summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* | Fix #2107 - flatten resource references arrays properlyBrice Figureau2009-03-272-0/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Resource parameters of the form [Res[a], Res[a,b]] ends being evaluated as [Res[a], [ Res[a], Res[b] ] This last form was not flattened when transfomed into RAL type, which in turn prevented the sub array to be converted in regular resource references. Thus the type was choking when encountering those native parser references instead of usual resource reference. The fix consists in flattening array of references before transformation to trans objects for RAL. Signed-off-by: Brice Figureau <brice-puppet@daysofwonder.com>
* | Fix #2101 - Return to recurse=0 == no recursion behaviorBrice Figureau2009-03-271-1/+10
| | | | | | | | | | | | | | | | | | Ticket #1469 introduced an incorrect change of behaviour where recurse = 0 was considered as an infinite recursion, unlike before. This patch restores this behavior, and makes some test pass. Signed-off-by: Brice Figureau <brice-puppet@daysofwonder.com>
* | Fix #2101 - fix failing testBrice Figureau2009-03-271-1/+5
| | | | | | | | | | | | | | | | This code hasn't been modified since the introduction of the fileset recurselimit parameter. Tests depending on this code were failing. Signed-off-by: Brice Figureau <brice-puppet@daysofwonder.com>
* | Fix #2101 - fix recurselimit == 0 bad behaviourBrice Figureau2009-03-272-5/+5
| | | | | | | | | | | | | | After the fix for #1469, recurselimit = 0 was considered as an infinite recursion which is the reverse of what it was before. Signed-off-by: Brice Figureau <brice-puppet@daysofwonder.com>
* | Merge branch 'tickets/master/2110'James Turnbull2009-03-271-1/+1
|\ \
| * | Fixed #2110 - versioncmp brokenJames Turnbull2009-03-261-1/+1
| | |
* | | Added rake ci:all taskJames Turnbull2009-03-261-0/+3
|/ /
* | CHANGELOG updatesJames Turnbull2009-03-211-0/+4
| |
* | Moved of puppetd, puppetca, puppetmasterd, puppetrun binary from bin to sbinJames Turnbull2009-03-215-4/+12
| |
* | Fixed #2086 - Fixes to make building tarballs easierJames Turnbull2009-03-211-2/+4
| |
* | Fix #1469 - Add an option to recurse only on remote sideBrice Figureau2009-03-205-45/+107
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When using recurse and a source, if the client side has many files it can take a lot of CPU/memory to checksum the whole client hierarchy. The idea is that it is not necessary to recurse on the client side if all we want is to manage the files that are sourced from the server. This changeset adds the "remote" recurse value which prevents recursing on the client side when a source is present. Since it also is necessary to limit the remote side recursion a new File{} parameter has been added called "recurselimit". Moreover, the Filetset API is changing to allow the new recurselimit parameter, and passing the recursion depth limit in the recurse parameter as an integer is now deprecated and not supported anymore. Signed-off-by: Brice Figureau <brice-puppet@daysofwonder.com>
* | Forbidding REST clients to set the node or IPLuke Kanies2009-03-203-1/+27
| | | | | | | | | | | | | | | | This is done for security reasons - if a client is unauthenticated, we don't want them to be able to just configure their own authentication information. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Fixing #1557 - Environments are now in REST URIsLuke Kanies2009-03-2022-185/+88
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit includes multiple, related changes, all in one commit because the whole thing was necessary to reach a functional tree again: * The URI starts with the environment, so: /production/certificate/foo /development/file_content/path/to/your/file * All REST handling is done by a single instance mounted at / for webrick and Mongrel, rather than having individual instances mounted at, say, /certificate. * All REST URI translation is done by an API module. Currently only the 'v1' module exists with no support for additional modules, but it's well-separated and will be easy to expand as we need it. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Adding explicit optional attribute to indirection requestsLuke Kanies2009-03-202-9/+41
| | | | | | | | | | | | | | | | | | | | | | | | | | Previously, any option that had a setter was treated as an attribute, but now we're specifying the list of attributes settable via options. We also have a to_hash method that will take all of the options and all of those attributes and join them back into a hash. This method is used by the REST Handler module, since it uses the indirection request internally. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Removing an unused source fileLuke Kanies2009-03-201-59/+0
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Adding a model accessor to the Request classLuke Kanies2009-03-202-4/+27
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Requests now use default environment when none is specifiedLuke Kanies2009-03-202-1/+12
