Regression tests fails to start on system without unix sockets.

Started by Kurt Roeckxover 22 years ago3 messages
#1Kurt Roeckx
Q@ping.be

The regression tests will fail to start on a system that doesn't
have, or wasn't compiled for, unix domain sockets.

The pg_regress script will start the postmaster with the -i
option in case of QNX and BEOS, but not for instance in case of
WIN32.

I see some options to fix this:
- Always start with -i
- Make the "unix_sockets" variable depend on
HAVE_UNIX_SOCKETS intead of listen the OSs.

Kurt

#2Tom Lane
tgl@sss.pgh.pa.us
In reply to: Kurt Roeckx (#1)
Re: Regression tests fails to start on system without unix sockets.

Kurt Roeckx <Q@ping.be> writes:

The regression tests will fail to start on a system that doesn't
have, or wasn't compiled for, unix domain sockets.
I see some options to fix this:
- Always start with -i
- Make the "unix_sockets" variable depend on
HAVE_UNIX_SOCKETS intead of listen the OSs.

The second way is the way it should have been done all along.

Probably the best fix is to add a command-line switch to pg_regress
to instruct it to use -i, and then have the makefile test
HAVE_UNIX_SOCKETS to decide whether to pass that switch. This way,
hand invocation of the script could easily run the test both ways,
on machines where that's possible.

regards, tom lane

#3Bruce Momjian
pgman@candle.pha.pa.us
In reply to: Tom Lane (#2)
Re: Regression tests fails to start on system without unix

Added to TODO:

Allow the regression tests to start postmaster with -i so the tests
can be run on systems that don't support unix-domain sockets

---------------------------------------------------------------------------

Tom Lane wrote:

Kurt Roeckx <Q@ping.be> writes:

The regression tests will fail to start on a system that doesn't
have, or wasn't compiled for, unix domain sockets.
I see some options to fix this:
- Always start with -i
- Make the "unix_sockets" variable depend on
HAVE_UNIX_SOCKETS intead of listen the OSs.

The second way is the way it should have been done all along.

Probably the best fix is to add a command-line switch to pg_regress
to instruct it to use -i, and then have the makefile test
HAVE_UNIX_SOCKETS to decide whether to pass that switch. This way,
hand invocation of the script could easily run the test both ways,
on machines where that's possible.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org

-- 
  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