Re: xmlconcat as variadic function

Tom Lane wrote:
> Peter Eisentraut <peter_e@gmx.net> writes:
>> Here is a patch to reimplement the xmlconcat functionality as a variadic 
>> function instead of a hardcoded special expression type.
> 
> What's the point of this?  I suppose making xmlconcat not a keyword is
> some small advantage, but having it behave subtly differently from the
> other xmlfoo functions isn't really all that nice.

Different from what and how?  We already have a mix of XML functions 
implemented in user space and some as built-in expressions.  And they 
all have different signatures and purposes, so I don't see a lot of need 
to align anything here.

If variadic functions had been available in 8.3, we would probably have 
implemented xmlconcat like this in the first place.

> And the change in
> this error message is not for the better:
> 
>> *** 55,63 ****
>>   (1 row)
>   
>>   SELECT xmlconcat(1, 2);
>> ! ERROR:  argument of XMLCONCAT must be type xml, not type integer
>>   LINE 1: SELECT xmlconcat(1, 2);
>> !                          ^
>>   SELECT xmlconcat('bad', '<syntax');
>>   ERROR:  invalid XML content
>>   LINE 1: SELECT xmlconcat('bad', '<syntax');
>> --- 55,64 ----
>>   (1 row)
>   
>>   SELECT xmlconcat(1, 2);
>> ! ERROR:  function xmlconcat(integer, integer) does not exist
>>   LINE 1: SELECT xmlconcat(1, 2);
>> !                ^
>> ! HINT:  No function matches the given name and argument types. You might need to add explicit type casts.
>>   SELECT xmlconcat('bad', '<syntax');
>>   ERROR:  invalid XML content
>>   LINE 1: SELECT xmlconcat('bad', '<syntax');

If you think this is a problem, then it can be fixed.  I never 
particularly liked this second error message in general.  A much better 
hint that would address this case and many others might be:

ERROR: function foo(type, type) does not exist
HINT: There is foo(type, type) and foo(type, type) instead.

There are a number of other distinguishable cases that call for better 
hints, e.g., no foo() exists at all.  I think in my experience, "you 
need to add explicit type casts" has almost never been the actual 
solution when this message comes up.