Обсуждение: -ltermcap needed for psql client build on OpenBSD 2.9

Doing a default ./configure (no additional options) on a freshly installed OpenBSD 2.9 results in a 'psql' binary that
isn'tlinked to termcap.  The error looks like this: 

$ ./psql
Welcome to psql, the PostgreSQL interactive terminal.

Type:  \copyright for distribution terms
       \h for help with SQL commands
       \? for help on internal slash commands
       \g or terminate with semicolon to execute query
       \q to quit

/usr/libexec/ld.so: Undefined symbol "_tgetent" called from psql:/usr/lib/librea
dline.so.0.0 at 0x40088964
$

-----snip-----

"_tgetent" is a termcap symbol, and simply appending

-ltermcap

...to the final gcc command during the 'psql' make fixes the problem.  This should probably work w/o human
intervention,however.  :-) 


PostgreSQL rocks!


Tom Stowell

"Tom Stowell" <jts@deforest.k12.wi.us> writes:
> "_tgetent" is a termcap symbol, and simply appending
> -ltermcap
> ...to the final gcc command during the 'psql' make fixes the problem.
> This should probably work w/o human intervention, however.  :-)

Indeed.  The configure script should have deduced the need to do this;
could you look into why it fails to do so on your platform?

            regards, tom lane

Tom Stowell writes:

> "_tgetent" is a termcap symbol, and simply appending
>
> -ltermcap
>
> ...to the final gcc command during the 'psql' make fixes the problem.  This should probably work w/o human
intervention,however.  :-) 

We have disqualified this as a linker bug pending OS developer feedback.
See NetBSD PR 13486:

http://www.NetBSD.org/cgi-bin/query-pr-single.pl?number=13486

No reponses yet.  Feel free to post this to an OpenBSD forum.

--
Peter Eisentraut   peter_e@gmx.net   http://funkturm.homeip.net/~peter

Tom Stowell writes:

> In the link you gave, it looked like provisions have been made in the
> development version of PostgreSQL 7.2 to work around this.

Not that I know of.

> Kind of a dumb linker bug... but our *BSD friends seem to think it's a
> feature.  :-)

We'll have to beat them into submission then.  I mean, the simplest thing
in the world would be to link libreadline against -lcurses, so the
dependency is recorded and the users of libreadline won't have to remember
it.

--
Peter Eisentraut   peter_e@gmx.net   http://funkturm.homeip.net/~peter