No, you're right. I seem to recall getting timeout errors, but we can certainly try it again... I'll remove the build agent requirement from the sql server build.
On Thu, Jul 28, 2011 at 10:02 AM, Richard Brown (gmail) <[email protected]> wrote: >> share the same files on disk. > > Do they?? I thought they had different paths. My bad. That'll be the > problem. > > -----Original Message----- From: Patrick Earl > Sent: Thursday, July 28, 2011 5:01 PM > To: [email protected] > Subject: Re: [nhibernate-development] Alternate Database Builds > > The way the tests are set up, they cannot run concurrently as they > share the same files on disk. I haven't investigated how to fix this, > but it may be as simple as changing the folder in the build > configuration page. > > Patrick Earl > > On Thu, Jul 28, 2011 at 9:55 AM, Julian Maughan > <[email protected]> wrote: >> >> Could we have the SQL Server tests run on a different agent and/or queue, >> so >> they run independently of the other database tests when a commit occurs? >> This would mean the SQL Server tests wouldn't be blocked, and several >> commits could be made and tested on SQL Server in the time it takes to >> perform one run of tests on all the other databases. >> On 28 July 2011 23:37, Richard Brown (gmail) <[email protected]> >> wrote: >>> >>> That's it turned on. >>> >>> Note, the only build this will (currently) affect is Firebird - the >>> others >>> (PostgreSQL and SQLite) run the complete suite correctly (with the >>> appropriate tests ignored programmatically as originally intended). >>> >>> We're hoping to get Oracle up and running shortly too, which will be >>> affected as it currently has failing tests. >>> >>> -----Original Message----- From: Patrick Earl >>> Sent: Wednesday, July 27, 2011 9:23 PM >>> To: [email protected] >>> Subject: Re: [nhibernate-development] Alternate Database Builds >>> >>> I say go for it, I can help clean up the mess. If nothing else, we >>> can temporarily add very problematic tests to the list of "acceptable >>> failures". >>> >>> On Wed, Jul 27, 2011 at 1:43 PM, Richard Brown <[email protected]> >>> wrote: >>>> >>>> It's easy to enable (I've added the two required lines but they're >>>> commented >>>> out just now). >>>> >>>> The only problem is that there's immediately a bunch of tests that fall >>>> into >>>> the category of added-but-failing. :( Not hard - just a logistical >>>> exercise. >>> >> >> > >
