
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
),
Parse comma-separated lists of values in all parameter types. This can be enabled for a specific parameter by setting the "csv" option to True. Remove "List" parameter type and replace all occurences of it with appropriate multi-valued parameter ("Str" in most cases) with csv enabled. Add new parameter type "Any", capable of holding values of any type. This is needed by the "batch" command, as "Str" is not suitable type for the "methods" parameter. ticket 2007
2011-11-21 09:50:27 -06:00
Str('permissions+',
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
cli_name='permissions',
label=_('Permissions'),
doc=_('Comma-separated list of permissions to grant ' \
'(read, write, add, delete, all)'),
Parse comma-separated lists of values in all parameter types. This can be enabled for a specific parameter by setting the "csv" option to True. Remove "List" parameter type and replace all occurences of it with appropriate multi-valued parameter ("Str" in most cases) with csv enabled. Add new parameter type "Any", capable of holding values of any type. This is needed by the "batch" command, as "Str" is not suitable type for the "methods" parameter. ticket 2007
2011-11-21 09:50:27 -06:00
csv=True,
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
),
Parse comma-separated lists of values in all parameter types. This can be enabled for a specific parameter by setting the "csv" option to True. Remove "List" parameter type and replace all occurences of it with appropriate multi-valued parameter ("Str" in most cases) with csv enabled. Add new parameter type "Any", capable of holding values of any type. This is needed by the "batch" command, as "Str" is not suitable type for the "methods" parameter. ticket 2007
2011-11-21 09:50:27 -06:00
Str('attrs*',
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
cli_name='attrs',
label=_('Attributes'),
doc=_('Comma-separated list of attributes'),
Parse comma-separated lists of values in all parameter types. This can be enabled for a specific parameter by setting the "csv" option to True. Remove "List" parameter type and replace all occurences of it with appropriate multi-valued parameter ("Str" in most cases) with csv enabled. Add new parameter type "Any", capable of holding values of any type. This is needed by the "batch" command, as "Str" is not suitable type for the "methods" parameter. ticket 2007
2011-11-21 09:50:27 -06:00
csv=True,
Don't use camel-case LDAP attributes in ACI and don't clear enrolledBy We keep LDAP attributes lower-case elsewhere in the API we should do the same with all access controls. There were two ACIs pointing at the manage_host_keytab permission. This isn't allowed in general and we have decided separately to not clear out enrolledBy when a host is unenrolled so dropping it is the obvious thing to do. ticket 597
2010-12-17 15:57:28 -06:00
normalizer=lambda value: value.lower(),
permission-mod prompts for all parameters ipa permission-mod was prompting for all parameters because they had specified flag 'ask_update'. The flag was removed. Additionally the exec_callback for permission-mod was updated to unify the behaviour with other ipa commands (raise exception when no modification was specified). https://fedorahosted.org/freeipa/ticket/2280
2012-05-16 06:36:35 -05:00
flags=('ask_create'),
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
),
StrEnum('type?',
cli_name='type',
label=_('Type'),
Fix 'ipa help permissions'; add 'dns' in allowed types.
2010-12-30 08:12:35 -06:00
doc=_('Type of IPA object (user, group, host, hostgroup, service, netgroup, dns)'),
Changed dns permission types Recent change of DNS module to version caused that dns object type was replaced by dnszone and dnsrecord. This patch corrects dns types in permissions class. https://fedorahosted.org/freeipa/ticket/646
2011-01-13 10:32:57 -06:00
values=(u'user', u'group', u'host', u'service', u'hostgroup', u'netgroup', u'dnsrecord',),
permission-mod prompts for all parameters ipa permission-mod was prompting for all parameters because they had specified flag 'ask_update'. The flag was removed. Additionally the exec_callback for permission-mod was updated to unify the behaviour with other ipa commands (raise exception when no modification was specified). https://fedorahosted.org/freeipa/ticket/2280
2012-05-16 06:36:35 -05:00
flags=('ask_create'),
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
),
Str('memberof?',
cli_name='memberof',
label=_('Member of group'), # FIXME: Does this label make sense?
