SearchWP 4.3.15

### 2.9.1
- **[Fix]** Fixed a potential issue with `sql_mode=only_full_group_by` support (added in 2.9)
- **[Fix]** Avoid error when parsing PDFs without `mbstring`

### 2.9
- **[New]** Redesigned engine configuration interface!
- **[New]** Index is now further optimized as per engine settings
- **[New]** New filter `searchwp_weight_max` to customize a maximum weight
- **[New]** New filter `searchwp_legacy_settings_ui` to use legacy settings UI
- **[New]** New filter `searchwp_indexer_apply_engines_rules` to control whether engine rules are considered during indexing
- **[New]** New filter `searchwp_indexer_additional_meta_exclusions` to control whether default additional Custom Fields are excluded from indexing
- **[New]** New filter `searchwp_supports_label_{post_type}_{support}` to customize the label used for a post type native attribute
- **[Improvement]** Additional debug statements output when enabled
- **[Improvement]** Better formatting of HTML comment debug output
- **[Fix]** Less aggressive handling of `pre_get_posts` during searching
- **[Fix]** Fix an issue with `sql_mode=only_full_group_by` (default in MySQL 5.7)

### 2.8.17
- **[Update]** Updated updater
- **[Fix]** Fixed an issue with database table creation

### 2.8.16
- **[Fix]** Fixed an issue where exclusionary weight was not properly applied to all Custom Fields
- **[Fix]** Resolved premature AND logic aggressiveness in some cases
- **[Fix]** Revised database creation schema to better cooperate with more database environments
- **[Improvement]** Better assumed sorting of results
- **[New]** New filter `searchwp_stats_table_class` to control CSS class applied to stats output

- **[Fix]** Fixed an issue where in some circumstances content blocks parsed from PDFs would not be properly separated, resulted in the last word of one section being lumped together with the first word of the next section
- **[Improvement]** Better handling of regex whitelist matches that result with multi-word tokens (e.g. spaces within) NOTE: having multi-word matches is not recommended
- **[Fix]** Fixed an issue where Heartbeat API-powered timestamp of last index was missing "Last Indexed" phrasing
- **[New]** New Filter: `searchwp_omit_meta_key` allows developers to omit specific meta keys from being indexed during indexing
- **[Improvement]** Added `$engine` parameter to `searchwp_query_orderby` filter
- **[Improvement]** Simplified the check for a stalled indexer
- **[Fix]** Prevent over-preparation of terms when performing AND logic refinement
- **[Fix]** Check for indexer being disabled when issuing delta updates