Re: Should pg_current_wal_location() become pg_current_wal_lsn()
Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>
From: Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>
To: robertmhaas@gmail.com
Cc: tgl@sss.pgh.pa.us, peter.eisentraut@2ndquadrant.com, bruce@momjian.us,
david.rowley@2ndquadrant.com, pgsql-hackers@postgresql.org
Date: 2017-04-17T05:54:52Z
Lists: pgsql-hackers
At Sat, 15 Apr 2017 12:56:32 -0400, Robert Haas <robertmhaas@gmail.com> wrote in <CA+TgmoZjDo9ckxf6aYrqyMoiSw5yfBB2gpMbrBtE9zr==uczhw@mail.gmail.com> > On Fri, Apr 14, 2017 at 6:52 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote: > > Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes: > >> If we're talking about making things easier to understand, wouldn't a > >> random user rather know what a WAL "location" is instead of a WAL "LSN"? > > > > I wouldn't object to standardizing on "location" instead of "lsn" in the > > related function and column names. What I don't like is using different > > words for the same thing. > > The case mentioned in the subject of this thread has been using the > word "location" since time immemorial. It's true that we've already > renamed it (xlog -> wal) in this release, so if we want to standardize > on lsn, now's certainly the time to do it. I'm worried that > pg_current_wal_lsn() is an identifier composed almost entirely of > abbreviations and therefore possibly just as impenetrable as > qx_current_pfq_dnr(), but maybe we should assume that LSN is a term of > art with which knowledgeable users are required to be familiar, much > as we are doing for "WAL". > > It appears, from grepping the 9.6 version of pg_proc.h, that both lsn > and location have some historical precedent. I'd better to have replied here. The detail is in my reply on another brandh of this thread. https://www.postgresql.org/message-id/20170417.143937.232025253.horiguchi.kyotaro@lab.ntt.co.jp After all, "location" seems better to me in user interface. We could rename almost all of %lsn% names into %location% except pg_lsn oprators as long as it doesn't become too long or complex. One annoyance is the historical function pg_xlog_location_diff(), which is currently just an alias of pg_lsn_mi. It is substantially an operator, but 'pg_wal_lsn_diff()' is so far from the historical name that it becomes totally useless. regards, -- Kyotaro Horiguchi NTT Open Source Software Center
Commits
-
Rename WAL-related functions and views to use "lsn" not "location".
- d10c626de47d 10.0 landed