Re: [PG-devel] no deactivation-hooks when killing fully asserted active buffer

2011-01-21 Thread Hendrik Tews
David Aspinall writes: Do you want to repair by adding a test for fully processed buffer? I reused an existing test case, see the description in coq/ex/test-cases/retract-completely-asserted/README. I would like to leave the work on the fix to you. Hendrik

[PG-devel] how to unlock ancestors

2011-01-21 Thread Hendrik Tews
Hi, for coq I lock ancestor buffers inside the processing of Require commands. Because these Require commands may stand at arbitrary positions in the sources, I should unlock ancestors when Require commands are retracted. I found the following solutions: 1. convert span-delete-action into