On 2024-Apr-22, Tom Lane wrote: > The main reason there's a delta is that people don't manage to > maintain the in-tree copy perfectly (at least, they certainly > haven't done so for this past year). So we need to do that > to clean up every now and then.
Out of curiosity, I downloaded the buildfarm-generated file and re-indented the whole tree. It turns out that most commits seem to have maintained the in-tree typedefs list correctly when adding entries (even if out of alphabetical order), but a few haven't; and some people have added entries that the buildfarm script does not detect. So the import from BF will delete those entries and mess up the overall indent. For example it does stuff like +++ b/src/backend/commands/async.c @@ -399,7 +399,7 @@ typedef struct NotificationList typedef struct NotificationHash { Notification *event; /* => the actual Notification struct */ -} NotificationHash; +} NotificationHash; There's a good half dozen of those. I wonder if we're interested in keeping a (very short) manually- maintained list of symbols that we know are in use but the scripts don't extract for whatever reason. The change of NotificationHash looks surprising at first sight: apparently 095d109ccd7 deleted the only use of that type as a variable anywhere. But then I wonder if that datatype is useful at all anymore, since it only contains one pointer -- it seems we could just remove it. But there are others: InjectionPointEntry, ResourceOwnerData, JsonNonTerminal, JsonParserSem, ... -- Álvaro Herrera PostgreSQL Developer — https://www.EnterpriseDB.com/ "This is what I like so much about PostgreSQL. Most of the surprises are of the "oh wow! That's cool" Not the "oh shit!" kind. :)" Scott Marlowe, http://archives.postgresql.org/pgsql-admin/2008-10/msg00152.php