On Fri, Jan 11, 2019 at 11:05 AM Andres Freund <and...@anarazel.de> wrote:
> Hi, > > On 2018-12-19 14:21:29 -0500, Robert Haas wrote: > > On Tue, Dec 18, 2018 at 11:17 PM Andres Freund <and...@anarazel.de> > wrote: > > > Another would be to be aggressive in renaming, and deconflict by > > > renaming functions like heap_create[_with_catalog] etc to sound more > > > accurate. I think that has some appeal, because a lot of those names > > > aren't describing their tasks particularly well. > > > > I like that option. > > I'd like to start doing that by moving the functions in the following > heapam.h block elsewhere: > > /* in heap/heapam.c */ > extern Relation relation_open(Oid relationId, LOCKMODE lockmode); > extern Relation try_relation_open(Oid relationId, LOCKMODE lockmode); > extern Relation relation_openrv(const RangeVar *relation, LOCKMODE > lockmode); > extern Relation relation_openrv_extended(const RangeVar *relation, > LOCKMODE lockmode, bool > missing_ok); > extern void relation_close(Relation relation, LOCKMODE lockmode); > > extern Relation heap_open(Oid relationId, LOCKMODE lockmode); > extern Relation heap_openrv(const RangeVar *relation, LOCKMODE lockmode); > extern Relation heap_openrv_extended(const RangeVar *relation, > LOCKMODE lockmode, bool > missing_ok); > > ISTM that the first block would best belong into new files like > access/rel[ation].h and access/common/rel[ation].h. I think the second > set should be renamed to be table_open() (with backward compat > redirects, there's way way too many references) and should go into > access/table.h access/table/table.c alongside tableam.[ch], but I could > also see just putting them into relation.[ch]. > > Comments? > Yes, that will be good. Regards, Haribabu Kommi Fujitsu Australia