doc=_('Target members of a group'),
permission-mod prompts for all parameters ipa permission-mod was prompting for all parameters because they had specified flag 'ask_update'. The flag was removed. Additionally the exec_callback for permission-mod was updated to unify the behaviour with other ipa commands (raise exception when no modification was specified). https://fedorahosted.org/freeipa/ticket/2280
2012-05-16 06:36:35 -05:00
flags=('ask_create'),
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
),
Str('filter?',
cli_name='filter',
label=_('Filter'),
doc=_('Legal LDAP filter (e.g. ou=Engineering)'),
permission-mod prompts for all parameters ipa permission-mod was prompting for all parameters because they had specified flag 'ask_update'. The flag was removed. Additionally the exec_callback for permission-mod was updated to unify the behaviour with other ipa commands (raise exception when no modification was specified). https://fedorahosted.org/freeipa/ticket/2280
2012-05-16 06:36:35 -05:00
flags=('ask_create'),
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
),
Str('subtree?',
cli_name='subtree',
label=_('Subtree'),
doc=_('Subtree to apply permissions to'),
permission-mod prompts for all parameters ipa permission-mod was prompting for all parameters because they had specified flag 'ask_update'. The flag was removed. Additionally the exec_callback for permission-mod was updated to unify the behaviour with other ipa commands (raise exception when no modification was specified). https://fedorahosted.org/freeipa/ticket/2280
2012-05-16 06:36:35 -05:00
flags=('ask_create'),
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
),
Str('targetgroup?',
cli_name='targetgroup',
label=_('Target group'),
doc=_('User group to apply permissions to'),
permission-mod prompts for all parameters ipa permission-mod was prompting for all parameters because they had specified flag 'ask_update'. The flag was removed. Additionally the exec_callback for permission-mod was updated to unify the behaviour with other ipa commands (raise exception when no modification was specified). https://fedorahosted.org/freeipa/ticket/2280
2012-05-16 06:36:35 -05:00
flags=('ask_create'),
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
),
)
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
# Don't allow SYSTEM permissions to be modified or removed
def check_system(self, ldap, dn, *keys):
try:
(dn, entry_attrs) = ldap.get_entry(dn, ['ipapermissiontype'])
except errors.NotFound:
self.handle_not_found(*keys)
if 'ipapermissiontype' in entry_attrs:
if 'SYSTEM' in entry_attrs['ipapermissiontype']:
return False
return True
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
api.register(permission)
class permission_add(LDAPCreate):
ticket 1669 - improve i18n docstring extraction This patch reverts the use of pygettext for i18n string extraction. It was originally introduced because the help documentation for commands are in the class docstring and module docstring. Docstrings are a Python construct whereby any string which immediately follows a class declaration, function/method declaration or appears first in a module is taken to be the documentation for that object. Python automatically assigns that string to the __doc__ variable associated with the object. Explicitly assigning to the __doc__ variable is equivalent and permitted. We mark strings in the source for i18n translation by embedding them in _() or ngettext(). Specialized extraction tools (e.g. xgettext) scan the source code looking for strings with those markers and extracts the string for inclusion in a translation catalog. It was mistakingly assumed one could not mark for translation Python docstrings. Since some docstrings are vital for our command help system some method had to be devised to extract docstrings for the translation catalog. pygettext has the ability to locate and extract docstrings and it was introduced to acquire the documentation for our commands located in module and class docstrings. However pygettext was too large a hammer for this task, it lacked any fined grained ability to extract only the docstrings we were interested in. In practice it extracted EVERY docstring in each file it was presented with. This caused a large number strings to be extracted for translation which had no reason to be translated, the string might have been internal code documentation never meant to be seen by users. Often the superfluous docstrings were long, complex and likely difficult to translate. This placed an unnecessary burden on our volunteer translators. Instead what is needed is some method to extract only those strings intended for translation. We already have such a mechanism and it is already widely used, namely wrapping strings intended for translation in calls to _() or _negettext(), i.e. marking a string for i18n translation. Thus the solution to the docstring translation problem is to mark the docstrings exactly as we have been doing, it only requires that instead of a bare Python docstring we instead assign the marked string to the __doc__ variable. Using the hypothetical class foo as an example. class foo(Command): ''' The foo command takes out the garbage. ''' Would become: class foo(Command): __doc__ = _('The foo command takes out the garbage.') But which docstrings need to be marked for translation? The makeapi tool knows how to iterate over every command in our public API. It was extended to validate every command's documentation and report if any documentation is missing or not marked for translation. That information was then used to identify each docstring in the code which needed to be transformed. In summary what this patch does is: * Remove the use of pygettext (modification to install/po/Makefile.in) * Replace every docstring with an explicit assignment to __doc__ where the rhs of the assignment is an i18n marking function. * Single line docstrings appearing in multi-line string literals (e.g. ''' or """) were replaced with single line string literals because the multi-line literals were introducing unnecessary whitespace and newlines in the string extracted for translation. For example: ''' The foo command takes out the garbage. ''' Would appear in the translation catalog as: "\n The foo command takes out the garbage.\n " The superfluous whitespace and newlines are confusing to translators and requires us to strip leading and trailing whitespace from the translation at run time. * Import statements were moved from below the docstring to above it. This was necessary because the i18n markers are imported functions and must be available before the the doc is parsed. Technically only the import of the i18n markers had to appear before the doc but stylistically it's better to keep all the imports together. * It was observed during the docstring editing process that the command documentation was inconsistent with respect to the use of periods to terminate a sentence. Some doc had a trailing period, others didn't. Consistency was enforced by adding a period to end of every docstring if one was missing.
2011-08-24 21:48:30 -05:00
__doc__ = _('Add a new permission.')
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
msg_summary = _('Added permission "%(value)s"')
Fix raw format for ACI commands ACI plugins (permission, selfservice and delegation) were not prepared to serve ACIs in a raw format, i.e. raw "aci" attribute taken from LDAP. This patch fixes all these plugins and their commands to provide provide this format. Few ACI raw format unit tests were added for all these plugins. https://fedorahosted.org/freeipa/ticket/2010 https://fedorahosted.org/freeipa/ticket/2223 https://fedorahosted.org/freeipa/ticket/2228 https://fedorahosted.org/freeipa/ticket/2232
2012-02-02 14:28:15 -06:00
has_output_params = LDAPCreate.has_output_params + output_params
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
def pre_callback(self, ldap, dn, entry_attrs, attrs_list, *keys, **options):
# Test the ACI before going any further
opts = copy.copy(options)
opts['test'] = True
opts['permission'] = keys[-1]
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
opts['aciprefix'] = ACI_PREFIX
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
self.api.Command.aci_add(keys[-1], **opts)
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
# Clear the aci attributes out of the permission entry
for o in options:
try:
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
if o not in ['objectclass']:
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
del entry_attrs[o]
except:
pass
return dn
def post_callback(self, ldap, dn, entry_attrs, *keys, **options):