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Moving the REST API functions into a moduleLuke Kanies2009-03-207-169/+197
| | | | | | | | | | | | | | | | | | | | | | | | | | This module is now used by the client and server side, rather than having a Handler module that's 90% server functionality but also used by the client. While we don't automatically get api choice from this, it at least provides a pattern for how we'll handle API development over time. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Using the Handler for the REST api on both sides of the connectionLuke Kanies2009-03-204-296/+124
| | | | | | | | | | | | | | | | | | | | Things are actually in a broken state here because we've got a conflict between how the two sides do their work and some extraction needs to get done. This commit is just a stopping-point so I can do the necessary refactoring. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Adding REST::Handler methods for converting between indirection and urisLuke Kanies2009-03-204-0/+155
| | | | | | | | | | | | | | | | | | | | This is the first main phase to having a common module for handling the REST api - this Handler module will be included by all of the web server REST modules and the Indirector Request class, so there's a common place that understands how the URI consists. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Adding environment support to the REST URILuke Kanies2009-03-204-17/+56
| | | | | | | | | | | | Also adding it to the Indirection Request. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Moving the query_string method to RequestLuke Kanies2009-03-204-100/+104
| | | | | | | | | | | | | | | | It required a request instance and didn't use the REST class it was in, so it makes more sense in the Request class. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Adding tests for the REST query string usageLuke Kanies2009-03-201-5/+21
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Fixing #2108 - pi should work againLuke Kanies2009-03-201-6/+8
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Fix #1088 - part2 - Add rspec testsBrice Figureau2009-03-144-37/+216
| | | | | | | | Signed-off-by: Brice Figureau <brice-puppet@daysofwonder.com>
* | Fix #1088 - Collections overridesBrice Figureau2009-03-144-721/+860
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This changeset defines a new syntax to override collection of resources (virtual or not). This feature is not constrained to the override in inherited context as usual resource override. The collection of resource supports a query like regular collection of virtual or exported resources. Usage example: file { "/tmp/testing": content => "whatever" } File<| |> { mode => 0600 } It also introduces a different behaviour for collection of catalog resources. Before this patch, only virtual resources were collected, now all resources (virtual or no) are collected and can be overriden. That means it is now possible to do: File <| |> { mode => 0600 } And all the Files resources will have mode 0600. It is then possible to have this puppet pattern: file { "/tmp/a": content => "a" } file { "/tmp/b": content => "b" } File <| title != "/tmp/a" |> { require => File["/tmp/b"] } which means that every File requires a file. Moreover it is now possible to define resource overriding without respecting the override on inheritance rule: class a { file { "/tmp/testing": content => "whatever" } } class b { include a File<| |> { mode => 0600 } } include b Signed-off-by: Brice Figureau <brice-puppet@daysofwonder.com>
* | Fixed #2071 - Updated LDAP schemaJames Turnbull2009-03-142-1/+4
| |
* | Fixing tests broken in previous commitsLuke Kanies2009-03-112-14/+5
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Fixing #1949 - relationships now use attributes instead of a labelLuke Kanies2009-03-104-69/+57
| | | | | | | | | | | | | | | | | | | | | | This was important because the use of the label to store attributes was a holdover from the GRATR library, and if we didn't cease its use before we switched to RESTful catalogs, then we'd be stuck with the @label instance variable forever, essentially. Now we can add and remove variables however we please. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Correctly handling numerical REST argumentsLuke Kanies2009-03-104-0/+25
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Not passing file sources on to child filesLuke Kanies2009-03-102-2/+2
| | | | | | | | | | | | | | This was sometimes causing files to get converted to directories. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Correctly handling non-string checksumsLuke Kanies2009-03-102-1/+16
| | | | | | | | | | | | | | | | | | | | | | | | Content is now returning the checksum rather than the actual content, and the method of creating the full checksum wasn't correctly handling timestamps, which aren't strings and can't be the right side of a String + call. I've opened #2064 as a better long-term fix. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Removing unnecessary calls to expire()Luke Kanies2009-03-102-2/+0
