I am asking because I suspect the bug may have been introduced just before releasing Beta 19.  I am working on Beta 19b updating the dialogs and other very minor improvements, such as Move to Top and Move to Bottom.  I did several in-depth code reviews to inspect what could have changed since Beta 19 and I am unable to find a single suspicious line of code.  All the code I added is not executed when I am getting this bug, so the bug cannot be the result of the new code I added.

I spent (wasted) the last 3 days chasing this bug.  I have never faced such a programming problem in my life.  I don't recall having a memory corruption bug in my code in more than a decade, and this time, I am having a heap corruption _HEAPBADNODE deep inside the memory allocator.

I would like to know if you ever had a crash using Beta 19.  I will be very grateful if you can describe the context you got the crash, for instance, saving to disk, or closing a dialog, or pressing the Undo button...

Thanks,
-- Dan Morin.

This usually happens if another program starts up such as email auto search or I try to stop the process after all the existing entries have been deleted and it continues to descend with nothing left to delete.

As my tree is very big this seems a waste of time, but I have found that I can only clear one column at a time and have to wait each time for it to descend all 20000 names even though only 500 have entries which I had grouped at the top.

This is due to my obsesion in removing the hundreds of custom tags that I don't need and which take up too much memory and has happened in both 18 and 19 and may not be a bug but a memory problem ?

By the way, I added a new option to delete a custom tag directly from the Table Layout.  With two mouse clicks, you can delete a whole custom tag with all its values.  This improvement will be in Beta 19b (hopefully tomorrow).

Alfi (10/13/2006)

---

Once; yesterday evening (October 12,2006).  My family tree has 51,500 individuals  (Size=XXXL)  !

GenoPro Beta 19 uses a native XML format which requires almost double the amount of memory than Beta 18.  There are huge spikes of memory when saving to disk because of the Unicode to UTF-8 encoding as well as the Zip compression.  In 2007 I will work optimizing this code.  In the meantime, I recommend having enough RAM to handle those memory spikes. I have an [old] family tree of yours (40,835 individuals). GenoPro Beta 19 needs about 400 MB of RAM when saving this family tree to disk.  My recommendation is having at least 512 MB since Windows XP uses about 100 MB, plus you may have other applications running.

A few hours later, I pinned down the bad code and fixed it by adding just a single line of code.  Nevertheless, I modified 70 source files adding data validation and sanity checks to pinpoint where the bug was lurking.  This was the toughest bug I had encountered in my life.

By the way, this bug was introduced in Beta 18(a), so GenoPro  2.b18, 2.b18b, 2.b18c, 2.b18d, 2.b18e, 2.b18f, 2.b18g and 2.b19 are lurking this bug.  I am wrapping up my code to release 2.b19b for tomorrow.