# Now actually add the aci.
opts = copy.copy(options)
opts['test'] = False
opts['permission'] = keys[-1]
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
opts['aciprefix'] = ACI_PREFIX
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
try:
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
result = self.api.Command.aci_add(keys[-1], **opts)['result']
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
for attr in self.obj.aci_attributes:
if attr in result:
entry_attrs[attr] = result[attr]
Catch ACI errors better when adding a permission. We create the aci with the --test flag to test its validity but it doesn't do the same level of tests that actually adding an aci to LDAP does. Catch any syntax errors that get thrown and clean up as best we can. ticket 621
2010-12-17 14:35:24 -06:00
except errors.InvalidSyntax, e:
# A syntax error slipped past our attempt at validation, clean up
self.api.Command.permission_del(keys[-1])
raise e
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
except Exception, e:
Catch ACI errors better when adding a permission. We create the aci with the --test flag to test its validity but it doesn't do the same level of tests that actually adding an aci to LDAP does. Catch any syntax errors that get thrown and clean up as best we can. ticket 621
2010-12-17 14:35:24 -06:00
# Something bad happened, clean up as much as we can and return
# that error
try:
self.api.Command.permission_del(keys[-1])
except Exception, ignore:
pass
try:
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
self.api.Command.aci_del(keys[-1], aciprefix=ACI_PREFIX)
Catch ACI errors better when adding a permission. We create the aci with the --test flag to test its validity but it doesn't do the same level of tests that actually adding an aci to LDAP does. Catch any syntax errors that get thrown and clean up as best we can. ticket 621
2010-12-17 14:35:24 -06:00
except Exception, ignore:
pass
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
raise e
return dn
api.register(permission_add)
class permission_del(LDAPDelete):
ticket 1669 - improve i18n docstring extraction This patch reverts the use of pygettext for i18n string extraction. It was originally introduced because the help documentation for commands are in the class docstring and module docstring. Docstrings are a Python construct whereby any string which immediately follows a class declaration, function/method declaration or appears first in a module is taken to be the documentation for that object. Python automatically assigns that string to the __doc__ variable associated with the object. Explicitly assigning to the __doc__ variable is equivalent and permitted. We mark strings in the source for i18n translation by embedding them in _() or ngettext(). Specialized extraction tools (e.g. xgettext) scan the source code looking for strings with those markers and extracts the string for inclusion in a translation catalog. It was mistakingly assumed one could not mark for translation Python docstrings. Since some docstrings are vital for our command help system some method had to be devised to extract docstrings for the translation catalog. pygettext has the ability to locate and extract docstrings and it was introduced to acquire the documentation for our commands located in module and class docstrings. However pygettext was too large a hammer for this task, it lacked any fined grained ability to extract only the docstrings we were interested in. In practice it extracted EVERY docstring in each file it was presented with. This caused a large number strings to be extracted for translation which had no reason to be translated, the string might have been internal code documentation never meant to be seen by users. Often the superfluous docstrings were long, complex and likely difficult to translate. This placed an unnecessary burden on our volunteer translators. Instead what is needed is some method to extract only those strings intended for translation. We already have such a mechanism and it is already widely used, namely wrapping strings intended for translation in calls to _() or _negettext(), i.e. marking a string for i18n translation. Thus the solution to the docstring translation problem is to mark the docstrings exactly as we have been doing, it only requires that instead of a bare Python docstring we instead assign the marked string to the __doc__ variable. Using the hypothetical class foo as an example. class foo(Command): ''' The foo command takes out the garbage. ''' Would become: class foo(Command): __doc__ = _('The foo command takes out the garbage.') But which docstrings need to be marked for translation? The makeapi tool knows how to iterate over every command in our public API. It was extended to validate every command's documentation and report if any documentation is missing or not marked for translation. That information was then used to identify each docstring in the code which needed to be transformed. In summary what this patch does is: * Remove the use of pygettext (modification to install/po/Makefile.in) * Replace every docstring with an explicit assignment to __doc__ where the rhs of the assignment is an i18n marking function. * Single line docstrings appearing in multi-line string literals (e.g. ''' or """) were replaced with single line string literals because the multi-line literals were introducing unnecessary whitespace and newlines in the string extracted for translation. For example: ''' The foo command takes out the garbage. ''' Would appear in the translation catalog as: "\n The foo command takes out the garbage.\n " The superfluous whitespace and newlines are confusing to translators and requires us to strip leading and trailing whitespace from the translation at run time. * Import statements were moved from below the docstring to above it. This was necessary because the i18n markers are imported functions and must be available before the the doc is parsed. Technically only the import of the i18n markers had to appear before the doc but stylistically it's better to keep all the imports together. * It was observed during the docstring editing process that the command documentation was inconsistent with respect to the use of periods to terminate a sentence. Some doc had a trailing period, others didn't. Consistency was enforced by adding a period to end of every docstring if one was missing.
