Fix unique constraint for orgs

[xoxys]

In #4817 we changed the org lookup from name="name" to name="name AND forge_id="forgeid" (https://github.com/woodpecker-ci/woodpecker/pull/4817/files#diff-eb104e63f04fbe20f88199ac5e15aa7164690f990773a43cedb458c083b16260R84) while the unique constraint was still "UQE_orgs_name" UNIQUE, btree (name).

With this change, the constraint was changes to "UQE_orgs_s" UNIQUE, btree (forge_id, name)

[qwerty287]

How can this fix the issue? Do we try to re-insert the org? There is no real multi-forge support yet where this could occur.

[xoxys]

Fair. I checked my local test instance again. I had an entry in the orgs table with forge_id=0. Don't ask me how it gets there, but this has triggered the error message from the issue with v3.2 while it has not with v3.1.

woodpecker_server=# select * from orgs;
 id | forge_id |    name    | is_user | private 
----+----------+------------+---------+---------
  1 |        0 | woodpecker | t       | f
  2 |        1 | testorg    | f       | f
  3 |        1 | testorg2   | f       | f

I might got trapped by this weird state, however, the constraint should still be fixed and match the changed lookup.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 26.25%. Comparing base (0ecd57d) to head (ca70160).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4923      +/-   ##
==========================================
+ Coverage   26.23%   26.25%   +0.01%     
==========================================
  Files         399      399              
  Lines       28397    28397              
==========================================
+ Hits         7451     7456       +5     
+ Misses      20273    20268       -5     
  Partials      673      673              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[qwerty287]

Doesn't fix the issue, see #4918 (comment) for the solution, but is good anyways.