summaryrefslogtreecommitdiffstats
path: root/spec/unit/network/client
Commit message (Collapse)AuthorAgeFilesLines
* Fix tests on #3347Jesse Wolfe2010-02-171-110/+0
|
* Feature #3347 REST-ified FileBucketJesse Wolfe2010-02-171-4/+98
| | | | | | | | | | | | | | | | FileBucket Files have been reimplemented as an indirector terminus so that they can be transmitted over REST. The old Network::Client.dipper has been replaced with a compatibility later in FileBucket::Dipper that uses the indirector to access filebucket termini. Slightly revised patch: * No longer allows nil contents in FileBucket outside of initialization * Uses File.exist? instead of the deprecated File.exists? * Tweaks JSON serialization and de-serialization to include "path" Deferred issues: * Feature #3371 "FileBucket should not keep files in memory". * Feature #3372 "Replace FileBucket Dipper with more idiomatic calls"
* Fixing #2094 - filebucket failures are clearer nowLuke Kanies2009-06-161-0/+16
| | | | | | We just add a bit of information to the exception. Signed-off-by: Luke Kanies <luke@madstop.com>
* Renmaing Puppet::Network::Client::Master to Puppet::AgentLuke Kanies2009-02-061-442/+0
| | | | | | | | | Made minor changes, including removing the parent class. The functionality hasn't changed yet -- that comes in later patches -- but all but a couple of the older tests pass. Signed-off-by: Luke Kanies <luke@madstop.com>
* Fixing #1749 - Splay now hopefully behaves "better" for small values.Luke Kanies2008-12-011-9/+9
| | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* Only apply splay the first runAndrew Shafer2008-09-211-0/+42
| | | | Issue 1491
* You can now select the encoding format when transferring the catalog,Luke Kanies2008-07-181-19/+77
| | | | | | | | | | with 'yaml' still being the default but 'marshal' being an option. This is because testing has shown drastic performance differences between the two, with up to 70% of compile time being spent in YAML code. Use the 'catalog_format' setting to choose your format, and the setting must be set on the client. Signed-off-by: Luke Kanies <luke@madstop.com>
* Fixing tests that are broken when running as root under OSX 10.5Paul Lathrop2008-05-231-0/+2
|
* Removed the code from the client that tries to avoid recompilingLuke Kanies2008-04-101-25/+0
| | | | | | the catalog. The client will now always recompile, assuming it can reach the server. It will still use the cached config if there's a failure.
* Renaming 'configuration' to 'catalog', fixing #954.Luke Kanies2007-12-111-38/+38
|
* Reverting the changes I'd made toward removing the globalLuke Kanies2007-11-191-112/+1
| | | | | | | | resources. These are commits: c19835ce9f8a5138b30a1a32ca741c996b0916d2 9290cc89a2206fb5204578f8e91208857a48b147 ffb4c2dbc7314b364d25e4f7be599ef05b767b44
* Fixed most failing tests, but there are still over thirty failing.Luke Kanies2007-11-181-1/+22
| | | | | | At this point, I'm holding the experiment until after the release, so I'm committing this for now and will take it back up after 0.24.0 is out.
* Modifying how default resources are created; they are nowLuke Kanies2007-11-151-4/+17
| | | | | added to the configuration by the master client, rather than by the creating types.
* This commit is the first run at removing all globalLuke Kanies2007-11-131-1/+78
| | | | | | | | | | references to resources. It deprecates the class-level [] and []= methods, used for so long to provide closure behaviour but now unnecessary with the node configuration's ability to function as a resource container. All of the spec tests pass, but there is much to do to make the test/ tests pass, I expect.
* Fixing #800 by refactoring how configurations are retrievedLuke Kanies2007-11-121-0/+365
from the server. The real problem was getting all of the validation done before any caching, which required a good bit more refactoring than I expected. In actuality, this commit is relatively small even though it covers many files; most of the changes just make the code clearer or shorter.