2011-08-24 21:48:30 -05:00
__doc__ = _('Delete a permission.')
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
msg_summary = _('Deleted permission "%(value)s"')
def pre_callback(self, ldap, dn, *keys, **options):
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
if not self.obj.check_system(ldap, dn, *keys):
raise errors.ACIError(info='A SYSTEM permission may not be removed')
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
# remove permission even when the underlying ACI is missing
try:
self.api.Command.aci_del(keys[-1], aciprefix=ACI_PREFIX)
except errors.NotFound:
pass
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
return dn
api.register(permission_del)
class permission_mod(LDAPUpdate):
ticket 1669 - improve i18n docstring extraction This patch reverts the use of pygettext for i18n string extraction. It was originally introduced because the help documentation for commands are in the class docstring and module docstring. Docstrings are a Python construct whereby any string which immediately follows a class declaration, function/method declaration or appears first in a module is taken to be the documentation for that object. Python automatically assigns that string to the __doc__ variable associated with the object. Explicitly assigning to the __doc__ variable is equivalent and permitted. We mark strings in the source for i18n translation by embedding them in _() or ngettext(). Specialized extraction tools (e.g. xgettext) scan the source code looking for strings with those markers and extracts the string for inclusion in a translation catalog. It was mistakingly assumed one could not mark for translation Python docstrings. Since some docstrings are vital for our command help system some method had to be devised to extract docstrings for the translation catalog. pygettext has the ability to locate and extract docstrings and it was introduced to acquire the documentation for our commands located in module and class docstrings. However pygettext was too large a hammer for this task, it lacked any fined grained ability to extract only the docstrings we were interested in. In practice it extracted EVERY docstring in each file it was presented with. This caused a large number strings to be extracted for translation which had no reason to be translated, the string might have been internal code documentation never meant to be seen by users. Often the superfluous docstrings were long, complex and likely difficult to translate. This placed an unnecessary burden on our volunteer translators. Instead what is needed is some method to extract only those strings intended for translation. We already have such a mechanism and it is already widely used, namely wrapping strings intended for translation in calls to _() or _negettext(), i.e. marking a string for i18n translation. Thus the solution to the docstring translation problem is to mark the docstrings exactly as we have been doing, it only requires that instead of a bare Python docstring we instead assign the marked string to the __doc__ variable. Using the hypothetical class foo as an example. class foo(Command): ''' The foo command takes out the garbage. ''' Would become: class foo(Command): __doc__ = _('The foo command takes out the garbage.') But which docstrings need to be marked for translation? The makeapi tool knows how to iterate over every command in our public API. It was extended to validate every command's documentation and report if any documentation is missing or not marked for translation. That information was then used to identify each docstring in the code which needed to be transformed. In summary what this patch does is: * Remove the use of pygettext (modification to install/po/Makefile.in) * Replace every docstring with an explicit assignment to __doc__ where the rhs of the assignment is an i18n marking function. * Single line docstrings appearing in multi-line string literals (e.g. ''' or """) were replaced with single line string literals because the multi-line literals were introducing unnecessary whitespace and newlines in the string extracted for translation. For example: ''' The foo command takes out the garbage. ''' Would appear in the translation catalog as: "\n The foo command takes out the garbage.\n " The superfluous whitespace and newlines are confusing to translators and requires us to strip leading and trailing whitespace from the translation at run time. * Import statements were moved from below the docstring to above it. This was necessary because the i18n markers are imported functions and must be available before the the doc is parsed. Technically only the import of the i18n markers had to appear before the doc but stylistically it's better to keep all the imports together. * It was observed during the docstring editing process that the command documentation was inconsistent with respect to the use of periods to terminate a sentence. Some doc had a trailing period, others didn't. Consistency was enforced by adding a period to end of every docstring if one was missing.
2011-08-24 21:48:30 -05:00
__doc__ = _('Modify a permission.')
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
msg_summary = _('Modified permission "%(value)s"')
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
has_output_params = LDAPUpdate.has_output_params + output_params
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
def pre_callback(self, ldap, dn, entry_attrs, attrs_list, *keys, **options):
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
if not self.obj.check_system(ldap, dn, *keys):
raise errors.ACIError(info='A SYSTEM permission may not be modified')
ipa permission-mod --rename does not work This patch fixes nonfunctional rename operation in permission plugin. Also makes sure, that no change is made to the underlying ACI in pre_callback() when the target permission already exists. Several tests for the rename operation have been created to ensure that the it won't break again unnoticed. https://fedorahosted.org/freeipa/ticket/814
2011-01-27 05:17:10 -06:00
# check if permission is in LDAP
Better handle permission object updates versus aci object updates. permissions are a real group pointed to by an aci, managed by the same plugin. Any given update can update one or both or neither. Do a better job at determining what it is that needs to be updated and handle the case where only the ACI is updated so that EmptyModList is not thrown. ticket 603
2010-12-10 21:52:44 -06:00
try:
(dn, attrs) = ldap.get_entry(
dn, attrs_list, normalize=self.obj.normalize_dn
)
except errors.NotFound:
self.obj.handle_not_found(*keys)
ipa permission-mod --rename does not work This patch fixes nonfunctional rename operation in permission plugin. Also makes sure, that no change is made to the underlying ACI in pre_callback() when the target permission already exists. Several tests for the rename operation have been created to ensure that the it won't break again unnoticed. https://fedorahosted.org/freeipa/ticket/814
2011-01-27 05:17:10 -06:00
# when renaming permission, check if the target permission does not
# exists already. Then, make changes to underlying ACI
if 'rename' in options:
Unable to rename permission object The update was failing because of the case insensitivity of permission object DN. Unit-tests added. https://fedorahosted.org/freeipa/ticket/2571
2012-04-11 02:37:15 -05:00
if options['rename']:
try:
new_dn = dn.replace(keys[-1].lower(), options['rename'], 1)
(new_dn, attrs) = ldap.get_entry(
new_dn, attrs_list, normalize=self.obj.normalize_dn
)
raise errors.DuplicateEntry()
except errors.NotFound:
pass # permission may be renamed, continue
else:
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
raise errors.ValidationError(
name='rename',error=_('New name can not be empty'))
ipa permission-mod --rename does not work This patch fixes nonfunctional rename operation in permission plugin. Also makes sure, that no change is made to the underlying ACI in pre_callback() when the target permission already exists. Several tests for the rename operation have been created to ensure that the it won't break again unnoticed. https://fedorahosted.org/freeipa/ticket/814
2011-01-27 05:17:10 -06:00
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
opts = copy.copy(options)
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
for o in ['all', 'raw', 'rights', 'test', 'rename']:
opts.pop(o, None)
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
setattr(context, 'aciupdate', False)
