Hi Brad

Thanks for adding the samples, I'm sure they will help people understand how to use launchapd lib. Your XSLT rule is broken; I suggest a fix.

> === modified file 'launchpadlib/wadl-to-refhtml.xsl'
> --- launchpadlib/wadl-to-refhtml.xsl	2009-01-13 17:30:59 +0000
> +++ launchpadlib/wadl-to-refhtml.xsl	2009-01-15 22:34:48 +0000
> @@ -379,6 +379,12 @@
>                  <xsl:text>/+wikiname/</xsl:text>
>                  <var>&lt;id&gt;</var>
>              </xsl:when>
> +            <xsl:when test="@id = 'commercial_subscription'">
> +                <xsl:text>/</xsl:text>
> +                <xsl:text>/+commercialsubscription/</xsl:text>
> +                <xsl:text>/</xsl:text>
> +                <var>&lt;commercial_subscription.id&gt;</var>
> +            </xsl:when>

This instruction builds a path by concatenating text, and it does not look
correct. I see

    //+commercialsubscription//<commercial_subscription.id>

/me looks at source
Yep, I am right. Since the convention is to have wrap the view name in
"/"s, this markup should look like the h_w_device rules

            <xsl:when test="@id = 'commercial_subscription'">
                <xsl:text>/+commercialsubscription/</xsl:text>
                <var>&lt;commercial_subscription.id&gt;</var>
            </xsl:when>

> === added directory 'samples'
> === added file 'samples/_pythonpath.py'
> --- samples/_pythonpath.py	1970-01-01 00:00:00 +0000
> +++ samples/_pythonpath.py	2009-01-15 22:34:48 +0000
> @@ -0,0 +1,14 @@
> +__metaclass__ = type
> +
> +import sys, os, os.path

Put each import on its own line.

> +
> +# See if we can find launchpadlib and waddlib
> +try:
> +    import launchpadlib
> +    import wadllib
> +except ImportError:
> +    # Modify sys.path to add trunk/lib.
> +    HOME = os.environ['HOME']
> +    trunk_path = os.environ.get('LP_TRUNK_PATH',
> +                                os.path.join(HOME, 'canonical/lp-branches/trunk'))

Wrap the code at 78 characters.

...

> === added file 'samples/lpapi.py'
> --- samples/lpapi.py	1970-01-01 00:00:00 +0000
> +++ samples/lpapi.py	2009-01-15 22:34:48 +0000
> @@ -0,0 +1,108 @@
> +#!/usr/bin/python2.4
> +import os
> +import sys
> +from urlparse import urljoin
> +import commands
> +
> +try:
> +    from launchpadlib.launchpad import (
> +        Launchpad, EDGE_SERVICE_ROOT, STAGING_SERVICE_ROOT)
> +    from launchpadlib.credentials import Credentials
> +    from launchpadlib.errors import *
> +    import launchpadlib
> +except ImportError:
> +    print >> sys.stderr, "Usage:"
> +    print >> sys.stderr, "   PYTHONPATH=somebranch/lib %s" % sys.argv[0]
> +    raise
> +
> +
> +class InvalidCredentials(Exception):
> +    pass
> +
> +
> +class Retry(Exception):
> +    pass
> +
> +
> +class LPSystem:

Can you add docstrings to these. This is a great base class and exceptions
that I expect other people will use and evolve.'


> +    endpoint = None
> +    auth_file = None
> +    def __init__(self, app_name='just testing', use_cache=True):
> +        home = os.environ['HOME']
> +        if use_cache:
> +            cachedir = os.path.join(home, '.launchpadlib', 'cache')
> +            if not os.path.exists(cachedir):
> +                cachedir = None
> +        else:
> +            cachedir = None

Is ``cachedir`` right? ``cache_dir`` looks PEP-8 friendly.

> +        try:
> +            # Load credentials from AUTH_FILE if it exists.
> +            self.auth_file = os.path.join(home, self.auth_file_name)
> +            self.credentials = Credentials()
> +            self.credentials.load(open(self.auth_file))
> +            print >> sys.stderr, "Loading credentials..."
> +            try:
> +                self.launchpad = Launchpad(self.credentials, self.endpoint,
> +                                           cache=cachedir)
> +            except launchpadlib.errors.HTTPError:
> +                raise InvalidCredentials, (
> +                    "Please remove %s and rerun %s to authenticate." % (
> +                    self.auth_file, sys.argv[0]))
> +        except IOError:
> +            # Prompt for authentication process, then save credentials to AUTH_FILE.

Wrap the code at 78 characters.

> +            try:
> +                self.launchpad = Launchpad.get_token_and_login(app_name,
> +                                                               self.endpoint,
> +                                                               cache=cachedir)
> +                self.launchpad.credentials.save(open(self.auth_file, "w"))
> +                print >> sys.stderr, "Credentials saved"
> +            except launchpadlib.errors.HTTPError, err:
> +                print >> sys.stderr, err.content
> +                raise
> +
> +    @property
> +    def url(self):
> +        return urljoin(self.endpoint, 'projects/')
> +
> +
> +class Edge(LPSystem):
> +    endpoint = EDGE_SERVICE_ROOT
> +    auth_file_name = 'edge.auth'
> +
> +
> +class Staging(LPSystem):
> +    endpoint = STAGING_SERVICE_ROOT
> +    auth_file_name = 'staging.auth'
> +
> +
> +class Production(LPSystem):
> +    endpoint = 'https://api.launchpad.net/beta/'
> +    auth_file_name = 'production.auth'
> +
> +
> +class Dev(LPSystem):
> +    endpoint = 'https://api.launchpad.dev/beta/'
> +    auth_file_name = 'dev.auth'
> +    def __init__(self, app_name='just testing'):
> +        LPSystem.__init__(self, app_name=app_name, use_cache=False)
> +
> +
> +systems=dict(dev=Dev,
> +             staging=Staging,
> +             edge=Edge,
> +             beta=Edge,
> +             production=Production,
> +             prod=Production,
> +             )

Wrap the "=" in spaces:
systems = dict(dev=Dev,

Is beta vestigial? Is it used>

> +
> +
> +def lp_factory(sysname, app_name='just_testing'):

I think system_name is clearer

> +    """Get a launchpad API instance for `sysname`."""

