*** SachaS has joined #gnuenterprise
*** SachaS_ has joined #gnuenterprise
*** SachaS has quit IRC
*** reinhard has joined #gnuenterprise
<reinhard> good morning all
*** sjc has joined #gnuenterprise
*** btami has joined #gnuenterprise
<btami> good morning
<reinhard> btami: thanks a lot for your work on reports designer
<reinhard> I think this is a very important and big step!
*** dimas has joined #gnuenterprise
<dimas> hello all
*** dimas has quit IRC
*** dimas has joined #gnuenterprise
*** ninja has joined #gnuenterprise
*** kilo has joined #gnuenterprise
<btami> reinhard: i'm just snooping it around :)
*** dimas has quit IRC
*** sjc has quit IRC
*** dimas has joined #gnuenterprise
*** yure_ has joined #gnuenterprise
*** btami has quit IRC
*** jamest has quit IRC
*** ninja has quit IRC
*** jamest has joined #gnuenterprise
*** yure has quit IRC
*** ninja has joined #gnuenterprise
*** kilo has quit IRC
<jamest> we list the post-query trigger at the block level as being ran once per record loaded
<jamest> but it appears that it's only once per query
<jamest> so are the docs wrong or is forms?
<reinhard> I'd suggest to define docs as wrong
<reinhard> as there is a new trigger now run once per record loaded
<reinhard> ON-RECORDLOADED IIRC
<jamest> actually it's only firing once per query request
<jamest> so i think there is a bug there in any case
<reinhard> what's the difference between once per query and once per query request?
<jamest> as if I attach it to a detail block it only fires once when the master is queried not on each query of the detail
<reinhard> well, anyway IMHO it should fire as many times as the PRE-QUERY trigger to be logical and consistent
<reinhard> ah
<jamest> yeah, i think it's a bug, i'll play around with it today some
<jamest> on-recordloaded works perfectly
<reinhard> yes
<reinhard> but note it was only introduced in last release
*** jcater has quit IRC
<jamest> lol
<jamest> seeing as how I use svn HEAD in production I don't think 1 release old features will bother me too much :)
<jamest> but i'm sure my QA team (aka the users) will report any issues encountered
*** jcater has joined #gnuenterprise
*** ninja has quit IRC
*** btami has joined #gnuenterprise
*** dcmwai has joined #gnuenterprise
*** dcmwai has quit IRC
*** btami has quit IRC
<jamest> reinhard: post commit trigger on blocks is in the same state
<jamest> i think if I add an event like 'dsCommited' to datasources/drivers/Base/RecordSet's _requery  function
<jamest> then hook into that into forms like the PRE-INSERT UPDATE DELETE triggers
<jamest> that would give me what I require
<jamest> is this the right spot in datasources to do that?
<reinhard> you want a record level post-commit trigger?
<reinhard> then probably yes
<reinhard> hmmm
<reinhard> however it might get tricky for details
<jamest> yes
<jamest> i've added in the code
<jamest> but it appears the block and/or datasource isn't pointed to the same record when _requery is called
<jamest> haven't traced out why yet
<jamest> near as I've been able to tell deleted records won't exist by the time we _requery so they aren't an issue, what other issue do you think might make it tricky?
<jamest> ah
<jamest> the requery in the resultset doesn't track the current record w/ the requery
<jamest> so it's always pointed at the last record
*** SachaS_ has quit IRC
*** SachaS has joined #gnuenterprise
<reinhard> yes, however I think it *might* be possible for the resultset to iterate through the records for requery again
<reinhard> the problem I see with the details is that _requery isn't called for details
<reinhard> you would have to do the same for _initialDataFromDict
<reinhard> also you would have to take care, as there are situations where _requery and _initialDataFromDict is done twice
<reinhard> (depending on the commit parameter)
<reinhard> so actually we would have to decide whether the trigger should be fired after the changes have been *posted* or after they have been *committed*
<jamest> in my case i need to rebuild a document from the commited changes
<jamest> i can do this via a db level trigger
<jamest> just figured it was something we might want forms to do
*** sjc has joined #gnuenterprise
<reinhard> I think it's a good idea
<reinhard> however off to bed now
<reinhard> night all
*** reinhard has quit IRC
*** jcater has quit IRC
*** jamest has quit IRC
*** SachaS has quit IRC
*** sjc has quit IRC
*** ninja has joined #gnuenterprise
*** ninja has quit IRC
*** ninja has joined #gnuenterprise