On 09.03.23 18:38, Tom Lane wrote:
> While reviewing this patch, I started to wonder why we don't eliminate
> the maintenance hassle of xml_1.out by putting in a short-circuit
> at the top of the test, similar to those in some other scripts:
>
> /* skip test if XML support not compiled in */
> SELECT '<value>one</value>'::xml;
> \if :ERROR
> \quit
> \endif
>
> (and I guess xmlmap.sql could get the same treatment).
>
> The only argument I can think of against it is that the current
> approach ensures we produce a clean error (and not, say, a crash)
> for all xml.c entry points not just xml_in. I'm not sure how much
> that's worth though. The compiler/linker would tell us if we miss
> compiling out every reference to libxml2.
>
> Thoughts?
>
> regards, tom lane
Hi Tom,
I agree it would make things easier and it could indeed save some time
(and some CI runs ;)).
However, checking in the absence of libxml2 if an error message is
raised, and checking if this error message is the one we expect, is IMHO
also a very nice test. But I guess I could also live with skipping the
whole thing.
Best, Jim
