On 08/07/2017 03:21 PM, Tom Lane wrote:
> Andrew Dunstan <andrew.dunstan@2ndquadrant.com> writes:
>> On 07/31/2017 01:02 PM, Tom Lane wrote:
>>> Record full paths of programs sought by "configure".
>> The problem with this commit, as jacana is demonstrating, is that on
>> Msys it finds the wrong prove. configure needs to run against the perl
>> we build against, i.e. a native Windows perl, but prove needs to run
>> with the perl from the MSys DTK that understands MSys virtualized
>> paths. I have a hack that will allow the buildfarm to overcome the
>> difficulty, (essentially it passes 'PROVE=prove' to make) but that's
>> fairly ugly and certainly non-intuitive for someone running an MSys
>> build and TAP tests without the buildfarm client.
> I'm confused. AFAIK, that commit did not change which "prove" would
> be used --- at least not unless you change PATH between configure and
> make. It only changed how specifically that program would be named in
> Makefile.global. Please clarify how that broke anything.
>
>
That's exactly what we do. See
<https://github.com/PGBuildFarm/client-code/blob/master/run_build.pl> at
line 1649.
cheers
andrew
--
Andrew Dunstan https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services