Search for text that includes underscore returns nothing

[mfigrs]

• Gitea version (or commit ref): 6df46c5
• Git version: 2.15.2
• Operating system: Docker on Synology
• Database (use [x]):
  • PostgreSQL
  • MySQL
  • MSSQL
  • [ x ] SQLite
• Can you reproduce the bug at https://try.gitea.io:
  • Yes (provide example URL)
  • No
  • Not relevant
• Log gist:

Description

I've enabled repository search in app.ini, however this feature is next to useless. If I search for a function name, say "list_users", the search returns nothing. Same for "list_" or "_users" - it seems as if the search doesn't like the underscore symbol.

[lunny]

It works here.

[mfigrs]

But here it does not:

[lunny]

If you just enabled the feature on app.ini, it will spend some time to create the index, maybe the index is not finished.

[mfigrs]

I've enabled this feature over a week ago. Anyway, if I just enter "user" without any underscore, I have 3 pages of results:

[zeripath]

Hmm... I bet this has to do with the way bleve segments out words. Could it be that it's seeing and indexing "list_users" as "list users"?

[mfigrs]

Hmm... I bet this has to do with the way bleve segments out words. Could it be that it's seeing and indexing "list_users" as "list users"?

A search for "list users" also returns nothing, unfortunately...

Some other points:

• search for "admin-users" text returns correct results
• search for "admin-users.php" returns nothing, but it should return identical results to the above.

Furthermore:
Searching for ".users", "/users", "\users" seem to return the same results as plain search for "users"...

[zeripath]

Hmm. OK a quick test of blevesearch/segment seems to imply that "list_users" would form a single segment.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs during the next 2 weeks. Thank you for your contributions.

[barryp]

I just tried it with gitea 1.6.4 and it's still an issue. Single word queries give results, but snake_cased queries return nothing even though the variable names are present in the repo.

[lunny]

The words splitter just uses spaces to split the words. I think that's why when you have list_users but cannot search list or users. The current code search is very simple which need improvements.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs during the next 2 weeks. Thank you for your contributions.

[bartwensley]

This issue renders the search functionality pretty much useless for code that regularly includes underscores (e.g. python). A couple more examples...

The following search returns a single result (there should be 33 matches from my count):
https://opendev.org/starlingx/config/search?q=_get_service_config

The following search returns no results (there should be 3 matches from my count):
https://opendev.org/starlingx/config/search?q=_get_admin_user_name