summaryrefslogtreecommitdiffstats
path: root/lib
Commit message (Collapse)AuthorAgeFilesLines
...
* Fix for #3690 failing to calculate error codesJesse Wolfe2010-02-174-3/+5
| | | | | | | | | | | | | | | | | | | | | This failure was getting caused by what I believe to be a bug in Puppet::Configurer where it always generated a new Puppet::Transaction::Report, even if one existed in an outer scope. In puppetd --test, a different Report was getting queried to generate the exit status than the one that was passed to the transaction -- this Report had no Metrics and would fail when queried. This obscured a second bug that Metrics could return nil for values if the Transaction had applied an empty Catalog, but Transaction::Report assumes that values will always be integers. It could be argued that an empty Report should be populated with empty Metrics before a Transaction is run, which would have prevented Report#exit_status from raising an exception ... which would have made these bugs much harder to track down. So, I've decided to leave that unchanged. Signed-off-by: Jesse Wolfe <jes5199@gmail.com>
* Fixing #3668 - fixed autoloading classes from modulesLuke Kanies2010-02-175-872/+895
| | | | | | | | | | | This involved essentially moving all of the importing and loading code out of the Parser and into a new 'TypeLoader' class. The parser and the ResourceTypeCollection classes now delegate to that class for all file handling. Most of the code paths are also now much cleaner, and a bit of redundancy was removed. Signed-off-by: Luke Kanies <luke@puppetlabs.com>
* Fix #3656 JSON serialization of dependenciesJesse Wolfe2010-02-171-1/+5
| | | | | | | | | | | | | | | | | | | | The pson serialization of resources was behaving incorrectly on parameters that are references to other resources: 1. Dependency parameters (require, subscribe, notify) were getting serialized as anonymous objects that looked like partially constructed resources 2. During de-serialization the pson parser would inflate them into hashes (rather than into resources) 3. The outer resource would try to coerce the hash into a resource by passing it to Resource.new 4. Resource.new would fail with a cryptic message, since it does not accept a hash as its first parameter (but the error is obfuscated by Resource.new's complicated argument handler) This patch solves the problem by explicitly converting dependency parameters into strings in the pson serialization. Signed-off-by: Jesse Wolfe <jes5199@gmail.com>
* Fixes #3582 - Adds dbport configuration option for specifying database portJames Turnbull2010-02-172-0/+3
|
* Adding indirector support to Resource TypesLuke Kanies2010-02-174-0/+73
| | | | | | | | | | Also adding JSON support. This is so that we can remotely retrieve information about resource types and classes, such as what arguments are required. Signed-off-by: Luke Kanies <luke@puppetlabs.com>
* Fix #3552 single executable should display usageJesse Wolfe2010-02-171-0/+15
| | | | | | | | | Added some tests to make the single executable command behavior explicit. Added logic to display the usage message if we're on a tty and no arguments are passed. Signed-off-by: Jesse Wolfe <jes5199@gmail.com>
* Added support for flat packages in the pkgdmg package provider.Roy Nielsen2010-02-171-21/+28
| | | | | | | | | | | | | | Added a test in: ./spec/unit/provider/package/pkgdmg.rb To test flat package support. The case where a package is a .pkg bundle, curl will attempt to download and not work. The "installer" command will then fail, as the source will be "not found" and the resource will fail. The puppet run will continue. Signed-off-by: Roy Nielsen <rsn@lanl.gov>
* Fix: the rcvar name is not necessarily the same as the service name. (More ↵Anselm Strauss2010-02-171-16/+39
| | | | and more I get the feeling that FreeBSD's rc system is totally unsuitable for Puppet. What about porting Upstart or SMF to FreeBSD ... ?)
* Added proper status commandAnselm Strauss2010-02-171-0/+4
|
* Re-included old BSD service provider, made new one default for FreeBSDAnselm Strauss2010-02-172-1/+56
|
* Rewrote FreeBSD service providerAnselm Strauss2010-02-171-39/+94
|
* Fixed documentation issues exposed in #3772James Turnbull2010-02-171-4/+4
|
* Renamed all references to Reductive Labs to Puppet LabsMarkus Roberts2010-02-172-2/+2
| | | | Find and replace to the rescue...
* Fix for #3107 Changing users on AIXMarkus Roberts2010-02-171-18/+4
| | | | | | | | | | | | | | There are several issues with changing the real, effective, and saved group and user ids in different environments (which methods to call, in what order, etc). While the code being replaced by this patch appeared to work for Linux, Solaris, and (with a special case test) Darwin; it was failing under AIX and may have had edge-case problems under the others. Ruby back to 1.8.1 has supported a higher level interface that deals with the problem and captures a broader range of OSes; it's a single call for group and one for user--the details of rid/eid/svid, etc ordering are handled internally. Switching to that simplifies our code and should improve/unify our support of various OSes.
