Alvaro Herrera wrote:
> Bruce Momjian wrote:
>> Alvaro Herrera wrote:
>
>>> Maybe this means that we should give pgindent a run over the back
>>> branches.
>> Yea, that thought has crossed our minds, but the problem is that there
>> is little testing of back branches so it would be risky.
>
> That's a fair point, though I wonder how can a code indenter be so
> broken that it broke code in ways not detected by our buildfarm.
Something like this:
if (foo)
{do something;do something else;
}
...
->
if (foo)do something;
do something else;
...
I wouldn't be too surprised if something like that happened with one of
the complex macros, like PG_TRY/CATCH.
-- Heikki Linnakangas EnterpriseDB http://www.enterprisedb.com