OpenGrok stuffs the repositories you have selected into the query string, which means that depending on the browser you're using, there is a maximum number of repositories that you can search at the same time. In theory, a repository with a long enough name could be unsearchable.
I've heard of three workarounds to this problem:
- Write a script that automates the process of "select N repos, search, select the next N repos, search, ....."
- Group repositories together in folders on the file system of the server on which OpenGrok is installed - or, in other words, use the file system to create a hierarchical structure. The top-level folders will be searchable in OpenGrok's UI. This architectural 'pattern' has additional benefits - it can be used to group all of the repositories used by a particular team under the same search heading, for example.
- Use OpenGrok Groupings to accomplish the same goal as in 2.
Regardless of the mechanism you choose, the end goal is to be able to squeeze all of the top-level search entries into the query string when you whack 'Select All' and search.