Posting time bug fix proposal (simplified)

Stuart D. Gathman stuart at gathman.org
Mon Aug 18 14:03:11 EDT 2008


On Mon, 18 Aug 2008, Charles Day wrote:

> That means there is no automatic way to tell a book closing transaction from
> a transaction shifted by the bug. So this patch, as is, would screw up
> anyone who has used the book closing feature. Is there anything in the data
> file, perhaps outside of transactions, that could at least show whether the
> book closing feature has been used?

It takes more than a simple patch, but the date bug can be completely
squashed by simply storing the timezone in the in memory register.  It could be
as compact as a time_t plus a 1 byte index into a timezone table.  All
register dates are displayed (and all scheduling done) relative to the
associated timezone, not the system timezone.  This preserves the date, *and*
the time of day across any changes to the system timezone, and is guaranteed
not to break old versions (any more than they were already broken) or book
closing code.

It might not even be that bad a change to the register.  Add the tz field
to the in memory data, and a tweak to the date display code.

-- 
 	      Stuart D. Gathman <stuart at bmsi.com>
Business Management Systems Inc.  Phone: 703 591-0911 Fax: 703 591-6154
"Confutatis maledictis, flamis acribus addictis" - background song for
a Microsoft sponsored "Where do you want to go from here?" commercial.


More information about the gnucash-devel mailing list