* Fixing #2864 Added support for AIX System Resource Controller (SRC) - ↵Andrew Forgue2010-02-171-0/+91
| | | | | | | | | | | | | service start stop This provider supports start/stop and restart of AIX services using the native AIX service manager, called the System Resource Controller. Currently it will not stop and start (but only refresh) a service that uses sockets or message queues as its communication method. It will run stopsrc and then startsrc for services that use signals as their communication method. Signed-off-by: Andrew Forgue <andrew.forgue@gmail.com>
* Fixes #2836. Add hold support to dpkg providerNigel Kersten2010-02-172-2/+41
|
* Fixes #1223 Add Zypper support for SuSE machinesRein Henrichs2010-02-171-0/+52
| | | | | | | | | | | | | | | | | | | Zypper is the replacement for `rug' from earlier SuSE releases. Zypper is backward compatible with the rug command (mostly) and supports most of the same commands that rug does. This version fixes a number of bugs in the original: * when installing with a specified version, fix bug where the package name was being doubled ("foo" became "foofoo"). * fix bug where package name and version were separated by a "=" when it should have been a "-". * Update specs to reflect the implementation's use of the "-l" flag as recommended in http://groups.google.com/group/puppet-dev/msg/d86416c079bd3faf Signed-off-by: Rein Henrichs <reinh@reinh.com>
* Fix for #3399 zone type should handle exclusive IP stacksRein Henrichs2010-02-172-30/+80
| | | | | | | | | | * corrected missing status * added cloning and support for default router * RH: Fix spec to return accurate value for @resource[:clone] * RH: Add spec for untested install case when @resource[:clone] returns a (non-falsy) value Signed-off-by: Rein Henrichs <rein@puppetlabs.com>
* Adding #3518 - basic dot format supportLuke Kanies2010-02-171-0/+3
| | | | | | | This allows you to ask for a node's catalog in dot format, for inputting into graphviz et al. Signed-off-by: Luke Kanies <luke@reductivelabs.com>
* Fixing #3533 - Removing all transaction cleanupLuke Kanies2010-02-172-15/+0
| | | | | | | | | | It's really slow and has no actual functionality any more, since we just remove the catalogs from memory anyway. This should be a good speed boost for very little effort. Signed-off-by: Luke Kanies <luke@puppetlabs.com>
* Fix for #2910 -- Tidy/matches is too tricky to useMarkus Roberts2010-02-171-29/+35
| | | | | | | | | | | | | The semantic interaction of tidy/matches and tidy/recurse is tricky to get right; it only makes sense to use matches with recursion (a fixed path will either statically match or it won't, no need for a run-time check) but there was nothing to warn users of this fact. To compound matters, the example in the matches parameter doc string even made this mistake. This patch: 1) fixes the doc string; 2) prohibits the use of match without a value of recurse capable of generating files to match, 3) fixes tests that were passing for the wrong reason and adds tests on the prohibition added in (2).
* Bug #3451: Don't leak the terminus class setting from ↵Jesse Wolfe2010-02-171-0/+4
| | | | | | | Puppet::Resource::Catalog's spec This issue causes other specs to fail, because they depend on the default terminus being unchanged.
* Fix to the fix for #3295Markus Roberts2010-02-171-6/+2
| | | | | | The output variable in the inner block wasn't visible in the outer block, and wasn't needed in any case, since the results are returned naturally if you just leave everything alone.
* Write ssh_authorized_keys as userMarkus Roberts2010-02-171-29/+9
| | | | | | | | | | | | This is a targeted fix to the issue of permissions when writing ssh authorized key files by 1) requiring that an existing users be specified on the resource and 2) doing the write as that user. It's based on Michael DeHaan's initial implementation of Luke's idea, but with a number of simplifications (mostly by testing necessary conditions as early as possible so the code isn't cluttered up with a lot of checks). The tests in this version are modified slightly to remove some additional implementation couplings that were added in master.
* Fix for #3558 -- source file reading speedupMarkus Roberts2010-02-171-7/+1
| | | | | | It's about 10x faster to read the whole file than to read each line and concatenate them (actually, it's O(n) vs. O(n^2), so the exact speedup depends on the file size).
* Fix for #3556 Plussignment value meldingMarkus Roberts2010-02-171-4/+10
| | | | | | | | | | | | | | The plussignment operator was constructing the new parameter value by modifying the param object's value in place (so as to preserve the file and line information for debugging). However, when multiple resources are overridden by the same plussignment this would result in all of the resources sharing the same value (the union of all the prior values and the new value), which is wrong. Instead, we need to give each resource its own copy of the value (e.g., a copy of the param object), which this patch implements. Signed-off-by: Markus Roberts <Markus@reality.com>
* Fixed #3655 - Puppet doesn't find installed packages with portage providerJames Turnbull2010-02-171-2/+2
| | | | Thanks to Benedikt Böhm for the patch
* Fixed #3672 - Error message on duplicate fileserver mounts incorrectJames Turnbull2010-02-171-2/+2
|
* Fixed stored configuration documentationJames Turnbull2010-02-171-3/+3
|
* Fixes #3653 - Changed default factpath value to better reflect plugins in ↵James Turnbull2010-02-171-1/+1
| | | | modules
* Partial fix to #2837 - changed warning message to debugJames Turnbull2010-02-171-1/+1
|
* Fix #3555 - fix state of purged listsPeter Meier2010-02-171-0/+10
| | | | | As purged is not contained by default within ensureable we need to extend the parameter to properly deal with purged lists.