I believe double-backquotes denotes a variable in ReST. single-backquotes
is a cross-reference.

...

=== added file 'samples/nopriv-api.py'
--- samples/nopriv-api.py	1970-01-01 00:00:00 +0000
+++ samples/nopriv-api.py	2009-01-15 22:34:48 +0000
@@ -0,0 +1,91 @@
+#!/usr/bin/python
+# -*-doctest-*-
+
+"""
+    >>> import _pythonpath, lpapi
+    >>> lp = lpapi.lp_factory('dev')
+
+    >>> bzr = lp.projects['bzr']
+    >>> print bzr.reviewer_whiteboard
+    tag:launchpad.net:2008:redacted
+    >>> bzr.reviewer_whiteboard = "Check on licensing"
+    >>> print bzr.reviewer_whiteboard
+    Check on licensing
+    >>> bzr.lp_save()
+    ...
+    Traceback (most recent call last):
+    ...
+    HTTPError: HTTP Error 401: Unauthorized
+
+    >>> from operator import attrgetter
+    >>> def print_projs(projs):
+    ...     for p in sorted(projs, key=attrgetter('name')):
+    ...         print p.name

Launchpad style to to never use single letters as variables (except for i
and j).

    >>> def print_projects(projects):
    ...     for project in sorted(projects, key=attrgetter('name')):
    ...         print project.name

This doesn't look like it is used.

> +
> +    >>> inactive = lp.projects.licensing_search(active=False)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> active = lp.projects.licensing_search(active=True)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(license_reviewed=False)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(has_zero_licenses=True)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(licenses=['Other/Proprietary'])
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> bzr.licenses
> +    []
> +
> +    >>> projs = lp.projects.licensing_search(has_zero_licenses=False)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(licenses=['GNU Affero GPL v3'])
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(subscription_expires_after="2005-01-01")

Wrap the code at 78 characters.

> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(has_zero_licenses=True, search_text="Bazaar")

Wrap the code at 78 characters.

> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +"""
> +
> +if __name__ == '__main__':
> +    # Remove existing credentials.
> +    import os
> +    try:
> +        os.unlink(os.path.join(os.environ['HOME'], 'dev.auth'))
> +    except OSError:
> +        pass
> +
> +    # Create correct credentials.
> +    print "Login as '<email address hidden>' in your browser."
> +    print "Press <Enter> when done."
> +    raw_input()
> +
> +    import _pythonpath, lpapi

I think a judicious comment is need to explain that those two imported
modules setup the scripting environment. It doesn't look like this uses
_pythonpath, but we know it does.

...

> === added file 'samples/sample-person-api.py'
> --- samples/sample-person-api.py	1970-01-01 00:00:00 +0000
> +++ samples/sample-person-api.py	2009-01-15 22:34:48 +0000
> @@ -0,0 +1,91 @@
> +#!/usr/bin/python
> +# -*-doctest-*-
> +
> +"""
> +    >>> import _pythonpath, lpapi
> +    >>> lp = lpapi.lp_factory('dev')
> +
> +    >>> bzr = lp.projects['bzr']
> +    >>> print bzr.reviewer_whiteboard
> +    tag:launchpad.net:2008:redacted
> +    >>> bzr.reviewer_whiteboard = "Check on licensing"
> +    >>> print bzr.reviewer_whiteboard
> +    Check on licensing
> +    >>> bzr.lp_save()
> +    ...
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> from operator import attrgetter
> +    >>> def print_projs(projs):
> +    ...     for p in sorted(projs, key=attrgetter('name')):
> +    ...         print p.name

This looks like the helper from the previous test. It should be the the
same as I suggested above, or extracted to a testing module that can
be imported. Oh and this too does not look like it is used.

> +
> +    >>> inactive = lp.projects.licensing_search(active=False)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> active = lp.projects.licensing_search(active=True)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(license_reviewed=False)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(has_zero_licenses=True)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(licenses=['Other/Proprietary'])
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> bzr.licenses
> +    []
> +
> +    >>> projs = lp.projects.licensing_search(has_zero_licenses=False)
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(licenses=['GNU Affero GPL v3'])
> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(subscription_expires_after="2005-01-01")

Wrap the code at 78 characters.

> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +    >>> projs = lp.projects.licensing_search(has_zero_licenses=True, search_text="Bazaar")

Wrap the code at 78 characters.

> +    Traceback (most recent call last):
> +    ...
> +    HTTPError: HTTP Error 401: Unauthorized
> +
> +"""
> +
> +if __name__ == '__main__':
> +    # Remove existing credentials.
> +    import os
> +    try:
> +        os.unlink(os.path.join(os.environ['HOME'], 'dev.auth'))
> +    except OSError:
> +        pass
> +
> +    # Create correct credentials.
> +    print "Login as '<email address hidden>' in your browser."
> +    print "Press <Enter> when done."
> +    raw_input()
> +
> +    import _pythonpath, lpapi

This too is doing something magical that a good comment would elucidate.

...

> === modified file 'setup.py'
> --- setup.py	2008-08-01 16:05:31 +0000
> +++ setup.py	2009-01-15 22:34:48 +0000

...