Rename current one queue processed before saving a record to pre_queue and add post_queue processed after saving. This is required in case we need to use id generated in Katello db in orchestration (e.g. make a refference in CP to object in Katello - environments orchestration). When a task on post_queue fails, also completed tasks in pre_queue are processed.
result of changes related to adding top level errata to a changeset.
This reverts commit 54db6b5.
…and space This commit addresses a couple of small issues observed w/ the search query: 1. support searching for repos containing dash or spaces in the repo name 2. fix the filter for environment_id to be id... this filter was incorrectly allowing matches based on env vs repo id
Handling a problem described also in 790356.
…P_HOST Search history is implemented to look in to the HTTP_REFERER to determine which page the user is requesting search history for. The HTTP_REFERER, however, contains the full URI (e.g. http://this.server.com/katello/organizations). For this to work, we need to strip anything prior to the katello prefix. Previously, we would do this by looking at the HTTP_HOST in the incoming request. The issue with this is that in production configuration, we see the host no longer matches what is in the referer. This is likely due to the apache front end. In order to address this, we can use the ruby URI class to parse out the path.
…ptions tab of UI
directly to a package filter.
Before this commit, it was possible that user would click 'delete all' to destroy notices and possibly see 'some' (e.g. handfull) notices still rendered. This is due to a race condition between the destroying of the notices (in DB+search server) and the re-rendering of the notices. In order to address this, we are converting the page to use ajax (consistent with the rest of the UI) and removing the content from the table vs asking the server for 'an updated list'.
repository will properly show up beside the Remove link.