Re: Small suggestion on build script - Mailing list pgsql-hackers

From Andrew Dunstan
Subject Re: Small suggestion on build script
Date
Msg-id 2020.24.211.141.25.1081507110.squirrel@www.dunslane.net
Whole thread Raw
In response to Re: Small suggestion on build script  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane said:
> "Dann Corbit" <DCorbit@connx.com> writes:
>> The response from the Mingw team:
>
>>> Symbolic links to files and directories do not work on Win32
>>> in general.  Support for symlink operation is limited to the source
>>> directory or file existing and being able to copy the source to the
>>> destination.
>
> The source does exist, and they should be able to copy it to the
> destination, and they do manage to do so much of the time.  I think you
> failed to convey the critical point, which is that they randomly fail
> to do what they claim to be able to do.
>
>

I had a couple of thoughts about ways of dealing with this problem in an
acceptable way:
1. make the symlinking more robust by doing it in a loop until it
succeeeds - surely we can test for that rather than just assuming it has
succeeded.
2. or instead of symlinking or making a lasting copy, make the dependency
back to the real file location and the corresponding compile action would
be copy/compile/delete copy.

cheers

andrew




pgsql-hackers by date:

Previous
From: Honza Pazdziora
Date:
Subject: Re: PostgreSQL configuration
Next
From: pgsql@mohawksoft.com
Date:
Subject: Re: PostgreSQL configuration