Re: Re: Email to hackers for test coverage

On Tue, 27 Aug 2019 14:07:48 +0800 michael@paquier.xyz wrote:

> There is a section in float4.sql which deals with overflow and

> underflow, so wouldn't it be better to move the tests there?  You
> could just trigger the failures with that:
> =# insert into float4_tbl values ('-10e-70'::float8);
> ERROR:  22003: value out of range: underflow
> LOCATION:  check_float4_val, float.h:145
> =# insert into float4_tbl values ('-10e70'::float8);
> ERROR:  22003: value out of range: overflow
> LOCATION:  check_float4_val, float.h:140

> I would also test all four patterns: 10e70, 10e-70, -10e70, -10e-70.

 

I think your way is much better, so I change the patch and it is  

'regression_20190827.patch' now.

> For the numeric part, this improves the case of

> ApplySortAbbrevFullComparator() where both values are not NULL.  Could

> things be done so as the other code paths are fully covered?  One

> INSERT is fine by the way to add the extra coverage.

There are code lines related to NULL values in

ApplySortAbbrevFullComparator(), but I think the code lines for

comparing a NULL and a NOT NULL can be never reached, because it is

handled in ApplySortComparator() which is called before

ApplySortAbbrevFullComparator(). So I think it is no use to INSERT

a NULL value.

--

Movead