Re: [PATCH] Rough cut at update-move support for properties

2012-12-08 Thread Stefan Sperling
On Sat, Dec 08, 2012 at 12:20:22AM -0800, Bert Huijben wrote: > You don't need a wq item to just update the db. Wq items are to keep > the wc and db synchronized. To keep the db itself synchronized we have > sqlite transactions Duh, of course! Thanks for the cluebat, Bert. So, Julian, your patch

RE: [PATCH] Rough cut at update-move support for properties

2012-12-08 Thread Bert Huijben
t: Re: [PATCH] Rough cut at update-move support for properties In r1418585 I committed a much better version of this, and a test which passes. It's still not right -- it's putting the props into the DB immediately, whereas it should be adding another work item to the work queue instead -- b

Re: [PATCH] Rough cut at update-move support for properties

2012-12-07 Thread Julian Foad
In r1418585 I committed a much better version of this, and a test which passes. It's still not right -- it's putting the props into the DB immediately, whereas it should be adding another work item to the work queue instead -- but it's a start. - Julian >> Julian Foad wrote: >>> I'm just po

Re: [PATCH] Rough cut at update-move support for properties

2012-12-07 Thread Julian Foad
Ben Reser > Julian Foad wrote: >> I'm just posting this so you (Philip et el.) can take a look if you >> want to.  I might proceed by extending the tests to check props and >> then committing it if it works, but really I want to factor it better >> -- maybe by reusing svn_wc_merge5() or the gut

Re: [PATCH] Rough cut at update-move support for properties

2012-12-06 Thread Ben Reser
On Thu, Dec 6, 2012 at 7:34 PM, Julian Foad wrote: > I'm just posting this so you (Philip et el.) can take a look if you want to. > I might proceed by extending the tests to check props and then committing it > if it works, but really I want to factor it better -- maybe by reusing > svn_wc_mer