Re: psql: customizable readline history filename - Mailing list pgsql-patches

From Bruce Momjian
Subject Re: psql: customizable readline history filename
Date
Msg-id 200506101628.j5AGSqt09706@candle.pha.pa.us
Whole thread Raw
In response to Re: psql: customizable readline history filename  (Andreas Seltenreich <seltenreich@gmx.de>)
List pgsql-patches
Andreas Seltenreich wrote:
> Bruce Momjian schrob:
>
> > I noticed your documentation example had a space before :DBNAME --- I
> > removed the space.
>
> This space is necessary for proper expansion of the parameter. The
> \set-Command concatenates all its arguments after expansion:
>
> --8<---------------cut here---------------start------------->8---
> scratch=# \set HISTFILE .psql_history-:DBNAME
> scratch=# \echo :HISTFILE
> .psql_history-:DBNAME
> scratch=# \set HISTFILE .psql_history- :DBNAME
> scratch=# \echo :HISTFILE
> .psql_history-scratch
> scratch=#
> --8<---------------cut here---------------end--------------->8---
>
> > You had the idea of not saving the history on exit if HISTFILE is not
> > set.  I don't think we have community agreement on that change, and I
> > bet we wouldn't get it.  (I have removed your documentation mention of
> > this.)  I think setting HISTSIZE to zero has that effect already if
> > people want it.
>
> I went for bash's behaviour back then, since we copied bash's
> semantics with the other history options too.

Ah, I see.  I tested ':' expansion in queries and found the space wasn't
needed, but I see now that with \set it is needed:

    test=> \set x a
    test=> \echo :x
    a
    test=> \set y b:x
    test=> \echo :y
    b:x
    test=> \set y b :x
    test=> \echo :y
    ba

I will update the documentation to add the space.  Thanks.

--
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

pgsql-patches by date:

Previous
From: Neil Conway
Date:
Subject: Re: final light versions of Oracle compatibility (SQLSTATE,
Next
From: Tatsuo Ishii
Date:
Subject: Re: bugfix: character-code conversion of MIC -> EUC_JP.