# If there are no options left we don't need to do anything to the
# underlying ACI.
if len(opts) > 0:
opts['permission'] = keys[-1]
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
opts['aciprefix'] = ACI_PREFIX
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
self.api.Command.aci_mod(keys[-1], **opts)
setattr(context, 'aciupdate', True)
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
# Clear the aci attributes out of the permission entry
for o in self.obj.aci_attributes:
try:
del entry_attrs[o]
except:
pass
return dn
def exc_callback(self, keys, options, exc, call_func, *call_args, **call_kwargs):
Refactor exc_callback invocation. Replace _call_exc_callbacks with a function wrapper, which will automatically call exception callbacks when an exception is raised from the function. This removes the need to specify the function and its arguments twice (once in the function call itself and once in _call_exc_callbacks). Add some extra checks to existing exception callbacks.
2012-04-19 07:06:32 -05:00
if call_func.func_name == 'update_entry':
if isinstance(exc, errors.EmptyModlist):
aciupdate = getattr(context, 'aciupdate')
permission-mod prompts for all parameters ipa permission-mod was prompting for all parameters because they had specified flag 'ask_update'. The flag was removed. Additionally the exec_callback for permission-mod was updated to unify the behaviour with other ipa commands (raise exception when no modification was specified). https://fedorahosted.org/freeipa/ticket/2280
2012-05-16 06:36:35 -05:00
if aciupdate:
Refactor exc_callback invocation. Replace _call_exc_callbacks with a function wrapper, which will automatically call exception callbacks when an exception is raised from the function. This removes the need to specify the function and its arguments twice (once in the function call itself and once in _call_exc_callbacks). Add some extra checks to existing exception callbacks.
2012-04-19 07:06:32 -05:00
return
raise exc
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
def post_callback(self, ldap, dn, entry_attrs, *keys, **options):
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
# rename the underlying ACI after the change to permission
ipa permission-mod --rename does not work This patch fixes nonfunctional rename operation in permission plugin. Also makes sure, that no change is made to the underlying ACI in pre_callback() when the target permission already exists. Several tests for the rename operation have been created to ensure that the it won't break again unnoticed. https://fedorahosted.org/freeipa/ticket/814
2011-01-27 05:17:10 -06:00
cn = keys[-1]
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
if 'rename' in options:
ipa permission-mod --rename does not work This patch fixes nonfunctional rename operation in permission plugin. Also makes sure, that no change is made to the underlying ACI in pre_callback() when the target permission already exists. Several tests for the rename operation have been created to ensure that the it won't break again unnoticed. https://fedorahosted.org/freeipa/ticket/814
2011-01-27 05:17:10 -06:00
self.api.Command.aci_mod(cn,aciprefix=ACI_PREFIX,
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
permission=options['rename'])
ipa permission-mod --rename does not work This patch fixes nonfunctional rename operation in permission plugin. Also makes sure, that no change is made to the underlying ACI in pre_callback() when the target permission already exists. Several tests for the rename operation have been created to ensure that the it won't break again unnoticed. https://fedorahosted.org/freeipa/ticket/814
2011-01-27 05:17:10 -06:00
self.api.Command.aci_rename(cn, aciprefix=ACI_PREFIX,
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
newname=options['rename'])
ipa permission-mod --rename does not work This patch fixes nonfunctional rename operation in permission plugin. Also makes sure, that no change is made to the underlying ACI in pre_callback() when the target permission already exists. Several tests for the rename operation have been created to ensure that the it won't break again unnoticed. https://fedorahosted.org/freeipa/ticket/814
2011-01-27 05:17:10 -06:00
cn = options['rename'] # rename finished
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
common_options = dict((k, options[k]) for k in ('all', 'raw') if k in options)
result = self.api.Command.permission_show(cn, **common_options)['result']
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
for r in result:
Display the value of memberOf ACIs in permission plugin. There were two problems: 1. memberof wasn't in the list of things we looked for in the return value from aci_show() 2. The value wasn't being translated into a group name. Use the DN class to retrieve the group name from the memberof URI. Note that I changed the parsing for targetgroup as well. We now save a lookup and potentially returning a NotFound if an aci points to a group that no longer exists. https://fedorahosted.org/freeipa/ticket/2100
2011-12-06 17:15:41 -06:00
if not r.startswith('member_'):
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
entry_attrs[r] = result[r]
return dn
api.register(permission_mod)
class permission_find(LDAPSearch):
