Re: [HACKERS] What happened to jsonb's JENTRY_ISFIRST?

2014-08-14 Thread Heikki Linnakangas
On 08/14/2014 02:45 AM, Peter Geoghegan wrote: On Wed, Aug 13, 2014 at 3:47 PM, Tom Lane wrote: If the bit is unused now, should we be worrying about reclaiming it for better use? Like say allowing jsonb's to be larger than just a quarter of the maximum datum size? Commit d9daff0e0cb15221789

Re: [HACKERS] What happened to jsonb's JENTRY_ISFIRST?

2014-08-13 Thread Peter Geoghegan
On Wed, Aug 13, 2014 at 3:47 PM, Tom Lane wrote: > If the bit is unused now, should we be worrying about reclaiming it for > better use? Like say allowing jsonb's to be larger than just a quarter > of the maximum datum size? Commit d9daff0e0cb15221789e6c50d9733c8754c054fb removed it. This is an

[HACKERS] What happened to jsonb's JENTRY_ISFIRST?

2014-08-13 Thread Tom Lane
jsonb.h claims that the high order bit of a JEntry word is set on the first element of a JEntry array. However, AFAICS, JBE_ISFIRST() is not used anywhere, which is a good thing because it refers to a constant JENTRY_ISFIRST that's not defined anywhere. Is this comment just a leftover from a conv