Peter Eisentraut said:
> Claudio Natoli wrote:
>> For application to HEAD, following community review.
>>
>> Supplement to previous win32 patch for dfmgr. Win32 replacement for
>> configure time constants like PKGLIBDIR.
>
> + /*
> + * Determine the directory of installed files, on the assumption that
> + * win32 will have a common root directory for all pgsql
> subdirectories + * (eg. lib, bin, share, etc).
> + */
>
> What happens if this assumption does not hold?
>
>
Things will break ...
The problem is that we need to be able to create a binary installer for
Windows where the location is not known until install time. This is common
Windows practice, and Windows admins rightly hate packages that have
hardcoded install paths. But to do that we need to be able to make a
couple of assumptions about things which in our current setup are
determined at configure time. In practice it is not likely to be a big
problem.
This was discussed in win32-hackers, and the consensus was that this was a
reasonable price to pay, IIRC.
cheers
andrew