Tom Lane wrote:
"Dave Page" <dpage@vale-housing.co.uk> writes:
Actually, my gripe about this one is that it wasn't detected promptly.
That patch went in two weeks ago; we should have known about
the problem
within a couple days at most. Seems like the Windows members of the
buildfarm don't run often enough. The whole point of the buildfarm is
to spot problems while the code is still fresh in mind, no?
I think that speaks for the current usage of the cygwin port. Snake runs
native builds daily, but like Magnus and his dev box there's no way I'm
letting Cygwin anywhere near it. Istr that the only vaguely active
Cygwin member is Andrew's laptop.
Well, "lack of interest" is certainly adequate reason to decommission a
port. If we can't find anyone who cares enough about Cygwin to host a
regularly-scheduled buildfarm member, I'm for blowing it off.
We used to have a couple of Cygwin boxes doing regular buildfarm runs. I
don't recall why Jim Buttafuoco stopped running ferret.
I have a shiny new set of components just waiting for me to put them
together in a machine. Sudden trips to Australia and bouts of ill
health have delayed this process far beyond what I wanted. My intention
is to put a couple of VMs on this box, one of which will be Windows, and
will run buildfarm regularly. Of course, if someone wanted to donate a
nice machine, either hosted by me or somewhere else, that would
shortcircuit things :-)
Anyway, the lack of daily Cygwin builds is not permanent.
There are several supported platforms not represented on the buildfarm -
e.g. the one HPUX member has never actually reported any results.
cheers
andrew
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
