CVS regression test problem

On Mon, Apr 25, 2005 at 01:18:29AM +0400, Oleg Bartunov wrote:

Just for information.
I see many regression tests failed (26 of 96). regression.diffs shows many
error messages like ! ERROR: unrecognized node type: 516

Did you initdb? I see no such problem with today's tip and my shared
row locking patch ... I think Tom committed something without
increasing catversion, but I'm not sure (my patch changes
PG_CONTROL_VERSION so I had to initdb anyway).

I assume you make distclean'd ...

--
Alvaro Herrera (<alvherre[@]dcc.uchile.cl>)
"Nunca se desea ardientemente lo que solo se desea por raz�n" (F. Alexandre)

On Sun, 24 Apr 2005, Alvaro Herrera wrote:

On Mon, Apr 25, 2005 at 01:18:29AM +0400, Oleg Bartunov wrote:

Just for information.
I see many regression tests failed (26 of 96). regression.diffs shows many
error messages like ! ERROR: unrecognized node type: 516

Did you initdb? I see no such problem with today's tip and my shared

yes, I did.

row locking patch ... I think Tom committed something without
increasing catversion, but I'm not sure (my patch changes
PG_CONTROL_VERSION so I had to initdb anyway).

I assume you make distclean'd ...

aha, make distclean helps ! Sorry for bothering

Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83

Oleg Bartunov <oleg@sai.msu.su> writes:

On Sun, 24 Apr 2005, Alvaro Herrera wrote:

error messages like ! ERROR: unrecognized node type: 516

I assume you make distclean'd ...

aha, make distclean helps ! Sorry for bothering

I added/removed some node types, so if you didn't do a full rebuild
you'd probably have various files out of sync about which node type
has which number.

AFAICS there are two reasonable ways to work with updating from CVS:
1. make distclean (or at least make clean) each time you update;
2. configure with --enable-depend and trust gcc to get it right.

If you don't do either then you *will* get burnt. I personally use
method #1 --- I don't know what the reliability of method #2 is.

regards, tom lane