Hello, this is the new version of this patch.
At Fri, 26 Feb 2016 13:17:26 +0900 (Tokyo Standard Time), Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp> wrote in
<20160226.131726.54224488.horiguchi.kyotaro@lab.ntt.co.jp>
> Hello, thank you for the comments.
>
> At Mon, 15 Feb 2016 15:43:57 +0300, Artur Zakirov <a.zakirov@postgrespro.ru> wrote in
<56C1C80D.7020101@postgrespro.ru>
> > I did some tests with your patch. But I am not confident in
> > tab-complete.c.
> >
> > And I have some notes:
> >
> > 1 - I execute git apply command and get the following warning:
> >
> > ../0001-Suggest-IF-NOT-EXISTS-for-tab-completion-of-psql.patch:302:
> > trailing whitespace.
> > /*
> > warning: 1 line adds whitespace errors.
> >
> > This is because of superfluous whitespace I think.
>
> Oops. I'll remove it.
>
> > 2 - In psql I write "create table if" and press <TAB>. psql adds the
> > following:
> >
> > create table IF NOT EXISTS
> >
> > I think psql should continue with lower case if user wrote query with
> > loser case text:
>
> Good catch! Hmm. COMPLETE_WITH_SCHEMA_QUERY() does it. For
> example, the following existing completion behaves in the same
> way.
>
> "drop index c" =<tab>=> "drop index CONCURRENTLY"
>
> The results of schema queries should be treated in case-sensitive
> way so the additional keywords by 'UNION' are also treated
> case-sensitively.
>
> COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_indexes,
> " UNION SELECT 'CONCURRENTLY'");
>
> Fixing this is another problem. So I'd like to leave this alone
> here.
>
> > create table if not exists
> >
> > 3 - Same with "IF EXISTS". If a write "alter view if" and press <TAB>
> > psql writes:
> >
> > alter view IF EXISTS
Finally, the only thing done in this new patch is removing one
dangling space.
reagrds,
--
Kyotaro Horiguchi
NTT Open Source Software Center