| | | | | | | | | | | | | | | | | | These calls were resulting in 1-2 extra round trips per file. For simple comparison, 200 empty files took 73s before this change and 20s after. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Passing checksums around instead of file contentsLuke Kanies2009-03-054-73/+158
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This switches the file's 'content' parameter to always use checksums, rather than always using content but switching to checksums whenever necessary. This greatly simplifies all the logging requirements (so that content doesn't show up in logs), but also simplifies insync comparisons, and much more. In the process, I found that the code was pulling down file content more often than was necessary, and fixing that cut 40% off of the time of a very small transaction. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Moving default fileserving mount creation to the Configuration classLuke Kanies2009-03-054-20/+24
| | | | | | | | | | | | | | | | | | | | It was previously in the parser, but the parser is only created if the fileserver.conf exists, so the default mounts weren't made if the file didn't exist. This is a bit less encapsulation, but not much. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Fixing #2028 - Better failures when a cert is found with no keyLuke Kanies2009-02-282-2/+91
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The problem was that the server had a certificate for the client. Initially the client just didn't have a key, because it assumed that if it had a certificate then it had a key. Upon fixing it to create the key, the key then did not match the found certificate. This commit fixes both of those: The key is always found before the certificate, and when the certificate is found it's verified against the private key and an exception is thrown if they don't match. It's always a failure, so this just makes the failure more informative. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Moving the clientyamldir setting into the puppetd sectionLuke Kanies2009-02-281-2/+2
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Fixed #1849 - Ruby 1.9 portability: `when' doesn't like colons, replace with ↵James Turnbull2009-02-2676-254/+256
| | | | | | | | semicolons
* | Fixed metaparameter reference to return strJames Turnbull2009-02-261-1/+11
| |
* | Merge branch 'tickets/master/2016'James Turnbull2009-02-242-36/+38
|\ \
| * | Fixed #2016 - Split metaparameters from types in reference documentationJames Turnbull2009-02-242-36/+38
| | |
* | | Fixed #2017 - incorrect requireJames Turnbull2009-02-241-1/+1
|/ /
* | Fixing #1904 - aliases are no longer inherited by child filesLuke Kanies2009-02-202-3/+3
| | | | | | | | Signed-off-by: Luke Kanies <luke@madstop.com>
* | Merge branch '0.24.x'Luke Kanies2009-02-203-12/+38
|\| | | | | | | | | | | Conflicts: lib/puppet/type/file/content.rb spec/unit/type/file/content.rb
| * Fixing #1871 once and for all - contents are never printedLuke Kanies2009-02-183-8/+43
| | | | | | | | | | | | | | | | | | They were still being printed in noop mode. The fix was to use is_to_s and should_to_s methods, rather than 'change_to_s'. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Removing or fixing old testsLuke Kanies2009-02-198-605/+22
| | | | | | | | | | | | | | | | Most of these were just obsolete tests that have been sitting around and broke with recent internal changes. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Correctly handling URI escaping throughout the REST processLuke Kanies2009-02-1913-97/+57
| | | | | | | | | | | | | | | | | | | | This means, at the least, that we can now serve files via REST when they have spaces and other weird characters in their names. This involves a small change to many files. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Refactoring how the Settings file is parsedLuke Kanies2009-02-196-113/+169
| | | | | | | | | | | | | | | | | | | | | | | | | | The goal of this refactor was to use a cached attribute for the LoadedFile instance we use to monitor whether the file needs reparsing. We were getting tests that affected later tests because they were holding on to LoadedFile stubs, somehow. The other main change here is that the Settings#parse method now knows how to look up its own file path. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Relying on 'should_parse_config' in the 'puppet' applicationLuke Kanies2009-02-192-15/+0
| | | | | | | | | | | | | | | | There was some redundancy here, as the app was marked as needing to parse the config, but then the setup hook manually parsed it. Signed-off-by: Luke Kanies <luke@madstop.com>
* | Providing better indirection authorization errorsLuke Kanies2009-02-193-1/+14
| | | | | | | | | | | | | | The errors now include the indirection and key, or the full URI used. Signed-off-by: Luke Kanies <luke@madstop.com>
generated by cgit (git 2.34.1) at 2025-09-26 13:39:56 +0000