ticket 1669 - improve i18n docstring extraction This patch reverts the use of pygettext for i18n string extraction. It was originally introduced because the help documentation for commands are in the class docstring and module docstring. Docstrings are a Python construct whereby any string which immediately follows a class declaration, function/method declaration or appears first in a module is taken to be the documentation for that object. Python automatically assigns that string to the __doc__ variable associated with the object. Explicitly assigning to the __doc__ variable is equivalent and permitted. We mark strings in the source for i18n translation by embedding them in _() or ngettext(). Specialized extraction tools (e.g. xgettext) scan the source code looking for strings with those markers and extracts the string for inclusion in a translation catalog. It was mistakingly assumed one could not mark for translation Python docstrings. Since some docstrings are vital for our command help system some method had to be devised to extract docstrings for the translation catalog. pygettext has the ability to locate and extract docstrings and it was introduced to acquire the documentation for our commands located in module and class docstrings. However pygettext was too large a hammer for this task, it lacked any fined grained ability to extract only the docstrings we were interested in. In practice it extracted EVERY docstring in each file it was presented with. This caused a large number strings to be extracted for translation which had no reason to be translated, the string might have been internal code documentation never meant to be seen by users. Often the superfluous docstrings were long, complex and likely difficult to translate. This placed an unnecessary burden on our volunteer translators. Instead what is needed is some method to extract only those strings intended for translation. We already have such a mechanism and it is already widely used, namely wrapping strings intended for translation in calls to _() or _negettext(), i.e. marking a string for i18n translation. Thus the solution to the docstring translation problem is to mark the docstrings exactly as we have been doing, it only requires that instead of a bare Python docstring we instead assign the marked string to the __doc__ variable. Using the hypothetical class foo as an example. class foo(Command): ''' The foo command takes out the garbage. ''' Would become: class foo(Command): __doc__ = _('The foo command takes out the garbage.') But which docstrings need to be marked for translation? The makeapi tool knows how to iterate over every command in our public API. It was extended to validate every command's documentation and report if any documentation is missing or not marked for translation. That information was then used to identify each docstring in the code which needed to be transformed. In summary what this patch does is: * Remove the use of pygettext (modification to install/po/Makefile.in) * Replace every docstring with an explicit assignment to __doc__ where the rhs of the assignment is an i18n marking function. * Single line docstrings appearing in multi-line string literals (e.g. ''' or """) were replaced with single line string literals because the multi-line literals were introducing unnecessary whitespace and newlines in the string extracted for translation. For example: ''' The foo command takes out the garbage. ''' Would appear in the translation catalog as: "\n The foo command takes out the garbage.\n " The superfluous whitespace and newlines are confusing to translators and requires us to strip leading and trailing whitespace from the translation at run time. * Import statements were moved from below the docstring to above it. This was necessary because the i18n markers are imported functions and must be available before the the doc is parsed. Technically only the import of the i18n markers had to appear before the doc but stylistically it's better to keep all the imports together. * It was observed during the docstring editing process that the command documentation was inconsistent with respect to the use of periods to terminate a sentence. Some doc had a trailing period, others didn't. Consistency was enforced by adding a period to end of every docstring if one was missing.
2011-08-24 21:48:30 -05:00
__doc__ = _('Search for permissions.')
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
msg_summary = ngettext(
Fix translatable strings in ipalib plugins. Needed for xgettext/pygettext processing.
2011-02-23 15:47:49 -06:00
'%(count)d permission matched', '%(count)d permissions matched', 0
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
)
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
has_output_params = LDAPSearch.has_output_params + output_params
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
def post_callback(self, ldap, entries, truncated, *args, **options):
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
if options.pop('pkey_only', False):
Create pkey-only option for find commands New option --pkey-only is available for all LDAPSearch based classes with primary key visible in the output. This option makes LDAPSearch commands search for primary attribute only. This may be useful when manipulating large data sets. User can at first retrieve all primary keys in a relatively small data package and then run further commands with retrieved primary keys. https://fedorahosted.org/freeipa/ticket/1262
2011-10-26 04:12:38 -05:00
return
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
for entry in entries:
(dn, attrs) = entry
try:
Fix raw format for ACI commands ACI plugins (permission, selfservice and delegation) were not prepared to serve ACIs in a raw format, i.e. raw "aci" attribute taken from LDAP. This patch fixes all these plugins and their commands to provide provide this format. Few ACI raw format unit tests were added for all these plugins. https://fedorahosted.org/freeipa/ticket/2010 https://fedorahosted.org/freeipa/ticket/2223 https://fedorahosted.org/freeipa/ticket/2228 https://fedorahosted.org/freeipa/ticket/2232
2012-02-02 14:28:15 -06:00
aci = self.api.Command.aci_show(attrs['cn'][0], aciprefix=ACI_PREFIX, **options)['result']
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
# copy information from respective ACI to permission entry
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
for attr in self.obj.aci_attributes:
if attr in aci:
attrs[attr] = aci[attr]
except errors.NotFound:
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
self.debug('ACI not found for %s' % attrs['cn'][0])
