Scheduled transactions not appearing beyond 19 Dec 2010

[John Ralls]

On Dec 12, 2010, at 12:34 PM, Phil Longstaff wrote:

> On Sun, 2010-12-12 at 12:25 -0800, John Ralls wrote:
>> On Dec 12, 2010, at 9:02 AM, Phil Longstaff wrote:
>> 
>>> On Mon, 2010-11-29 at 23:05 -0500, Volker Englisch wrote:
>>>> On 11/27/2010 11:58 AM, Mark Sansom wrote:
>>>>> Also as an example, if I choose one of my regular scheduled transactions
>>>>> (weekly) and go into Edit Scheduled Transaction, on the Overview tab
>>>>> under Occurrences it says Last Occurred: 09/01/2011, which suggests
>>>>> it should appear in the account, but it doesn't.  The last automatic
>>>>> entry for this weekly transaction is 19/12/2010, and there are no
>>>>> other scheduled transactions appearing after this date.
>>>> 
>>>> Just wondering:  Are you certain that your SX is enabled?
>>>> When you open the transaction editor there is a check box to 
>>>> enable/disable a SX.
>>>> 
>>>> Also, are your transactions listed in the calendar below the list of 
>>>> transactions?  Click on a date in the calendar to see all of the 
>>>> upcoming transactions that are scheduled for that date.
>>>> 
>>> 
>>> I just ran into this as well, where scheduled transactions appeared in
>>> the "Since Last Run" dialog but no transaction was created.  In
>>> my /tmp/gnucash.trace file (on windows, it's under C:\Documents and
>>> Settings\phil\Local Settings\Temp, I think), there was the line:
>>> 
>>> CRIT   Null account kvp value for SX [Transfer to savings], cancelling
>>> creation.
>>> 
>>> which seemed to be new.  I don't know if something was corrupted or if
>>> this is a byproduct of a recent change.  I'll try deleting and
>>> re-creating the sx, after updating my source to include recent fixes.
>>> 
>> 
>> Are you using a sql backend, and if so, with which version was the database created?
> 
> Sqlite3 back end, currently using 2.3.17.  Don't know which version
> created it.  Early 2.3.X and upgraded with new versions.  I haven't
> really checked to see if this is a symptom of the slots change you made
> recently.
> 
> Speaking of the slots change, src/backend/dbi/test has some tests for
> the sql backends.  Basically, creates a session with some data.  Saves
> to sql file.  Reloads into a new session, then compares whether data
> loaded is the same as data saved.  It does test different slot types but
> doesn't included any nested frames (has gint64, double, numeric, string
> and guid).  Maybe one should be added to ensure your fix works and isn't
> broken in the future.

The OP's problem was with the slot handling between 2.3.15 and 2.3.17, fixed with r19911. Because of that warning,
I suspect that you're seeing the same problem.

I've been using the attachment Christian made for 627831 as a round-trip test (save to sql, quit, restart, save to xml, diff the xml and the original). The dbi test seems a bit easier to automate, I'll look at expanding it to cover more cases. I want to expand test coverage in general for engine, qof, and the backends before starting on the not-loading-everything redesign. TDD and all that. Expanding the slots coverage is as good a place to start as any.

Regards,
John Ralls