Add setting to configure TrigramMax

Background

TrigramMax limit, which controls the maximum number of unique trigrams that can be indexed per file. If a file has more trigrams than this limit, then only the filename gets indexed. Zoekt's default TrigramMax is 20,000 unique trigrams per file. When a file exceeds this limit:

  • The filename is still indexed. The filename search works
  • The file content is NOT indexed. The content search does not work.

GitLab currently does NOT configure this value, so it uses Zoekt's default of 20,000.

Proposal

Add a json zoekt application setting to configure Zoekt's TrigramMax limit. It should be of integer type. Default will be 20000.

Edited by Ravi Kumar