Re: Further _bt_first simplifications for parallel index scans

2025-01-07 Thread Peter Geoghegan
On Tue, Jan 7, 2025 at 6:56 AM Matthias van de Meent wrote: > Apart from comments on comment contents and placement, no specific issues: Pushed this just now. Thanks for the review! > > + * > > + * Initialize arrays during first (unscheduled) primitive index scan. > > + */ > > I thin

Re: Further _bt_first simplifications for parallel index scans

2025-01-07 Thread Matthias van de Meent
On Thu, 2 Jan 2025 at 17:38, Peter Geoghegan wrote: > > Attached patch goes a bit further with simplifying _bt_first's > handling of seizing the parallel scan. This continues recent work from > commits 4e6e375b and b5ee4e52. > > Aside from requiring less code, the new structure relieves _bt_first

Further _bt_first simplifications for parallel index scans

2025-01-02 Thread Peter Geoghegan
Attached patch goes a bit further with simplifying _bt_first's handling of seizing the parallel scan. This continues recent work from commits 4e6e375b and b5ee4e52. Aside from requiring less code, the new structure relieves _bt_first from having separate calls to _bt_start_array_keys for the seria