- May 13, 2025
-
- May 10, 2025
- May 08, 2025
-
-
semantic-release-ci authored
# [4.2.0](v4.1.0...v4.2.0) (2025-05-08) ### Features * Add support for .tfignore file to ignore files in module archives when supplying modules as archives ([73d7d956](73d7d956)), closes [#561](#561)
-
- May 07, 2025
- May 06, 2025
-
-
semantic-release-ci authored
# [4.1.0](v4.0.0...v4.1.0) (2025-05-06) ### Bug Fixes * Add columns to group by for module search for tables that are included in the select ([96a531e4](96a531e4)), closes [#562](#562) * Cast action enum to a string to fix ordering of action in audit ([e81fd5e2](e81fd5e2)), closes [#562](#562) * **db:** Fix enum migrations for Postgres ([2354fb33](2354fb33)), closes [#562](#562) * **db:** Handle constraints not existing (by checking if they exist) ([60e7d113](60e7d113)), closes [#562](#562) * **db:** Mark all 'id' columns as incremental, ensuring that incrementing IDs are used in Postgres ([ccba6962](ccba6962)), closes [#562](#562) * Ensure audit action column is cast where performing string comparison ([48149fcd](48149fcd)), closes [#562](#562) * Fix enum creation for Postgres, as the Enum must be created beforehand ([e5b4d541](e5b4d541)), closes [#562](#562) * Fix non-consistent ordering of results in Postgresl in search when relevant scores are the same. ([58086a5f](58086a5f)), closes [#562](#562) * Fix provider search base query to include all columns in WHERE clause in the GROUP BY clause ([0eba7dc8](0eba7dc8)), closes [#562](#562) * Fix sql query, where table is being selected, which treats first argument as a where clause, resulting in the column being added to "WHERE" without an operator. ([c787ab1b](c787ab1b)), closes [#562](#562) * Handle differences in true/false values in Postgres for DB migration scripts ([180cdaa6](180cdaa6)), closes [#562](#562) * Handle git_provider_id being an empty string and do not treat as a set value in create module provider endpoint ([8387d663](8387d663)), closes [#562](#562) * Include all selected fields that are grouped by in analytics base and module usage count queries ([a31e8d1e](a31e8d1e)), closes [#562](#562) * Include ID column in analytics query, as this is used in the group_by, meaning it must be included in the selected fields, otherwise Postgres will throw an error ([8878ff63](8878ff63)), closes [#562](#562) * Lower-case all remaining non-timestamp fields when ordering audit data. Update tests to expect the fact that the engines still don't match. ([2231ba73](2231ba73)), closes [#562](#562) * Only attempt to resolve GitRepository when updating module provider if the git provider ID is non-empty ([376bf6a4](376bf6a4)), closes [#562](#562) * Order rows returned for prometheus metrics ([fc3c1db0](fc3c1db0)), closes [#562](#562) * Perform lower-case comparison of username in audit filtering ([43cf6966](43cf6966)), closes [#562](#562) * Replace use of `lastrowid` with `inserted_primary_key` when obtaining primary key values of inserted rows. ([bc001856](bc001856)), closes [#562](#562) * Update DB migrations to use LargeBinary as variant to "blob" for Postgresql as blob is not available for Postgresql ([01739087](01739087)), closes [#562](#562) ### Features * Add package for postgresql connector ([74c792c3](74c792c3)), closes [#562](#562) * **database:** Add support for Postgres server ([1197088b](1197088b)), closes [#562](#562)
-
- May 04, 2025
-
- May 03, 2025
-
-
Matt authored
fix: Fix non-consistent ordering of results in Postgresl in search when relevant scores are the same. This resulted in results being duplicated between pages, because the results did not always get returned in the same order, the pagination was effectively broken. Also sort by provider and module name, to ensure a consistent result. Handle test differences between engines and update expected values based on the new results Issue #562
- Apr 28, 2025
-
- Apr 25, 2025
-
- Apr 24, 2025
-
-
Matt authored
fix: Lower-case all remaining non-timestamp fields when ordering audit data. Update tests to expect the fact that the engines still don't match. Not many people use this project, even fewer likely setup OIDC to identify users. Almost no-one probably uses the audit page, let alone the ordering of columns, let alone cares that empty/symbols aren't ordered when they switched from MySQL to Postges... If they do, I'm sorry I didn't spend an extra hour on tweaking SQL to work the same on all three (but hey, at least it's not degrading audit-search performance to performe a load more casts/case statements to fake/replace data for ordering)... but the search is still there, so this can easily filter data rather than relying on hundreds of rows of audit sorting in a particular way that the thing you wants happens to be on the top (or bottom). ... Now I've wasted an extra minute getting caught up in this. Well, I got distracted by https://cookieplmonster.github.io/2025/04/23/gta-san-andreas-win11-24h2-bug/ and the ...
-
- Apr 23, 2025
-
-
Matt authored
The Postgresql varient caused MySQL migrations to throw the following: ``` sqlalchemy.exc.UnsupportedCompilationError: Compiler <sqlalchemy.dialects.mysql.base.MySQLDDLCompiler object at 0x7f47b1a89220> can't render element of type <class 'sqlalchemy.dialects.postgresql.base.CreateEnumType'> (Background on this error at: https://sqlalche.me/e/14/l7de) ``` Issue #562
- Apr 22, 2025
-