From b3055022e97d62ae2754429d87d9784730d68307 Mon Sep 17 00:00:00 2001 From: Bruce Allen Date: Thu, 13 Jan 2005 06:23:49 +0000 Subject: [PATCH] Comments for David svn path=/trunk/boinc/; revision=5094 --- checkin_notes | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/checkin_notes b/checkin_notes index 93e4ab2f07..1c0bc566ce 100755 --- a/checkin_notes +++ b/checkin_notes @@ -22551,6 +22551,28 @@ Bruce 12 Jan 2005 more days, when I get some confidence that there are not other problems to fix! Please read embedded comments/questions in code. + - Some more comemnts added later. We do sometimes get non-zero + return values from possibly_send_result(). I think that this is + because you've failed to match up the start_transaction and + commit_transaction pairs. Be careful: sleeps AND modifies the + DB. Right now I am too tired to fix this without making it worse. + Your turn. + + - A few other comments. Currently, I am 'advertising' filenames + returned by get_working_set_filename() for TWO types of files: + (1) those for which I can make more work + (2) those for which I can NOT make more work, but where you have + not flagged with flag_for_possible_removal(). + I am not sure that I understand the intent of + flag_for_possible_removal(). You surely are not guaranteeing + that no results will ever appear in the future for these files + (outstanding results might fail). So what is my WU generator + supposed to do with flag_for_possible_removal? Is (2) above + correct? + + - Finally, I have not yet seen a SINGLE instance of a + flag_for_possible_removal. Is the logic right? + sched/ sched_locality.C