Merged PRs

dolt

• 6595: Feature: Allow * wildcard in @@dolt_replicate_heads
  This change allows customers to use * as a wildcard in the @@dolt_replicate_heads system variable to match zero or more characters in a branch name. For example, to replicate the main branch and all branches that start with feature, the following configuration can now be applied:

  dolt config --local --add sqlserver.global.dolt_replicate_heads main,feature*
  

  Documentation: dolthub/docs#1698
  Fixes: #6486
• 6562: Bug fix: Building secondary prefix indexes from existing data
  When building a secondary index from existing table data and the secondary index contains a prefix of a TEXT/BLOB column, we don't currently dereference the TEXT/BLOB pointer to get the content. Instead, we take the TEXT/BLOB pointer address and incorrectly put that into the secondary index. This manifested as errors in collation sorting, since one string was not a valid, encoded string and was instead a pointer address. Additionally, we weren't applying the prefix to the content for any existing data when building a prefix index – i.e. any VARCHAR data was inserted in full to the prefix index without trimming.
  It's difficult to test indexes directly, so I updated the index validation code in our test suite to support validating prefix indexes. I added some more tests for prefix indexes in GMS, too: dolthub/go-mysql-server#1971
  Fixes #6520
• 6517: Automatically merge schema changes that widen VARCHAR fields to TEXT
  This enhancement allows more schema changes to merge cleanly (instead of being reported as a schema conflict) when we can determine that a type is being widened and all existing data is guaranteed to be safe to convert. This includes:
  • VARCHAR → TINYTEXT, TEXT, MEDIUMTEXT, LONGTEXT
  • CHAR → TINYTEXT, TEXT, MEDIUMTEXT, LONGTEXT
  • BINARY → TINYBLOB, BLOB, MEDIUMBLOB, LONGBLOB
  • VARBINARY → TINYBLOB, BLOB, MEDIUMBLOB, LONGBLOB
    Trying to merge a change to a VARCHAR column larger than 255 characters to make it a TINYTEXT column will still generate a schema conflict, since TINYTEXT has a maximum size of 255 characters.
    One of the more interesting conversions is converting to VARBINARY(N) or TINYTEXT to BINARY(N). MySQL right-pads all BINARY(N) values with null bytes, up to N, which requires us to call Type.Convert as part of remapping values in a schema merge, and also to rebuild secondary indexes in some cases.
    Fixes: #6442

go-mysql-server

• 1978: Fix table lookup bug for nil asOf
  Edit: Show keys for info schema table should work the same way as MySQL now.
  Prev: This error message is not correct, MySQL finds the table and does not error. But this at least reverts the behavior to what Dolt did before the name resolution refactors.
• 1976: Refactor prepareds to cache AST
  Prepared statements were rewritten recently to cache a query string
  with marked bind variable locations. Executing prepared statements
  would generate a full SQL query string for re-parsing, binding, and
  analysis. This update caches the AST. Executing a prepared statement
  starts at binding an AST. Bind variables will be substituted in the
  process of converting AST expression to plan expressions.
  GMS bump: #6593
• 1975: Union NULL types cast to same type as neighbor
  3+-way unions are subject to a bug where a NULL projection will be cast to a char type, and then have a type clash with the subsequent table relation. Rearranging the query to have the NULL projection last appears to fix the bug because the convert cast is moved to a position where it cannot have a type clash with a table relation to the right.
• 1974: Re-implement generated stored columns with new analyzer
  Testing revealed a bug in explicit DEFAULT values in insert statements, added a skipped test for same.
  Also rewrote column default tests as scripts.
• 1945: server: Add a warning message at startup if the server is run without secure_file_priv being set to a non-empty value.

Closed Issues

• 6442: Automatically merge changes that enlarge varchar fields to text types
• 6486: Allow pattern-matching in replicate_heads variable
• 6497: LAST_INSERT_ID() always returns 0
• 6520: Panic when inserting into prefix index on TEXT column
• 6462: go.mod causes go install to fail (must not contain directives ... interpreted differently than if it were the main module)
• 6224: When executing a query in SQL, the error message is unrelated to the SQL statement.

Latency