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
# Now find all the ACIs that match. Once we find them, add any that
# aren't already in the list along with their permission info.
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
Fix raw format for ACI commands ACI plugins (permission, selfservice and delegation) were not prepared to serve ACIs in a raw format, i.e. raw "aci" attribute taken from LDAP. This patch fixes all these plugins and their commands to provide provide this format. Few ACI raw format unit tests were added for all these plugins. https://fedorahosted.org/freeipa/ticket/2010 https://fedorahosted.org/freeipa/ticket/2223 https://fedorahosted.org/freeipa/ticket/2228 https://fedorahosted.org/freeipa/ticket/2232
2012-02-02 14:28:15 -06:00
opts = copy.copy(options)
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
opts['aciprefix'] = ACI_PREFIX
Fix raw format for ACI commands ACI plugins (permission, selfservice and delegation) were not prepared to serve ACIs in a raw format, i.e. raw "aci" attribute taken from LDAP. This patch fixes all these plugins and their commands to provide provide this format. Few ACI raw format unit tests were added for all these plugins. https://fedorahosted.org/freeipa/ticket/2010 https://fedorahosted.org/freeipa/ticket/2223 https://fedorahosted.org/freeipa/ticket/2228 https://fedorahosted.org/freeipa/ticket/2232
2012-02-02 14:28:15 -06:00
try:
# permission ACI attribute is needed
del opts['raw']
except:
pass
Fix overlapping cn param/option issue, pass cn as aciname in find permission-find --name wasn't working for two reasons. The first was that the cn to search on in options ended up overlapping the primary key name causing the request to fail. The second reason was aci uses aciname, not cn, as its name field. So searching on --name matched everything because it was as if you were searching on nothing. https://fedorahosted.org/freeipa/ticket/2320
2012-05-11 14:28:03 -05:00
if 'cn' in options:
# the attribute for name is difference in acis
opts['aciname'] = options['cn']
Fix raw format for ACI commands ACI plugins (permission, selfservice and delegation) were not prepared to serve ACIs in a raw format, i.e. raw "aci" attribute taken from LDAP. This patch fixes all these plugins and their commands to provide provide this format. Few ACI raw format unit tests were added for all these plugins. https://fedorahosted.org/freeipa/ticket/2010 https://fedorahosted.org/freeipa/ticket/2223 https://fedorahosted.org/freeipa/ticket/2228 https://fedorahosted.org/freeipa/ticket/2232
2012-02-02 14:28:15 -06:00
aciresults = self.api.Command.aci_find(*args, **opts)
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
truncated = truncated or aciresults['truncated']
results = aciresults['result']
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
Fix overlapping cn param/option issue, pass cn as aciname in find permission-find --name wasn't working for two reasons. The first was that the cn to search on in options ended up overlapping the primary key name causing the request to fail. The second reason was aci uses aciname, not cn, as its name field. So searching on --name matched everything because it was as if you were searching on nothing. https://fedorahosted.org/freeipa/ticket/2320
2012-05-11 14:28:03 -05:00
if 'cn' in options:
# there is an option/param overlap if --name is in the
# search list, we don't need cn anymore so drop it.
options.pop('cn')
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
for aci in results:
found = False
if 'permission' in aci:
for entry in entries:
Add plugin for manage self-service ACIs This is just a thin wrapper around the aci plugin, controlling what types of ACIs can be added. Right now only ACIs in the basedn can be managed with this plugin. ticket 531
2010-12-08 12:33:40 -06:00
(dn, attrs) = entry
ACI plugin supports prefixes When more than one plugin produce ACIs, they share common namespace of ACI name. This may lead to name collisions between the ACIs from different plugins. This patch introduces a mandatory "prefix" attribute for non-find ACI operations which allow plugins to use their own prefixes (i.e. namespaces) which is then used when a name of the ACI is generated. Permission, Delegation and Selfservice plugins has been updated to use their own prefixes thus avoiding name collisions by using their own namespaces. Default ACIs in LDIFs has been updated to follow this new policy. Permission plugin now uses its CN (=primary key) instead of description in ACI names as Description may not be unique. This change requires an IPA server reinstall since the default ACI set has been changed. https://fedorahosted.org/freeipa/ticket/764
2011-01-21 02:20:01 -06:00
if aci['permission'] == attrs['cn'][0]:
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
found = True
break
Catch ACI errors better when adding a permission. We create the aci with the --test flag to test its validity but it doesn't do the same level of tests that actually adding an aci to LDAP does. Catch any syntax errors that get thrown and clean up as best we can. ticket 621
2010-12-17 14:35:24 -06:00
if not found:
Fix raw format for ACI commands ACI plugins (permission, selfservice and delegation) were not prepared to serve ACIs in a raw format, i.e. raw "aci" attribute taken from LDAP. This patch fixes all these plugins and their commands to provide provide this format. Few ACI raw format unit tests were added for all these plugins. https://fedorahosted.org/freeipa/ticket/2010 https://fedorahosted.org/freeipa/ticket/2223 https://fedorahosted.org/freeipa/ticket/2228 https://fedorahosted.org/freeipa/ticket/2232
2012-02-02 14:28:15 -06:00
permission = self.api.Command.permission_show(aci['permission'], **options)['result']
dn = permission['dn']
del permission['dn']
if (dn, permission) not in entries:
entries.append((dn, permission))
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
api.register(permission_find)
class permission_show(LDAPRetrieve):