* Fix #3540 - name methods correctlyPeter Meier2010-02-171-2/+2
| | | | | The "Ensure Parameter" tries to call destroy on the provider and not delete.
* Fixes #3295 - generate() now sets the working directory to the directory ↵Paul Lathrop2010-02-171-2/+4
| | | | | | containing the specified command. Also adds rspec tests for generate().
* Fixed #3532 - Typo in lib/puppet/ssl/host.rbJames Turnbull2010-02-171-1/+1
| | | | Thanks to Jasper Lievisse Adriaanse for the fix.
* Updated version and CHANGELOG to 0.25.5rc1James Turnbull2010-02-171-1/+1
|
* Fixes #3460 - Makes Puppet FHS compliant by moving /var/puppet to ↵James Turnbull2010-02-172-2/+2
| | | | /var/lib/puppet
* Fix for #3101 (bug in MRI 1.8.7)Markus Roberts2010-02-171-1/+12
| | | | | | | | | | | Due to a bug in Ruby 1.8.7 net/http will attempt to close a connection that wasn't successfully opened (it's nil), first checking to see if the connection is already close, and thus raising a method missing exception. This bug causes error messages that are confusing / misleading. To get around this, we add a closed? method to nil such that a nil (unopened) connection is always considered closed, allowing the real problem to be reported.
* Fixes #3419. OS X 10.6 Ruby doesn't set supplementary groupsNigel Kersten2010-02-171-2/+26
|
* Revert the guts of #2890Markus Roberts2010-02-174-32/+40
| | | | | | | | | | This patch reverts the semantically significant parts of #2890 due to the issues discussed on #3360 (security concerns when used with autosign, inconsistency between REST & XMLRPC semantics) but leaves the semantically neutral changes (code cleanup, added tests) in place. This patch is intended for 0.25.x, but may also be applied as a step in the resolution of #3450 (refactored #2890, add "remove_certs" flag) in Rolwf.
* Fail gracefully on packages that don't have the HOMEPAGE variable set (e.g. ↵Gábor Vészi2010-02-171-1/+1
| | | | dev-lang/php).
* Fixed #3443 - Typo in mount typeJames Turnbull2010-02-171-2/+3
|
* Fixes #3135 - darwin doesn't support 'mount -o remount'James Turnbull2010-02-171-1/+2
| | | | Thanks for Eric Sorenson for the patch.
* Adding :catalog_terminus settingLuke Kanies2010-02-174-6/+15
| | | | | | | | | | This will rarely be used, but it enables even more architectural flexibility, such as precompiling catalogs and storing them in memcached or equivalent. With this setup, a single host can probably serve all catalogs and you would then just have as many compiling hosts as needed. Signed-off-by: Luke Kanies <luke@reductivelabs.com>
* fixing obsolete comment in puppetdLuke Kanies2010-02-171-3/+3
| | | | Signed-off-by: Luke Kanies <luke@reductivelabs.com>
* Adding support for only using cached catalogsLuke Kanies2010-02-172-1/+5
| | | | | | | | | This basically allows a sysadmin to control when a client will compile a new catalog - with this option enabled, the client will use the cached catalog as long as it has one, only recompiling when run with the option disabled. Signed-off-by: Luke Kanies <luke@reductivelabs.com>
* Refactoring Configurer to enable the next featureLuke Kanies2010-02-171-27/+30
| | | | Signed-off-by: Luke Kanies <luke@reductivelabs.com>
* Fix #3155 - prevent error when using two matching regex in cascadeBrice Figureau2010-02-174-12/+68
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The following manifest: case $var { /match/: { if $var =~ /matchagain/ { } } } is failing because the "=~" operators when matching sets an ephemeral variable in the scope. But the case regex also did it, and since they both belong to the same scope, and Puppet variables are immutables, the scope raises an error. This patch fixes this issue by adding to the current scope a stack of ephemeral symbol tables. Each new match operator or case/selector with regex adds a new scope. When we get out of the case/if/selector structure the scope is reset to the ephemeral level we were when entering it. This way the following manifest produces the correct output: case $var { /match(rematch)/: { notice("1. \$0 = $0, \$1 = $1") if $var =~ /matchagain/ { notice("2. \$0 = $0, \$1 = $1") } notice("3. \$0 = $0, \$1 = $1") } } notice("4. \$0 = $0") And the output is: 1. $0 = match, $1 = rematch 2. $0 = matchagain, $1 = rematch 3. $0 = match, $1 = rematch 4. $0 = Signed-off-by: Brice Figureau <brice-puppet@daysofwonder.com>
* Fixed link typoJames Turnbull2010-04-221-1/+1
|
* Fixed #3384 - Updated broken linkJames Turnbull2010-04-221-3/+3
|
generated by cgit (git 2.34.1) at 2025-09-26 14:07:10 +0000