Claudio Natoli said:
>> Strictly speaking, the pwd flags test should probably depend
>> in the test
>> platform rather than the target platform. Something like:
>>
>> case `uname` in
>> MINGW32*)
>> PWDFLAGS=-W;;
>> *)
>> PWDFLAGS=;;
>> esac
>
> There's three other tests that use $host_platform in pg_regress.sh in
> the exact same manner. Not sure I understand the objection.
>
if you tested a mingw build from some other platform pwd might not
understand the -W flag. It's probably a minor nit at best.
cheers
andrew