ticket 1669 - improve i18n docstring extraction This patch reverts the use of pygettext for i18n string extraction. It was originally introduced because the help documentation for commands are in the class docstring and module docstring. Docstrings are a Python construct whereby any string which immediately follows a class declaration, function/method declaration or appears first in a module is taken to be the documentation for that object. Python automatically assigns that string to the __doc__ variable associated with the object. Explicitly assigning to the __doc__ variable is equivalent and permitted. We mark strings in the source for i18n translation by embedding them in _() or ngettext(). Specialized extraction tools (e.g. xgettext) scan the source code looking for strings with those markers and extracts the string for inclusion in a translation catalog. It was mistakingly assumed one could not mark for translation Python docstrings. Since some docstrings are vital for our command help system some method had to be devised to extract docstrings for the translation catalog. pygettext has the ability to locate and extract docstrings and it was introduced to acquire the documentation for our commands located in module and class docstrings. However pygettext was too large a hammer for this task, it lacked any fined grained ability to extract only the docstrings we were interested in. In practice it extracted EVERY docstring in each file it was presented with. This caused a large number strings to be extracted for translation which had no reason to be translated, the string might have been internal code documentation never meant to be seen by users. Often the superfluous docstrings were long, complex and likely difficult to translate. This placed an unnecessary burden on our volunteer translators. Instead what is needed is some method to extract only those strings intended for translation. We already have such a mechanism and it is already widely used, namely wrapping strings intended for translation in calls to _() or _negettext(), i.e. marking a string for i18n translation. Thus the solution to the docstring translation problem is to mark the docstrings exactly as we have been doing, it only requires that instead of a bare Python docstring we instead assign the marked string to the __doc__ variable. Using the hypothetical class foo as an example. class foo(Command): ''' The foo command takes out the garbage. ''' Would become: class foo(Command): __doc__ = _('The foo command takes out the garbage.') But which docstrings need to be marked for translation? The makeapi tool knows how to iterate over every command in our public API. It was extended to validate every command's documentation and report if any documentation is missing or not marked for translation. That information was then used to identify each docstring in the code which needed to be transformed. In summary what this patch does is: * Remove the use of pygettext (modification to install/po/Makefile.in) * Replace every docstring with an explicit assignment to __doc__ where the rhs of the assignment is an i18n marking function. * Single line docstrings appearing in multi-line string literals (e.g. ''' or """) were replaced with single line string literals because the multi-line literals were introducing unnecessary whitespace and newlines in the string extracted for translation. For example: ''' The foo command takes out the garbage. ''' Would appear in the translation catalog as: "\n The foo command takes out the garbage.\n " The superfluous whitespace and newlines are confusing to translators and requires us to strip leading and trailing whitespace from the translation at run time. * Import statements were moved from below the docstring to above it. This was necessary because the i18n markers are imported functions and must be available before the the doc is parsed. Technically only the import of the i18n markers had to appear before the doc but stylistically it's better to keep all the imports together. * It was observed during the docstring editing process that the command documentation was inconsistent with respect to the use of periods to terminate a sentence. Some doc had a trailing period, others didn't. Consistency was enforced by adding a period to end of every docstring if one was missing.
2011-08-24 21:48:30 -05:00
__doc__ = _('Display information about a permission.')
Add new schema to store information about permissions. There are some permissions we can't display because they are stored outside of the basedn (such as the replication permissions). We are adding a new attribute to store extra information to make this clear, in this case SYSTEM. ticket 853
2011-02-01 10:57:18 -06:00
has_output_params = LDAPRetrieve.has_output_params + output_params
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
def post_callback(self, ldap, dn, entry_attrs, *keys, **options):
try:
Do not use extra command options in ACI, permission, selfservice Allowing Commands to be called with ignored unknown options opens the door to problems, for example with misspelled option names. Before we start rejecting them, we need to make sure IPA itself does not use them when it calls commands internally. This patch does that for ACI-related plugins. Part of the work for https://fedorahosted.org/freeipa/ticket/2509
2012-04-17 11:48:33 -05:00
common_options = dict((k, options[k]) for k in ('all', 'raw') if k in options)
aci = self.api.Command.aci_show(keys[-1], aciprefix=ACI_PREFIX, **common_options)['result']
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
for attr in self.obj.aci_attributes:
if attr in aci:
entry_attrs[attr] = aci[attr]
except errors.NotFound:
Rename permissions and privileges to be more readable. This also drops description from permissions since it seems redundant and fixes up the help text a little. ticket 792
2011-01-31 10:01:56 -06:00
self.debug('ACI not found for %s' % entry_attrs['cn'][0])
Provide attributelevelrights for the aci components in permission_show. Since the broken-out components are just part of the aci just copy right access rights for aci. ticket 943
2011-04-20 12:26:20 -05:00
if options.get('rights', False) and options.get('all', False):
# The ACI attributes are just broken-out components of aci so
# the rights should all match it.
for attr in self.obj.aci_attributes:
entry_attrs['attributelevelrights'][attr] = entry_attrs['attributelevelrights']['aci']
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
return dn
api.register(permission_show)
class permission_add_member(LDAPAddMember):
"""
Add members to a permission.
"""
Rename INTERNAL to NO_CLI for commands we hide from the cli. Also make i18n_messages and json_metadata NO_CLI. ticket 821
2011-01-20 14:07:43 -06:00
NO_CLI = True
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
api.register(permission_add_member)
class permission_remove_member(LDAPRemoveMember):
"""
Remove members from a permission.
"""
Rename INTERNAL to NO_CLI for commands we hide from the cli. Also make i18n_messages and json_metadata NO_CLI. ticket 821
2011-01-20 14:07:43 -06:00
NO_CLI = True
Re-implement access control using an updated model. The new model is based on permssions, privileges and roles. Most importantly it corrects the reverse membership that caused problems in the previous implementation. You add permission to privileges and privileges to roles, not the other way around (even though it works that way behind the scenes). A permission object is a combination of a simple group and an aci. The linkage between the aci and the permission is the description of the permission. This shows as the name/description of the aci. ldap:///self and groups granting groups (v1-style) are not supported by this model (it will be provided separately). This makes the aci plugin internal only. ticket 445
2010-12-01 10:23:52 -06:00
api.register(permission_remove_member)
Reference in New Issue Copy Permalink